You are on page 1of 215

CHEST PAIN UNITS

CONTENTS

Foreword xi
Michael H. Crawford

Preface xiii
Ezra A. Amsterdam and J. Douglas Kirk

Acute Coronary Syndromes: Risk Stratification and Initial Management 401


Christopher P. Cannon
For patients who have acute coronary syndromes (ACS), risk stratification is key to ini-
tiating appropriate treatment. For ST-segment elevation MI, immediate reperfusion
therapy is needed, and thus rapid identification of ST elevation on the ECG is critical.
Then, having a standardized protocol for rapid treatmentwith either primary percuta-
neous coronary intervention or thrombolysisis critical. For unstable angina/nonST
elevation ACS, after first identifying the patients who have a higher likelihood of actu-
ally having an ACS (as opposed to noncardiac chest pain) stratification to high versus
lower risk is needed to choose appropriate therapies. Thus, it is important for risk stra-
tification to be a central part of all management of patients who have ACS.

Chest Pain Unit Concept: Rationale and Diagnostic Strategies 411


Andra L. Blomkalns and W. Brian Gibler
Each year in the United States, over 8 million patients present to the emergency depart-
ment (ED) with complaints of chest discomfort or other symptoms consistent with pos-
sible acute coronary syndrome (ACS). While over half of these patients are typically
admitted for further diagnostic evaluation, fewer than 20% are diagnosed with ACS.
With hospital beds and inpatient resources scarce, these admissions can be avoided by
evaluating low- to moderate-risk patients in chest pain units. This large, undifferentiated
patient population represents a potential high-risk group for emergency physicians re-
quiring a systematic approach and specific ED resources. This evaluation is required
to appropriately determine if a patient is safe to be discharged home with outpatient
follow-up versus requiring admission to the hospital for monitoring and further testing.

Acute Coronary Syndromes in the Emergency Department: Diagnostic


Characteristics, Tests, and Challenges 423
J. Hector Pope and Harry P. Selker
Failure to diagnose patients who have acute coronary syndromes (ACSs)either acute
myocardial infarction (AMI) or unstable angina pectoris (UAP)who present to the

VOLUME 23 NUMBER 4 NOVEMBER 2005 v


emergency department (ED) remains a serious public health issue. Better understanding
of the pathophysiology of coronary artery disease has allowed the adoption of a unifying
hypothesis for the cause of ACSs: the conversion of a stable atherosclerotic lesion to a
plaque rupture with thrombosis. Thus, physicians have come to appreciate UAP and
AMI as parts of a continuum of ACSs. This article reviews the state of the art regarding
the diagnosis of ACSs in the emergency setting and suggests reasons why missed diag-
nosis continues to occur, albeit infrequently.

Use of Biomarkers in the Emergency Department and Chest Pain Unit 453
Allan S. Jaffe
The use of biomarkers of cardiac injury in the emergency department (ED) and observa-
tion unit settings has several nuances that are different and, therefore, worthy of its own
set of use guidelines. The markers that are used, however, are the same. The primary
marker of choice continues to be cardiac troponin (Tn). Other markers that have been
used because of the need in the ED for rapid triage have been myoglobin and fatty acid
binding protein. In addition, some centers still prefer less sensitive and less specific
markers such as creatine kinase myocardial band (CK-MB). More recently, a push has
occurred to develop markers of ischemia, such as ischemia modified albumin (IMA),
to determine which patients have ischemia, even in the absence of cardiac injury. As tro-
ponin assays become more sensitive and method for use becomes better understood, the
use of these other markers are being relegated to lesser and lesser roles. Markers of ische-
mia are useful, but at present, despite some enthusiasm, are not ready for routine use.
Before describing the recommendations for clinical use of biomarkers in the ED, a basic
understanding of some of the science and measurement issues related to these analytes is
helpful.

Point-of-Care Testing and Cardiac Biomarkers: The Standard of Care and


Vision for Chest Pain Centers 467
Gerald J. Kost and Nam K. Tran
Point-of-care testing (POCT) is defined as testing at or near the site of patient care. POCT
decreases therapeutic turnaround time (TTAT), increases clinical efficiency, and improves
medical and economic outcomes. TTAT represents the time from test ordering to patient
treatment. POC technologies have become ubiquitous in the United States, and, there-
fore, so has the potential for speed, convenience, and satisfaction, strong advantages
for physicians, nurses, and patients in chest pain centers. POCT is applied most benefi-
cially through the collaborative teamwork of clinicians and laboratorians who use inte-
grative strategies, performance maps, clinical algorithms, and care paths (critical
pathways). For example, clinical investigators have shown that on-site integration of
testing for cardiac injury markers (myoglobin, creatinine kinase myocardial band [CK-
MB], and cardiac troponin I [cTnI]) in accelerated diagnostic algorithms produces effec-
tive screening, less hospitalization, and substantial savings. Chest pain centers, which
now total over 150 accredited in the United States, incorporate similar types of protocol-
driven performance enhancements. This optimization allows chest pain centers to im-
prove patient evaluation, treatment, survival, and discharge. This article focuses on
cardiac biomarker POCT for chest pain centers and emergency medicine.

Markers of Cardiac Ischemia and Inflammation 491


Tracy Y. Wang, Wael A. AlJaroudi, and L. Kristin Newby
Because biomarkers of myocardial necrosis only become positive in the setting of myo-
cardial necrosis and disruption of cellular integrity, the diagnosis of myocardial infarc-
tion can only be made in retrospect. Ideally, one would like to identify patients at risk
for complications before myocardial necrosis occurs. Insights into the pathophysiology
of atherothrombosis have allowed development of novel markers to detect not only early
ischemia without myocyte death but also early indicators of coronary inflammation in
patients who have preclinical atherosclerosis.

vi CONTENTS
Exercise Testing in Chest Pain Units: Rationale, Implementation, and Results 503
Ezra A. Amsterdam, J. Douglas Kirk, Deborah B. Diercks, William R. Lewis,
and Samuel D. Turnipseed
Chest pain units are now established centers for assessment of low-risk patients pre-
senting to the emergency department with symptoms suggestive of acute coronary
syndrome. Accelerated diagnostic protocols, of which treadmill testing is a key
component, have been developed within these units for efficient evaluation of these pa-
tients. Studies of the last decade have established the utility of early exercise testing,
which has been safe, accurate, and cost-effective in this setting. Specific diagnostic pro-
tocols vary, but most require 6 to12 hours of observation by serial electrocardiography
and cardiac injury markers to exclude infarction and high-risk unstable angina before
proceeding to exercise testing. However, in the chest pain unit at UC Davis Medical Cen-
ter, the approach includes "immediate" treadmill testing without a traditional process to
rule out myocardial infarction. Extensive experience has validated this approach in a
large, heterogeneous population. The optimal strategy for evaluating low-risk patients
presenting to the emergency department with chest pain will continue to evolve based
on current research and the development of new methods.

Imaging in the Evaluation of the Patient with Suspected Acute


Coronary Syndrome 517
Michael C. Kontos and James L. Tatum
Over the last decade, major advances have been made in the treatment of acute coronary
syndromes (ACSs). However, effective implementation of these treatments requires
timely and accurate identification of the high-risk patient among all those presenting
to the emergency department (ED) with symptoms suggestive of ACS. The opportunity
for improving outcomes is time-dependent, so that early identification of the patient
who has true ACS is essential. This necessity further increases the need for rapid triage
tools, especially in the current setting of ED and hospital overcrowding that has become
the norm in large urban centers.

Echocardiography in the Evaluation of Patients in Chest Pain Units 531


William R. Lewis
Using percutaneous angioplasty to induce the ischemic cascade in the cardiac catheter-
ization laboratory, echocardiographic wall motion abnormalities have been documented
to precede electrocardiographic abnormalities and angina. Therefore, detection of car-
diac wall motion abnormalities is potentially more sensitive than the history, physical
examination, and ECG for identification of myocardial ischemia. Echocardiography is
highly reliable for assessing cardiac wall motion and, thus, it has been used for diagnosis
and risk assessment in patients presenting to the emergency department (ED) with
symptoms suggestive of myocardial ischemia. In patients who have acute ST-elevation
myocardial infarction (MI), echocardiography is comparable to invasive left ventriculo-
graphy for detecting wall motion abnormalities. However, the usefulness of echocardio-
graphy in the low-risk population that has chest pain of uncertain origin and a
nondiagnostic initial presentation is less well established.

Newer Imaging Methods for Triaging Patients Presenting to the Emergency


Department with Chest Pain 541
James McCord and Ezra A. Amsterdam
The usefulness of electron beam CT (EBCT) for the risk stratification of patients in the
emergency department (ED) who have possible acute coronary syndrome has been eval-
uated in three small studies. The results of these studies are promising, as patients who
have no coronary calcium detected by EBCT essentially had no adverse cardiac events.

CONTENTS vii
Although the negative predictive value of EBCT was excellent, the limited positive pre-
dictive value that would lead to further diagnostic testing makes this strategy less attrac-
tive if applied to a broad population. Further larger studies may help define which
patients in the ED who have chest pain and nondiagnostic ECGs can be effectively eval-
uated by EBCT. Recent advances in noninvasive coronary angiography by multislice
computed tomography are of considerable interest in the ED evaluation of patients with
undefined chest pain, but the utility of this method in this setting awaits clinical studies.

Chest Pain Units: Management of Special Populations 549


Deborah B. Diercks, J. Douglas Kirk, and Ezra A. Amsterdam
Chest pain units provide an important alternative to traditional hospital admission for
patients who present to the emergency department with symptoms compatible with
acute coronary syndrome and a normal or inconclusive initial evaluation. Although pa-
tient subgroups such as women, diabetics, those with established coronary artery dis-
ease, and those with symptoms related to stimulant use present unique challenges,
management in a chest pain unit appears to be appropriate in these populations. Judi-
cious application of accelerated diagnostic protocols and current testing methods can
promote safe, accurate, and cost-effective risk stratification of special populations to
identify patients who can be safely discharged and patients who require hospital admis-
sion for further evaluation.

Management of the Patient with Chest Pain and a Normal Coronary


Angiogram 559
Eric H. Yang and Amir Lerman
Angina in the setting of a normal coronary angiogram (NOCAD) occurs in 20% to 30% of
patients undergoing coronary angiography. The management of these patients can be
challenging and requires a correct diagnosis of the etiology. The differential diagnosis
of NOCAD can be classified anatomically into three categories: (1) epicardial disease,
(2) coronary microvascular dysfunction, and (3) noncoronary disease. The pathophysiol-
ogy of NOCAD and a systematic diagnostic approach to these patients is reviewed. Po-
tential therapeutic options and prognosis are also discussed.

Using the Emergency Department Clinical Decision Unit for Acute


Decompensated Heart Failure 569
W. Frank Peacock
Acute decompensated heart failure (ADHF) is a complex disease of epidemic propor-
tions. In the United States, it accounts for more than 1 million hospitalizations annually,
and heart failure represents the single greatest cost to the Centers for Medicaid and Medi-
care Studies. Half of the annual costs are estimated to be the result of hospitalization.
Compared with other pathology, heart failure has a very high hospitalization rate, with
80% of emergency department ADHF patients being admitted. This high rate has re-
sulted from the lack of successful management predictors available to the emergency
physician and the lack of any disposition option other than hospitalization for the ADHF
patient. The emergency department observation unit offers an alternative to hospitaliza-
tion for patients with ADHF. Validated protocols have demonstrated that in ADHF, in-
tensive short-term therapeutic, diagnostic, and educational protocols result in a marked
improvement in hospitalization rates, while at the same time decreasing costs. New risk
stratification data can aid in the identification of the appropriate candidate. The observa-
tion unit now represents a nonhospitalization disposition option for patients presenting
to the emergency department with ADHF.

viii CONTENTS
Cost Effectiveness of Chest Pain Units 589
Sandra Sieck
Health care facilities face many challenges in their attempts to provide cost-effective care
without sacrificing quality. One key factor in producing quality outcomes while main-
taining economic profitability is the establishment of a cost-effective outpatient care en-
vironment. Chest Pain Units (CPUs) have evolved to provide a streamlined approach to
acute cardiac care that emphasizes optimal efficiency initiated at the point of entry. The
Centers for Medicare and Medicaid Services have structured new reimbursement ap-
proaches designed to shift care from the inpatient setting and reward efficient and ap-
propriate care delivered in the outpatient arena. These new reimbursement strategies
have transformed the CPU into an economically viable entity for the acute care facility
and also have afforded opportunities to enhance the quality of care delivered to the acute
cardiac patient.

Medical Error Prevention in ED Triage for ACS: Use of Cardiac Care


Decision Support and Quality Improvement Feedback 601
Denise H. Daudelin and Harry P. Selker
Medical errors in the care of patients who present with acute coronary syndrome (ACS)
include errors in emergency department (ED) triage, such as the decision to send home a
patient who presents with ACS or to hospitalize a patient who does not have ACS to the
cardiac care unit (CCU), and errors in treatment, such as the failure to promptly use re-
perfusion therapy for patients who present with ST-elevation acute myocardial infarc-
tion (AMI). ECG-based acute cardiac ischemia time-insensitive predictive instrument
(ACI-TIPI) and thrombolytic predictive instruments (TPIs), with a linked TIPI informa-
tion system (TIPI-IS), provide real-time, concurrent, and retrospective decision support
tools and feedback for the prevention of medical errors in the care of patients who pre-
sent with ACS. In real-time, ACI-TIPI probabilities printed on the ECG header for the ED
physician, provide an additional piece of information for triage decision making, and the
ACI-TIPI Risk Management form reduces liability risk by prompting consideration and
documentation of key clinical factors in the diagnosis of ACI. Also in real-time, the TPI
increases overall coronary reperfusion therapy use. Concurrent flagging by TIPI-IS uses
electronically acquired ECG and hospital data to provide concurrent alerts about po-
tential misdiagnosis or mis-triage of patients with ACS. Retrospectively TIPI-IS-based
feedback reports allow performance improvement. These examples of information tech-
nology tools integrated into ECG equipment already used in hospitals to deliver patient
care demonstrate the potential to adapt other existing equipment or other patient care
activities to enhance patient safety and error reduction.

Cumulative Index 2005 615

CONTENTS ix
FORTHCOMING ISSUES

February 2006
Emergency Cardiac Care: From ED to CCU
Amal Mattu, MD, and Mark Kelemen, MD, MSc
Guest Editors

May 2006
Interventional Cardiology
Samin K. Sharma, MD, Guest Editor

August 2006
Advanced 12-Lead Electrocardiography
S. Serge Barold, MD, Guest Editor

RECENT ISSUES

August 2005
Cardio-Renal Disease
Ragavendra R. Baliga, MD, MBA, FACC, FRCP(Edin),
Sanjay Rajagopalan, MD, and
Rajiv Saran, MD, MS, MRCP, Guest Editors

May 2005
Therapeutic Strategies in Diabetes
and Cardiovascular Disease
Prakash C. Deedwania, MD, FACC, FACP, FCCP, FAHA
Guest Editor

February 2005
Patent Foramen Ovale
Edward A. Gill, Jr, MD, and John D. Carroll, MD
Guest Editors

VISIT OUR WEB SITE

The Clinics are now available online!


Access your subscription at www.theclinics.com
Cardiol Clin 23 (2005) xi

Foreword
Chest Pain Units

Michael H. Crawford, MD
Consulting Editor

Chest pain units have been enthusiastically em- including articles on cost eectiveness, medical
braced by some medical centers and disdained as errors, and the employment of various types of
marketing ploys by critics. Whether you have one diagnostic imaging. New noninvasive methods for
or not, this issue should be of interest. If you al- imaging the coronary arteries are adding a new
ready have one, you may get some ideas for dimension to the evaluation of chest pain patients.
improvement or widening the scope of your opera- Newer biomarkers are also changing the triage of
tion. If you dont have one, this issue will help with patients with chest pain. They are discussed in
the decision of whether to pursue establishing one. three articles. After more than 102 years of use, is
One clear benet of chest pain centers is that they the ECG ready for the medical museum? Will
decrease unnecessary hospital admissions by al- biomarkers and imaging replace the ECG in
lowing patients to be evaluated further without triaging chest pain patients? Read on.
bogging down the emergency department or being
admitted. As pointed out in one of the articles, this Michael H. Crawford, MD
concept is not unique to chest pain but can also be Division of Cardiology
applied to heart failure exacerbation patients. Department of Medicine
Drs. Amsterdam and Kirk from cardiology University of California
and emergency medicine, respectively, at the San Francisco Medical Center
University of California at Davis School of 505 Parnassus Avenue, Box 0124
Medicine are leaders in this area and both are San Francisco, CA 94143-0124, USA
very active in the growing Society of Chest Pain
E-mail address: michael.crawford@
Centers. They have assembled a panel of experts
ucsfmedctr.org
to cover every aspect of chest pain centers,

0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.09.012 cardiology.theclinics.com
Cardiol Clin 23 (2005) xiiixiv

Preface
Chest Pain Units

Ezra A. Amsterdam, MD, FACC J. Douglas Kirk, MD, FACEP


Guest Editors

Although time-dependent therapy is an over- of this year. Although these units dier in certain
arching principle of the management of acute cor- respects, their common basis is a protocol-driven
onary syndromes (ACS), fulllment of this patient evaluation by an accelerated diagnostic
concept remains incomplete [1]. Delays in recogni- strategy, including clinical observation, sequential
tion of ACS and initiation of therapy entail pa- ECGs, and serial cardiac injury markers. These
tient, physician, and systemic factors [2]. This basic methods are being extended by rapidly
problem is particularly pertinent to the treatment emerging technologies with the promise of earlier
of non-ST elevation (non-STE) ACS in which the and more accurate diagnostic capability, thereby
history and ECG may be ambiguous. Although further enhancing the use of CPUs. This volume
the high-risk patient is the primary clinical focus, of Cardiology Clinics encompasses these advances
most patients who present with chest pain have in presenting the contemporary status of CPUs
a noncardiac etiology that is frequently benign. and associated methodology from the perspectives
The clinical dilemma posed by this situation has of an outstanding group of clinician scientists who
often been resolved by unnecessary admissions have made major contributions to this eld.
to avert missed diagnoses of ACS, resulting in In his introductory chapter, Cannon synthe-
suboptimal patient care and resource use [3]. sizes current understanding of ACS in terms
The development and growth of chest pain units pathophysiology, clinical presentation, and risk
(CPUs) is a response to this problem by applica- stratication. To achieve the latter goal, he con-
tion of a systematic approach to optimize man- siders the roles of basic clinical tools, guidelines,
agement of patients presenting with symptoms and critical pathways while maintaining focus on
compatible with ACS by aording: (1) prompt the individual patient for optimal management.
identication and treatment of those with an Blomkalns and Gibler elucidate the rationale of
ischemic syndrome and (2) early and accurate the CPU concept and the practical aspects of its
discharge of individuals without evidence of myo- implementation. They relate how these strategies
cardial ischemia. can dier based on institutional variation in
Through its recently implemented process, the methods of stang, structure, and administration
Society of Chest Pain Centers has accredited to meet the CPU goals. In the next two articles,
almost 150 CPUs in this country [4] and, based Pope and Selker (1) delve further into the di-
on the current growth rate, it is anticipated that agnostic challenges and current systems for eval-
this number will reach 200 or more by the end uating patients presenting with chest pain and (2)
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.017 cardiology.theclinics.com
xiv PREFACE

explore the crucial problem of missed ACS and decompensated heart failure, which is detailed by
inadvertent discharge from the emergency depart- Peacock in his presentation of the rationale, in-
ment. In the latter context, they identify the dications, and results of this innovation.
principal factors associated with these clinical Reduction of medical errors is an imperative of
errors and methods to prevent them. The pivotal the CPU and Daudelin and Selker clarify current
role of cardiac injury markers is developed by Jae research in this area. Although their clinical utility
who describes the most recent advances in this area has been demonstrated, the complex question of
and charts a rational application based on an whether CPUs are cost-eective has received
understanding of the value and limitations of these limited attention. Sieck reviews this issue in terms
essential clinical tools. These concepts are ex- of its medical and administrative aspects and their
tended by Kost and Tran who analyze the rapidly relation to the larger framework of the health care
advancing area of point-of-care testing for cardiac system.
injury markers and the remarkably rapid results This volume provides a current understanding
while maintaining diagnostic accuracy. Wang, of the implementation and ecacy of CPUs and
Aroudi, and Newby then explore the increased the integration of new technology into this ap-
understanding of inammation in ACS and the proach. These results have provided the basis for
growing body of markers that reect this process, continuing advances toward the optimal strategy
some of which are already being applied clinically. for the management of patients presenting to the
The next series of articles is devoted to the emergency department with chest pain.
varieties of cardiac stress testing and imaging as
applied in the CPU setting. Our group reviews the Ezra A. Amsterdam, MD, FACC
extensive clinical experience in multiple centers Professor, Department of Internal Medicine
conrming the utility of treadmill testing as the Associate Chief (Academic Aairs)
nal element of accelerated diagnostic protocols Division of Cardiovascular Medicine
and we note our results with the immediate exercise University of California School of Medicine
test in low-risk patients. The central role of myo- (Davis) Medical Center, Sacramento, CA, USA
cardial scintigraphy in many CPUs is addressed by E-mail address: eaamsterdam@ucdavis.edu
Kontos and Tatum who demonstrate the ecacy of
this method in their innovative approach to risk J. Douglas Kirk, MD, FACEP
stratication. Rest and stress echocardiography has Associate Professor and Vice Chair
had substantial application in the assessment of Department of Emergency Medicine
patients who present with chest pain and the Medical Director, Chest Pain Service
indications and accuracy of this technique are University of California School of Medicine
evaluated by Lewis. New imaging methods, such (Davis) Medical Center, Sacramento, CA, USA
as coronary artery calcium scoring and the emer- E-mail address: jdkirk@ucdavis.edu
gence of noninvasive coronary angiography by
computed tomography, are receiving major atten-
tion and McCord and Amsterdam provide a per-
spective on the potential of these techniques.
Because of their unique clinical characteristics, References
populations such as women, the elderly, and those [1] Braunwald E, Antman EM, Beasley JW, et al. ACC/
with documented coronary heart disease, have AHA 2002 guideline update for the management of
required specic investigation to determine if the patients with unstable angina and non-ST-segment
CPU concept applies to their presentations. Diercks elevation myocardial infarction. Available at: http://
and colleagues demonstrate that the CPU strategy www.acc.org/clinical/guidelines/unstable/unstable.pdf.
is appropriate in these other groups with special [2] Gibler WB, Cannon CP, Blomkalns AL, et al. Prac-
clinical features. A large proportion of patients who tical implementation of the guidelines for unstable
present to the emergency department with chest angina/non-ST-segment elevation myocardial infarc-
tion in the emergency department. Ann Emerg Med
pain do not have obstructive coronary artery
2005;46:18597.
disease. Yang and Lerman delineate these distinc- [3] Amsterdam EA, Lewis WR, Kirk JD, et al. Acute is-
tive syndromes and impart a systematic approach chemic syndromes. Chest pain center concept. Car-
to their etiology, prognosis, and management. A diol Clin 2002;20:11736.
new dimension that has recently been added to the [4] Society of Chest Pain Centers. Available at: http://
CPU is the management of patients with acute www.scpcp.org/accreditation/accreditedlist.html
Cardiol Clin 23 (2005) 401409

Acute Coronary Syndromes: Risk Stratication


and Initial Management
Christopher P. Cannon, MD
Cardiovascular Division, Brigham and Womens Hospital, 75 Francis Street, Boston, MA 02115, USA

In the United States, there are approximately treatments recommended for all patients. The
1.68 million patients admitted every year to treatments recommended include anti-ischemic,
hospitals with acute coronary syndromes (ACSs) antithrombotic therapy and what strategy should
[1]. Of these, one quarter present with acute myo- be followed (ie, whether to pursue an invasive
cardial infarction (MI) associated with ECG versus conservative approach), and reperfusion
ST-segment elevation (STEMI), whereas three therapy for STEMI.
quarters, or approximately 1.3 million patients,
have unstable angina/nonST elevation MI (UA/
NSTEMI) [1]. The former is most commonly Assessment of likelihood of coronary artery
caused by acute total occlusion of a coronary disease
artery and therefore urgent reperfusion is the
mainstay of therapy, whereas UA/NSTEMI is Approximately 6 to 7 million persons per year
usually associated with a nonocclusive thrombus in the United States present to EDs or chest pain
[2]. Among patients who have UA/NSTEMI, be- units with a complaint of chest pain or other
tween 40% and 60% will have evidence of myo- symptoms suggestive of possible ACS. Of these,
cardial necrosis with elevated troponin, and are approximately 20% to 25% have a nal diagnosis
thus diagnosed with a NSTEMI [3,4]. of unstable angina or MI [5,6]. Thus, the rst step
In the past several years, there have been many in evaluating patients who have possible UA/
advances in the diagnosis and management of this NSTEMI is to determine the likelihood that coro-
patient population, as summarized in part in the nary artery disease (CAD) is the cause of the
2002 update of the American College of Cardiol- presenting symptoms. The 2002 ACC/AHA
ogy (ACC) and the American Heart Association guidelines list factors associated with increased
(AHA) UA/NSTEMI guidelines, and the 2004 likelihood that the patient actually has unstable
ACC/AHA STEMI guidelines. angina (Table 1).
The approach to the patient, especially at the Numerous prediction rules using clinical and
rst presentation in the emergency department ECG variables have also been developed to assess
(ED) or chest pain unit, involves careful risk the likelihood of CAD in patients presenting to
stratication. This assessment actually involves the ED with chest pain [714]. One algorithm, the
two steps: assessment of the likelihood that the acute cardiac ischemia time-insensitive prediction
patients symptoms represent ACS (as opposed to instrument (ACI-TIPI) has been integrated into
noncardiac chest pain), and risk stratication of ECG devices. It provides a quantitative likelihood
the patients who have ACS to identify high- of the patient having ACSs (ie, STEMI and UA/
versus lower-risk patients. The guidelines recom- NSTEMI) [12,13]. One randomized trial has
mend that specic therapies and treatment can be shown that ACI-TIPI reduced unnecessary hospi-
targeted to only high-risk patients, with other tal and coronary care unit admissions [13]. Other
studies have suggested that good clinical judgment
is as good as the computer algorithms [9,10], em-
E-mail address: cpcannon@partners.org phasizing the importance of obtaining a good
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.006 cardiology.theclinics.com
402 CANNON

Table 1
Estimating the likelihood that the suspected acute coronary syndrome event is secondary to coronary artery disease
High likelihood Intermediate likelihood Low likelihood
Feature Any below: No high-likelihood features No high- or intermediate-
but any below: likelihood features but
may have:
History Typical angina; known history Probable angina; age O70 Atypical symptoms
of coronary artery disease years; male diabetes
including myocardial infarction mellitus
Examination Congestive heart failure Peripheral vascular disease, Pain on plapation
cerebrovascular accident
ECG New ECG changes Old ECG abnormalities Normal ECG
Cardiac markers Positive Normal Normal
From Braunwald E, Mark DB, Jones RH, et al. Unstable angina: diagnosis and management. Rockville MD: Agency
for Health Care Policy and Research and the National Heart, Lung, and Blood Institute, US Public Health Service, US
Department of Health and Human Services; 1994; AHCPR Publication No. 94-0602.

clinical history when assessing patients who have to higher-risk patients. Factors associated with
chest pain. a high risk for death or nonfatal MI in patients
who have UA/NSTEMI are prolonged rest pain,
ST-segment changes, elevated cardiac biomarkers
Risk stratication
(eg, troponin), diabetes, evidence of congestive
Based on data from a global registry, the heart failure, and age over 75 years. Low-risk
outcomes of patients who have ACS fall on patients present without rest pain [22], ECG
a spectrum of risk, ranging from 30-day mortality changes, or evidence of heart failure (see Box 3).
of 1.7% for patients who have unstable angina to Other factors are highlighted as markers of
7.4% for those who have NSTEMI and 11.1% for long-term risk, such as extent of CAD, left ven-
those who have STEMI [15]. Within clinical trials tricular dysfunction, and elevated markers of
in which inclusion criteria select higher-risk pa- inammation.
tients, rates of death by 30 days ranged from In UA/NSTEMI, use of troponin alone has
3.5% to 4.5% and of new MI from 6% to 12% been a powerful tool for risk stratication and
[1618]. However, patents who have UA/ targeting of therapies. Studies have found that
NSTEMI constitute a wide spectrum of patients high-risk patients benet from the more aggres-
and risk [1921]; thus, the ACC/AHA guidelines sive treatments. For low molecular weight he-
strongly recommend risk stratication as impor- parin (LMWH), glycoprotein (GP) IIb/IIIa
tant to characterize the patients prognosis and inhibitors, and an early invasive strategy, there
to select appropriate therapy. Boxes 1 through 3 is a greater benet of these interventions in pa-
list the three recommended stratications for tients who have a positive troponin, and almost
risk of death/MI in UA/NSTEMI. no benet in patients who have a negative
The rationale for risk stratication is to target troponin [3,2326]. For example, with the GP
some of the more aggressive therapies to the IIb/IIIa inhibitors, there was a 50% to 70% re-
higher-risk patients. This approach has been seen duction in death or MI in patients who were tro-
in multiple studies with several of the agents for ponin-positive and were receiving GP IIb/IIIa
management of UA/NSTEMI. In STEMI, most of inhibitors compared with patients not receiving
the treatments are similar based on the patients these agents, with no benet of GP IIb/IIIa in-
risk, beginning with reperfusion therapy and in- hibitors in those who did not have a positive tro-
cluding antithrombotic and anti-ischemic thera- ponin [24,25]. A dierent pattern has been seen
pies, which are similar to that of UA/NSTEMI. with the oral antiplatelet agents, in that aspirin
Thus, among patients who have been identied and clopidogrel benet low-, intermediate-, and
as having a moderate or high likelihood of having high-risk patients and those who have positive
an ACS, the ACC/AHA guidelines highlight risk or negative cardiac markers [2729]. Thus, aspi-
stratication as a key step that can target the rin and clopidogrel have been recommended
intensity of medical and interventional therapies for all patients regardless of risk, whereas GP
ACUTE CORONARY SYNDROMES 403

Box 1. Short-term high risk for death/ Box 2. Short-term intermediate risk for
myocardial infarction in unstable death/myocardial infarction in unstable
angina/NSTEMI angina/NSTEMI

At least one of: No high-risk features, but must have


History one of:
Accelerating tempo of ischemic History
symptoms in preceding 48 hours Prior myocardial infarction, peripheral
vascular disease, or cerebrovascular
Character of pain disease
Prolonged ongoing (>20 min) Coronary artery bypass grafting
rest pain Prior aspirin use

Clinical findings Character of pain


Pulmonary edema, most likely caused Prolonged (>20 min) rest angina,
by ischemia now resolved, with moderate or
New or worsening mitral regurgitation high likelihood of coronary artery
murmur disease
S3 or new/worsening rales Rest angina (<20 min) or relieved with
Hypotension, bradycardia, tachycardia rest or sublingual nitroglycerin
Age over 75 years
Clinical findings
ECG Age over 70 years
Rest angina with transient ST changes
greater than 0.05 mV ECG
Bundle branch block T-wave inversions greater than 0.2 mV
New sustained ventricular tachycardia Pathologic Q waves

Cardiac markers Cardiac markers


Markedly elevated (eg, Troponin T or Slightly elevated (eg, Troponin
Troponin I >0.1 ng/mL) T >0.01 ng/mL but <0.1 ng/mL)

From Braunwald E, Antman EM, Beasley From Braunwald E, Antman EM, Beasley
JW, et al. ACC/AHA guidelines for the man- JW, et al. ACC/AHA guidelines for the man-
agement of patients with unstable angina agement of patients with unstable angina
and non-ST-segment elevation myocardial and non-ST-segment elevation myocardial
infarction: a report of the American College infarction: a report of the American College
of Cardiology/American Heart Association of Cardiology/American Heart Association
Task Force on Practice Guidelines (Committee Task Force on Practice Guidelines (Committee
on the Management of Patients With Unstable on the Management of Patients With Unstable
Angina). J Am Coll Cardiol 2000;36:9701062; Angina). J Am Coll Cardiol 2000;36:9701062;
with permission. with permission.

IIb/IIIa inhibitors are recommended for use only developed using clinical variables and ECG and
in high-risk patients or those undergoing percu- initial serum cardiac marker data [30,31]. The
taneous coronary intervention (PCI) (Fig. 1). Thrombolysis in Myocardial Infarction (TIMI)
risk score was developed using multivariate analysis
to predict the occurrence of death, MI, or recurrent
Risk scores
ischemia leading to urgent revascularization in the
Integrating all the risk factors identied in the TIMI 11B trial. Seven independent risk factors
ACC/AHA guidelines as low-, intermediate-, or emerged: age over 65 years, more than three
high-risk, comprehensive risk scores have been risk factors for CAD, documented CAD at
404 CANNON

Definite ACS
With Cath and PCI Likely/Definite Possible
Box 3. Short-term low risk for death/ or High-risk (IIa) ACS ACS
myocardial infarction in unstable angina
Aspirin
Aspirin Aspirin
No high or intermediate risk feature, but + +
IV heparin/SQ LMWH* SQ LMWH*
may have any of: + or
History IV platelet IV heparin
GP IIb/IIIa
None indicated antagonist clopidogrel

clopidogrel
Character of pain
New-onset Canadian Cardiovascular Fig. 1. 2002 American College of Cardiology/American
Society class III or IV angina Heart Association guideline recommendations for
in past 2 weeks without prolonged antithrombotic therapy for unstable angina/nonST
(>20 min) rest pain, but with elevation acute coronary syndrome. (From Braunwald
moderate or high likelihood of E, Antman EM, Beasley JW, et al. ACC/AHA guide-
coronary artery disease lines for the management of patients with unstable angina
and non-ST-segment elevation myocardial infarction:
Clinical findings a report of the American College of Cardiology/American
Heart Association Task Force on Practice Guidelines
None indicated
(Committee on the Management of Patients With Un-
stable Angina). J Am Coll Cardiol 2000;36:9701062;
ECG with permission.)
Normal or unchanged ECG during
chest discomfort

Cardiac markers
Normal Antithrombotic and medical therapy
Initial treatment for patients who have UA/
From Braunwald E, Antman EM, Beasley NSTEMI should include aspirin, which leads to
JW, et al. ACC/AHA guidelines for the man- a 50% to 70% reduction in death or MI as
agement of patients with unstable angina compared with placebo [34]. Current data on aspi-
and non-ST-segment elevation myocardial rin indicate the drug is benecial for long-term
infarction: a report of the American College treatment at doses as low as 75 mg/d [27]. New
of Cardiology/American Heart Association
data from the Clopidogrel in Unstable Angina to
Task Force on Practice Guidelines (Committee
on the Management of Patients With Unstable
Prevent Recurrent Events (CURE) trial indicates
Angina). J Am Coll Cardiol 2000;36:9701062; that lower doses of aspirin (eg, 81 mg) are associ-
with permission. ated with a 50% lower rate of major bleeding over
1 year of treatment than doses of 200 to 325 mg
[35]. Thus, for acute management in-hospital, 160
to 325 mg daily is recommended, but at hospital dis-
catheterization, prior ASA, more than two episodes charge and during follow-up, the new data suggest
of angina in the last 24 hours, ST deviation more a dose of 81 mg may be safer.
than 0.5 mm, and elevated cardiac markers. This The 2002 ACC/AHA guideline included a new
scoring system was able to risk stratify patients Class I recommendation for the use of clopidogrel
across a ten-fold gradient of risk, from 4.7% to in addition to aspirin. Clopidogrel blocks the ADP
40.9% (P ! .001) [31]. More importantly, the rela- pathway by blocking the P2Y12 component of the
tive benet of enoxaparin as compared with unfrac- ADP receptor, which in turn decreases platelet ac-
tionated heparin increased as the risk increased tivation and aggregation. The CURE trial found
[31]. Similar ndings have now been seen using the that clopidogrel plus aspirin led to a 20% relative
TIMI risk score to predict the benet of GP IIb/ risk reduction in cardiovascular death, MI, or
IIIa inhibitors [32] and an early invasive strategy stroke compared with aspirin alone [28]. This ben-
[3]. Thus, these ndings support the ACC/AHA et was seen in low- and high-risk patients [29] and
guideline recommendations that risk stratication was seen as early as 24 hours [36], with the Kaplan-
be the rst task in evaluating patients who present Meier curves for event rate in the two treatment
with UA/NSTEMI [33]. groups diverging after just 2 hours.
ACUTE CORONARY SYNDROMES 405

This benet has also been seen in two other have evidence of recurrent ischemia either at rest
trials. In the CREDO trial of patients undergoing or on provocative testing. Of these, six of the last
PCI, clopidogrel was associated with a signicant seven have all shown a signicant benet of the
27% relative reduction in the combined risk for invasive strategy (Fig. 2), especially in higher-risk
death, MI, or stroke at 1 year (P .02) [37]. Simi- patients [3,35,47]. Accordingly, the 2002 ACC/
larly, in the CAPRIE trial, clopidogrel alone had AHA guideline has added ST-segment changes
a signicant reduction in these events versus aspirin and positive troponin to the list of high-risk indi-
through 3 years of follow-up in patients who had cators that would lead to a Class I recommenda-
prior atherothrombotic disease [38]. Thus, there is tion for an early invasive strategy, as shown in
strong evidence from large double-blind, random- Box 4 [41]. With regard to the timing of an inva-
ized trials that clopidogrel plus aspirin is the new sive strategy, the results from the Intracoronary
optimal long-term antithrombotic regimen. Stenting with Antithrombotic Regimen Cooling-
Intravenous GP IIb/IIIa inhibitors have also O study found a benet of an immediate invasive
been shown to be benecial in treating UA/ strategy with an average time to catheterization
NSTEMI [39]. For upstream management (ie, of 2 hours, compared with a delayed invasive strat-
initiating therapy when the patient rst presents egy with an average time to catheterization of 4
to the hospital), the small molecule inhibitors ep- days [48]. Randomized studies have not yet evalu-
tibatide and tiroban clearly show benet, ated whether an immediate invasive approach is
whereas abciximab was of no benet in an unse- better than catheterization 24 to 48 hours
lected UA/NSTEMI patient population [40] and postadmission.
is in fact contraindicated for patients approached
with a noninvasive strategy [41].
Unfractionated heparin (UFH) or LMWH is Summarydacute therapy for nonST elevation
recommended for patients who have UA/NSTEMI acute coronary syndrome
[42]. Comparative trials of enoxaparin (a LMWH)
versus UFH have demonstrated superiority of Summarizing the ACC/AHA guidelines, there
enoxaparin in reducing recurrent cardiac events are now ve baseline therapies for all patients:
[43,44]. Based on these data, the 2002 Updated aspirin; clopidogrel; heparin or LMWH (with
ACC/AHA UA/NSTEMI practice guidelines enoxaparin the preferred antithrombin); beta-
have made a Class IIA recommendation that enox- blockers; and nitrates (Fig. 3) [23]. Then there
aparin is the preferred antithrombin over UFH [41]. are two treatments that are best targeted by risk
Anti-ischemic therapy with nitrates is also
recommended, with the use of intravenous ni-
trates for ongoing ischemic pain and beta-block- ISAR-COOL
ISAR-COOL
ade early and during long-term follow-up [33].
In STEMI, benets of early angiotensin- RITA-3
RITA-3

VINO
converting enzyme inhibition are considerable. VANQWISH ICTUS
ICTUS VINO
VANQWISH TRUCS
TRUCS
In the GISSI-3 [45] trial, early initiation of lisino- MATE
MATE

pril was associated with a 12% mortality reduc- TACTICS-


TACTICS-
TIMI IIIB
TIMI IIIB TIMI 18
TIMI 18
tion, and in ISIS-4 [46], captopril therapy
resulted in a 7% mortality reduction. Benets FRISC
FRISC II II
were more pronounced in patients who had ante-
rior MI. However, in patients who had NSTEMI,
no benet was observed in the ISIS-4 study. Conservative Invasive
# Pts: 920 2874 7018

Invasive versus conservative strategy: nonST Fig. 2. The weight of the evidence showing benet of
elevation acute coronary syndrome an invasive versus conservative strategy in patients who
have unstable angina/nonST elevation acute coronary
For patients who have UA/NSTEMI, nine
syndrome. The size of the boxes for each of the nine ran-
randomized trials have assessed the merits of an domized trials corresponds to the number of patients en-
invasive strategy involving routine cardiac cathe- rolled. (Modied from Cannon CP, Turpie AG. Unstable
terization with revascularization (if feasible) ver- angina and non-ST-elevation myocardial infarction: ini-
sus a conservative strategy where angiography and tial antithrombotic therapy and early invasive strategy.
revascularization are reserved for patients who Circulation 2003;107:26405; with permission.)
406 CANNON

Aspirin, Clopidogrel, Heparin/LMWH,


Box 4. ACC/AHA guideline Beta-blocker, Nitrates
recommendations for an early invasive High-Risk Lower-Risk
strategy for unstable angina/nonST +Troponin, ST s, TIMI Risk >3 ECG, Markers
elevation myocardial infarction Rec Ischemia, CHF, prior Revasc

Class I GP IIb/IIIa inhibitor Conservative


Any of the high-risk indicators (level of Strategy

evidence: A) Invasive
Strategy
 Recurrent angina at rest/low-level
activity despite treatment Fig. 3. Evidence-based risk stratication to target thera-
 Elevated Troponin T or Troponin I pies in unstable angina/nonST elevation acute coronary
 New ST-segment depression syndrome, as recommended in the American College of
 Recurrent angina/ischemia with Cardiology/American Heart Association guidelines.
congestive heart failure symptoms, (From Cannon CP. Evidence-based risk stratication
rales, mitral regurgitation to target therapies in acute coronary syndromes. Circu-
lation 2002;106:158891; with permission.)
 Positive stress test
 EF less than 0.40
 Decreased blood pressure Many randomized controlled trials have com-
 Sustained tidal volume pared pharmacologic and mechanical reperfusion
 PCI less than 6 months, prior coronary during STEMI. A meta-analysis of 23 of these
artery bypass grafting trials found that primary PCI was superior to
thrombolytic therapy in reducing mortality; non-
From Braunwald E, Antman EM, Beasley
fatal reinfarction; stroke; and the combined end-
JW, et al. ACC/AHA guidelines for the man- point of death, nonfatal reinfarction, and stroke
agement of patients with unstable angina [49]. The benets of PCI in reducing mortality
and non-ST-segment elevation myocardial and recurrent MI were particularly striking.
infarction: a report of the American College At present, however, only 20% to 25% of
of Cardiology/American Heart Association hospitals have primary PCI capabilities. There-
Task Force on Practice Guidelines (Committee fore, hospitals that do not have PCI capabilities
on the Management of Patients With Unstable can either treat patients who have STEMI with
Angina). J Am Coll Cardiol 2000;36:9701062; immediate thrombolysis, or emergently transfer
with permission.
them to a site where PCI can be performed.
Recent clinical trials have also found that if
stratication: the invasive strategy and the IIb/ door-to-balloon times are within 2 to 3 hours,
IIIa inhibitor, where the benet is in the higher- that primary PCI is superior to brinolysis with
risk patients. a 40% to 50% reduction of the combined end
point [5052].
ECG ST-segment elevation: reperfusion therapy
In STEMI, the coronary artery is in most cases Thrombolysis
100% occluded with a fresh thrombus at the site Thrombolysis has been shown to reduce mor-
of a ruptured plaque. To restore perfusion to the tality in several large placebo-controlled trials
myocardium, immediate reperfusion of the in- using streptokinase (SK) and tissue plasminogen
farct-related artery is needed. activator (t-PA). These benets persist through at
There are two approaches: use of brinolytic least 10 years of follow up. The Fibrinolytic
therapy or primary PCI. Time is critical with Therapy Trialists overview of all the major
either therapy, where the sooner reperfusion is placebo-controlled studies showed a 2.6% abso-
achieved, the greater the chance of reducing lute reduction in mortality for patients who have
morbidity and increasing survival. STEMI treated within the rst 12 hours after the
onset of symptoms [53]. Patients presenting with
Primary percutaneous coronary intervention
new left bundle branch block and a strong clinical
At present, the preferred method of achieving history for acute MI also derive a substantial ben-
coronary reperfusion is the use of primary PCI. et from thrombolysis. Patients who have nonST
ACUTE CORONARY SYNDROMES 407

elevation ACS, however, do not benet from and thus rapid identication of ST elevation on
thrombolysis. the ECG is critical. Having a standardized pro-
SK was the initial brinolytic drug, with tocol for rapid treatmentdwith either primary
newer agents later developed. The brin-specic PCI or thrombolysisdis critical. For UA/
thrombolytic agents, such as alteplase (t-PA), NSTEMI, one rst has to identify the patients
were seen in the TIMI-1 trial to have twice the who have a higher likelihood of actually having an
rate of infarct-related artery patency as SK [54]. In ACS as opposed to noncardiac chest pain. There
the GUSTO I trial, accelerated tPA plus heparin are many modalities for this assessment, as
led to a highly signicant 14% reduction in mor- reviewed in this issue. Among patients who have
tality. Reteplase and tenecteplase are molecular a moderate or high likelihood of ACS, stratica-
modications of t-PA designed to have longer tion to high versus lower risk is needed to choose
plasma half-lives that allow double or single bolus appropriate therapies. Thus, it is important for
administration, respectively. However, the newer risk stratication to be a central part of all
agents did not lead to a further reduction in mor- management of patients who have ACS.
tality rates compared with t-PA.
Similarly, use of the combination of half-dose
brinolysis and GP IIb/IIIa receptor inhibitors did References
not improve mortality, and this more complicated
[1] American Heart Association. Heart and stroke sta-
regimen is not widely used. The use of other
tistical update. American Heart Association, in
adjunctive agents with full-dose brinolysis is being press.
actively studied, with enoxaparin being compared [2] The TIMI IIIA Investigators. Early eects of tissue-
with UFH in the EXTRACT-TIMI 25 trial, and type plasminogen activator added to conventional
clopidogrel being added to standard regimens therapy on the culprit lesion in patients presenting
versus placebo in the CLARITY-TIMI 28 trial. with ischemic cardiac pain at rest. Results of the
Thrombolysis in Myocardial Ischemia (TIMI IIIA)
Trial. Circulation 1993;87:3852.
Choice of reperfusion therapy [3] Cannon CP, Weintraub WS, Demopoulos LA, et al.
In summary, although thrombolysis is more Comparison of early invasive and conservative
strategies in patients with unstable coronary syn-
widely available, is simple to administer, and
dromes treated with the glycoprotein IIb/IIIa inhib-
produces results that are independent of operator itor tiroban. N Engl J Med 2001;344(25):187987.
experience, primary PCI in experienced hands is [4] Fox KA, Goodman SG, Klein W, et al. Manage-
associated with less-recurrent infarction and is- ment of acute coronary syndromes. Variations in
chemia; lower short-term mortality; and less practice and outcome; ndings from the Global Reg-
stroke and intracranial hemorrhage than throm- istry of Acute Coronary Events (GRACE). Eur
bolysis. PCI also results in higher early infarct- Heart J 2002;23(15):117789.
related artery patency rates and reduced residual [5] Pope JH, Ruthazer R, Beshansky JR, et al. Clinical
stenosis compared with thrombolysis. The current features of emergency department patients present-
ACC/AHA guidelines recommend invasive strat- ing with symptoms suggestive of acute cardia ische-
mia: a multicenter study. J Thromb Thrombolysis
egy with PCI if a timely intervention can be
1998;6:6374.
performed; that is, if door-to-balloon time is less [6] Kontos MC, Ornato JP, Tatum JL, et al. How many
than 90 minutes or if the delay to PCI compared patients are eligible for treatment with GP IIb/IIIa
with on-site thrombolysis is less than 1 hour. inhibitors? Results from a clinical database. Circula-
Primary PCI is also generally preferred for pa- tion 1999;100(Suppl 1):1775.
tients in cardiogenic shock, who have contra- [7] Pozen MW, DAgostino RB, Selker HP, et al. A pre-
indications to thrombolysis, and in whom dictive instrument to improve coronary-care-unit
thrombolysis is unlikely to produce meaningful admission practices in acute ischemic heart disease.
benet (eg, symptom onset more than 3 hours A prospective multicenter clinical trial. N Engl J
before presentation). Med 1984;310:12738.
[8] Tierney WM, Roth BJ, Psaty B, et al. Predictors of
myocardial infarction in emergency room patients.
Summary Crit Care Med 1985;13:52631.
[9] Goldman L, Cook EF, Brand DA, et al. A computer
For patients who have ACS, risk stratication protocol to predict myocardial infarction in emer-
is key to initiating appropriate treatment. For gency department patients with chest pain. N Engl
STEMI, immediate reperfusion therapy is needed, J Med 1988;318:797803.
408 CANNON

[10] Lee TH, Juarez G, Cook EF, et al. Ruling out acute to Open Occluded Coronary Arteries in Acute Cor-
myocardial infarction. A prospective multicenter onary Syndromes IIb Investigators. N Engl J Med
validation of a 12-hour strategy for patients at low 1999;341(4):22632.
risk. N Engl J Med 1991;324:123946. [22] Scirica BM, Cannon CP, Gibson CM, et al. Assess-
[11] Selker HP, Grith JL, DAgostino RB. A time- ing the eect of publication of clinical guidelines on
insensitive predictive instrument for acute myocar- the management of unstable angina and non-ST-
dial infarction mortality: a multicenter study. Med elevation myocardial infarction in the TIMI III
Care 1991;29(12):1196211. (199093) and the GUARANTEE (199596) regis-
[12] Sarasin FP, Reymond JM, Grith JL, et al. Impact tries. Critical Pathways in Cardiology 2002;1:
of the acute cardiac ischemia time-insensitive predic- 15160.
tive instrument (ACI-TIPI) on th speed of triage de- [23] Cannon CP. Evidence-based risk stratication to
cision making for emergency department patients target therapies in acute coronary syndromes. Circu-
presenting with chest pain: a controlled clinical trial. lation 2002;106(13):158891.
J Gen Intern Med 1994;9:18794. [24] Hamm CW, Heeschen C, Goldmann B, et al. Benet
[13] Selker HP, Beshansky JR, Grith JL, et al. Use of of abciximab in patients with refractory unstable an-
the acute cardiac ischemia time-insensitive predictive gina in relation to serum troponin T levels. N Engl J
instrument (ACI-TIPI) to assist with triage of Med 1999;340(21):16239.
patients with chest pain or other symptoms sugges- [25] Heeschen C, Hamm CW, Goldmann B, et al. Tropo-
tive of acute cardiac ischemia. A multicenter, con- nin concentrations for stratication of patients with
trolled clinical trial. Ann Intern Med 1998;129(11): acute coronary syndromes in relation to therapeu-
84555. tic ecacy of tiroban. Lancet 1999;354(9192):
[14] Goldman L, Cook EF, Johnson PA, et al. Prediction 175762.
of the need for intensive care in patients who come [26] Morrow DA, Cannon CP, Rifai N, et al. Ability of
to emergency departments with acute chest pain. minor elevations of troponin I and T to predict ben-
N Engl J Med 1996;334(23):1498504. et from an early invasive strategy in patients with
[15] Fox KA, Cokkinos DV, Deckers J, et al. The unstable angina and non-ST elevation myocardial
ENACT study: a pan-European survey of acute infarction: results from a randomized trial. JAMA
coronary syndromes. European Network for 2001;286:240512.
Acute Coronary Treatment. Eur Heart J 2000;21: [27] Antithrombotic Trialists Collaboration. Collabora-
14409. tive meta-analysis of randomised trials of antiplate-
[16] The Platelet Receptor Inhibition for Ischemic Syn- let therapy for prevention of death, myocardial
drome Management in Patients Limited by Unstable infarction, and stroke in high risk patients. BMJ
Signs and Symptoms (PRISM-PLUS) Study Investi- 2002;324(7329):7186.
gators. Inhibition of the platelet glycoprotein IIb/ [28] Clopidogrel in Unstable Angina to Prevent Recur-
IIIa receptor with tiroban in unstable angina and rent Events Trial Investigators. Eects of clopidog-
non-Q-wave myocardial infarction. N Engl J Med rel in addition to aspirin in patients with acute
1998;338:148897. coronary syndromes without ST-segment elevation.
[17] The PURSUIT Trial Investigators. Inhibition of N Engl J Med 2001;345:494502.
platelet glycoprotein IIb/IIIa with eptibatide in [29] Budaj A, Yusuf S, Mehta SR, et al. Benet of clopi-
patients with acute coronary syndromes. N Engl J dogrel in patients with acute coronary syndromes
Med 1998;339:43643. without ST-segment elevation in various risk
[18] Antman EM, McCabe CH, Gurnkel EP, et al. groups. Circulation 2002;106:16226.
Enoxaparin prevents death and cardiac ischemic [30] Boersma E, Pieper KS, Steyerberg EW, et al. Predic-
events in unstable Angina/Non-Q-wave myocardial tors of outcome in patients with acute coronary syn-
infarction: results of the Thrombolysis In Myocar- dromes without persistent ST-segment elevation.
dial Infarction (TIMI) 11B trial. Circulation 1999; Results from an international trial of 9461 patients.
100(15):1593601. Circulation 2000;101:255767.
[19] The Global Use of Strategies to Open Occluded [31] Antman EM, Cohen M, Bernink PJ, et al. The TIMI
Coronary Arteries (GUSTO) IIb Investigators. risk score for unstable angina/non-ST elevation MI:
A comparison of recombinant hirudin with heparin A method for prognostication and therapeutic deci-
for the treatment of acute coronary syndromes. sion making. JAMA 2000;284:83542.
N Engl J Med 1996;335:77582. [32] Morrow DA, Antman EM, Snapinn SM, et al. An
[20] Hochman JS, McCabe CH, Stone PH, et al. Out- integrated clinical approach to predicting the benet
come and prole of women and men presenting of tiroban in non-ST elevation acute coronary syn-
with acute coronary syndromes: a report from dromes: application of the TIMI Risk Score for US/
TIMI IIIB. J Am Coll Cardiol 1997;30:1418. NSTEMI in PRISM-PLUS. Eur Heart J 2002;23(3):
[21] Hochman JS, Tamis JE, Thompson TD, et al. Sex, 2239.
clinical presentation, and outcome in patients with [33] Braunwald E, Antman EM, Beasley JW, et al. ACC/
acute coronary syndromes. Global Use of Strategies AHA guidelines for the management of patients
ACUTE CORONARY SYNDROMES 409

with unstable angina and non-ST segment elevation high-risk patients with non-ST-segment elevation
myocardial infarction: a report of the American acute coronary syndromes receiving the glycopro-
College of Cardiology/American Heart Association tein IIb/IIIa inhibitor eptibatide. Circulation
Task Force on Practice Guidelines (Committee on 2003;107:23844.
the Management of Unstable Angina and Non-ST [45] Gruppo Italiano per lo Studio della Sopravvivenza nel-
Segment Elevation Myocardial Infarction). J Am linfarto Miocardico. GISSI-3: eects of lisinopril and
Coll Cardiol 2000;36:9701056. transdermal glyceryl trinitrate singly and together on
[34] Theroux P, Ouimet H, McCans J, et al. Aspirin, hep- 6-week mortality and ventricular function after acute
arin or both to treat unstable angina. N Engl J Med myocardial infarction. Lancet 1994;343:111522.
1988;319:110511. [46] ISIS-4 (Fourth International Study of Infarct Sur-
[35] Peters R, Zhao F, Lewis BS, et al. Aspirin dose and vival) Collaborative Group. ISIS-4: a randomized
bleeding events in the CURE study. Eur Heart J factorial trial assessing early oral captopril, oral
2002;4(Suppl):510. mononitrate, and intravenous magnesium sulphate
[36] Yusuf S, Mehta SR, Zhao F, et al. Early and late in 58,050 patients with suspected acute myocardial
eects of clopidogrel in patients with acute coronary infarction. Lancet 1995;345:66985.
syndromes. Circulation 2003;107(7):96672. [47] FRagmin and Fast Revascularisation during InSta-
[37] Steinhubl SR, Berger PB, Mann JT III, et al. Early bility in Coronary artery disease Investigators. Inva-
and sustained dual oral antiplatelet therapy follow- sive compared with non-invasive treatment in
ing percutaneous coronary intervention: a random- unstable coronary-artery disease: FRISC II prospec-
ized controlled trial. JAMA 2002;288(19):241120. tive randomised multicentre study. Lancet 1999;
[38] CAPRIE Steering Committee. A randomised, 354(9180):70815.
blinded, trial of clopidogrel versus aspirin in patients [48] Neumann FJ, Kastrati A, Pogatsa-Murray G, et al.
at risk of ischaemic events (CAPRIE). Lancet 1996; Evaluation of prolonged antithrombotic pretreat-
348:132939. ment (cooling-o strategy) before intervention in
[39] Boersma E, Harrington RA, Moliterno DJ, et al. patients with unstable coronary syndromes: a ran-
Platelet glycoprotein IIb/IIIa inhibitors in acute cor- domized controlled trial. JAMA 2003;290:15939.
onary syndromes: a meta-analysis of all major rand- [49] Keeley EC, Boura JA, Grines CL. Primary angio-
omised clinical trials. Lancet 2002;359:18998. plasty versus intravenous thrombolytic therapy for
[40] The GUSTO IV-ACS Investigators. Eect of glyco- acute myocardial infarction: a quantitative review of
protein IIb/IIIa receptor blocker abciximab on 23 randomised trials. Lancet 2003;361(9351):1320.
outcome in patients with acute coronary syn- [50] Andersen HR, Nielsen TT, Rasmussen K, et al. A
dromes without early coronary revascularisation: comparison of coronary angioplasty with brinolytic
the GUSTO IV-ACS randomised trial. Lancet therapy in acute myocardial infarction. N Engl J
2001;357(9272):191524. Med 2003;349(8):73342.
[41] Braunwald E, Antman EM, Beasley JW, et al. ACC/ [51] Widimsky P, Budesinsky T, Vorac D, et al. Long
AHA guideline update for the management of distance transport for primary angioplasty vs im-
patients with unstable angina and non-ST-segment mediate thrombolysis in acute myocardial infarc-
elevation myocardial infarction-2002: summary arti- tion. Final results of the randomized national
cle: a report of the American College of Cardiology/ multicentre trialPRAGUE-2. Eur Heart J 2003;
American Heart Association Task Force on Practice 24(1):94104.
Guidelines (Committee on the Management of [52] Grines CL, Westerhausen DR Jr, Grines LL, et al. A
Patients with Unstable Angina). Circulation 2002; randomized trial of transfer for primary angioplasty
106(14):1893900. versus on-site thrombolysis in patients with high-risk
[42] Oler A, Whooley MA, Oler J, et al. Adding heparin myocardial infarction: the Air Primary Angioplasty
to aspirin reduces the incidence of myocardial infarc- in Myocardial Infarction study. J Am Coll Cardiol
tion and death in patients with unstable angina. A 2002;39(11):17139.
meta-analysis. JAMA 1996;276:8115. [53] Fibrinolytic Therapy Trialists (FTT) Collaborative
[43] Antman EM, Cohen M, Radley D, et al. Assessment Group. Indications for brinolytic therapy in sus-
of the treatment eect of enoxaparin for unstable pected acute myocardial infarction: collaborative
angina/non-Q-wave myocardial infarction: TIMI overview of early mortality and major morbidity
11B-ESSENCE meta-analysis. Circulation 1999; results from all randomised trials of more than
100:16028. 1000 patients. Lancet 1994;343(8893):31122.
[44] Goodman SG, Fitchett D, Armstrong PW, et al. [54] TIMI Study Group. The Thrombolysis in Myocar-
Randomized evaluation of the safety and ecacy dial Infarction (TIMI) Trial; Phase I ndings.
of enoxaparin versus unfractionated heparin in N Engl J Med 1985;312:9326.
Cardiol Clin 23 (2005) 411421

Chest Pain Unit Concept: Rationale


and Diagnostic Strategies
Andra L. Blomkalns, MD*, W. Brian Gibler, MD
Department of Emergency Medicine, University of Cincinnati College of Medicine,
Mail Location 0769, 231 Albert Sabin Way, Cincinnati, OH 45267-0769, USA

Each year in the United States, over 8 million present with possible ACS. The evaluation of
patients present to the emergency department patients who exhibit symptoms consistent with
(ED) with complaints of chest discomfort or other ACS requires a protocol that includes testing for
symptoms consistent with possible acute coronary myocardial necrosis, rest ischemia, and exercise-
syndrome (ACS). While over half of these patients induced ischemia.
are typically admitted for further diagnostic
evaluation, fewer than 20% are diagnosed with Implementation
ACS. With hospital beds and inpatient resources
scarce, these admissions can be avoided by eval- Implementation of a CPU approach in an
uating low- to moderate-risk patients in chest pain emergency setting is nearly as challenging as
units (CPUs) [15]. evaluating the patients themselves. Several syner-
This large, undierentiated patient population gistic components are required to adequately,
represents a potential high-risk group for emer- eciently, and cost-eectively evaluate this pa-
gency physicians requiring a systematic approach tient population, including
and specic ED resources. This evaluation is
1) Sucient and dedicated space to evaluate
required to appropriately determine if a patient
and observe the patients
is safe to be discharged home with outpatient
2) Nursing sta and other ED personnel that
follow-up versus requiring admission to the hos-
are knowledgeable and dedicated to this
pital for monitoring and further testing.
approach
The concept of rapid diagnostic and treatment
3) Emergency physician sta that can be re-
protocols for specic disease processes is not new.
sponsible for these patients
For instance, analogous approaches for trauma
4) Collaboration of cardiologist colleagues on
patients have been accepted widely for many
the protocols, testing, follow-up, and admis-
years, resulting in decreased morbidity and im-
sion procedures
proved survival for these patients. The same
5) Availability of nuclear cardiology and radiol-
tenants of early evaluation, appropriate risk
ogy colleagues for the emergent performance
stratication, and intervention are appropriate
and reading of nuclear imaging studies
for the patient presenting with possible ACS to
6) Laboratory personnel to run serial cardiac
the ED. Just as trauma centers have adopted strict
biomarker measurements
protocols for diagnosis and treatment of the
7) Primary care physicians that understand the
injured patient, CPUs should have basic principles
scope and limitations of a CPU evaluation
for evaluation and treatment of patients who
for follow-up
Various physical models exist for CPUs. Sev-
* Corresponding author. eral of these models are illustrated in Table 1 and
E-mail address: Andra.Blomkalns@uc.edu Figs. 1 and 2. In Table 1, the Medical College of
(A.L. Blomkalns). Virginia protocol integrates the use of radionuclide
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.011 cardiology.theclinics.com
412 BLOMKALNS & GIBLER

Table 1
Medical College of Virginia CPU protocol
Level AMI risk ACS risk Strategy Disposition
1 Very high Very high Fibrinolysis/PCI CCU
2 High High ASA, heparin, NTG, GP IIb/IIIa CCU
3 Moderate Moderate Markers nuclear imaging 9-h observations
4 Low Mod or low Nuclear imaging Home and OP stress test
5 Very low Very low As needed Home
Abbreviation: AMI, acute myocardial infarction.

testing into a ve-level risk stratication program. be prohibitive for some hospital physicians,
Figs. 1 and 2 demonstrate protocols for the nurses, and administrations considering the
University of Cincinnati and the Mayo Clinic, re- start-up of a new CPU and, thus, some of these
spectively. Both protocols integrate myocardial protocols are performed within the regular de-
necrosis testing by cardiac biomarkers and func- partment using basic ED resources. This alterna-
tional exercise testing. Each model takes into ac- tive can be suboptimal as the increased time
count the specic patient population, cardiologist required for CPU protocols can hamper ED pa-
collaboration, nuclear imaging availability, and tient throughput, lowering patient satisfaction.
ED resources to design the optimal CPU for that A successful CPU also requires a nursing sta
hospital. dedicated to the care of these patients with special
Most of these units are located in an area needs. These patients require frequent assessment,
adjacent to, but separate from, the general emer- close monitoring, noninvasive testing, and serial
gency patient population. Often the CPU protocol cardiac biomarker testing and ECG acquisition.
patients are evaluated in conjunction with other These tasks are generally in excess of those
observation unit (OU) protocols. As these pa- expected for a typical emergency patient and
tients are frequently in the ED for 6 hours or may require additional training and stang as
more, it is desirable to provide many of the in- appropriate for the individual institution and ED.
hospital comforts generally not available in a tra- The overall success of the unit depends on
ditional ED. These comforts may include hospital adherence to the systematic evaluation protocol
beds, meals, television sets, and telephones, which by nursing personnel.
greatly improve patient satisfaction. This space Although most of these patients are placed on
allocation and additional patient resources may CPU protocols that do not require constant

Fig. 1. University of Cincinnati Heart ER strategy. Data from [5] Gibler et al, Ann Emerg Med 1995;25:18 and [17]
Storrow et al, Circulation 1998;98:I-425.
CHEST PAIN UNIT CONCEPT 413

Fig. 2. Mayo Clinic Strategy.

emergency physician interaction, they must be Radionuclide perfusion imaging has emerged as
continually re-evaluated and re-assessed while in an important tool in many centers for assessing
the CPU. ACS is a dynamic process and can easily patients who present with and without known
progress over the course of the patients stay in cardiovascular disease and also in evaluating the
the CPU. Symptom changes, cardiac biomarker patient presenting to the ED with possible ACS.
positivity, and ECG ST-segment variation should Technetium-99 (Tc-99) sestamibi tomographic
prompt immediate therapy and potential interven- imaging is the most common agent currently
tion in the cardiac catheterization laboratory. The being used in patients presenting to the ED with
emergency physician must stay integrally involved suspected ACS [11,12]. Studies have shown that
with the patient to provide this increased level of positive rest perfusion imaging accurately identi-
therapy. A successful CPU requires the coopera- es patients at high risk for adverse cardiac events
tion of many health care professionals in the ED [13]. Radionuclide perfusion imaging has thus be-
and around the participating institution. An abso- come a cornerstone of evaluation for many units
lutely critical component of successful CPU oper- [14,15].
ation is the collaboration and cooperation of The diagnostic sensitivity of myocardial perfu-
cardiologists and radiologists. Cardiology in- sion imaging is, in part, dependent on the timing
volvement in CPU protocols is necessary for of injection in relation to the cessation of chest
many reasons. Formulation, knowledge, and ac- pain. The major obstacles for radionuclide studies
ceptance of CPU protocols particular for each in- have traditionally included cost and accessibility.
stitution requires active participation and The perfusion agents have a 6- to 12-hour shelf
expertise of the cardiologists. In addition, many life and, thus, have to be prepared several times
CPU protocols successfully use provocative test- a day to be available for acute imaging. Collab-
ing with radionuclide echocardiography, or stan- oration with nuclear cardiology and radiology
dard graded exercise testing, and these tests colleagues is paramount for this reason. Timely
generally require the interpretation of a cardiolo- evaluation of these patients requires immediate
gist [510]. access to these studies and timely reading and
While collaboration between emergency med- reporting of the results by expert readers.
icine and cardiology is intuitive, integration with The laboratory must also be supportive of
nuclear cardiologists and radiologists is also a CPU. Ideally, bedside point-of-care testing for
pivotally important to the success of every CPU. cardiac biomarkers provides rapid data collection
414 BLOMKALNS & GIBLER

for these patients as creatine-kinase myocardial low- to moderate-risk patients who present with
band (CK-MB) and, in particular, the troponins I possible ACS to the ED.
(TnI) and T (TnT) allow risk stratication to be Other institutions also have developed eective
performed expeditiously [16]. A collaborative rela- chest pain center strategies. The Medical College
tionship between emergency physicians and labo- of Virginia has an elegant protocol that triages
ratorians can ensure a consistent and eective chest pain patients into ve distinct levels (see
evaluation of patients in the CPU. Table 1) [14,18]. Level 1 patients have ECG crite-
Finally, primary care physicians must be sup- ria for acute myocardial infarction (AMI) while
portive of the CPU and willing to have patients level 5 patients have clearly noncardiac chest dis-
evaluated in this protocol-driven process. Com- comfort. The triage level dictates further diagnos-
munication between emergency physicians and the tic measures and treatment. Intermediate level
patients primary care physician can then assure patients (levels 2 and 3) include individuals with
careful follow-up and patient compliance with variable probability of unstable angina. These pa-
lifestyle changes and treatment recommendations tients are admitted to the CCU for diagnostic test-
after evaluation. ing for ACS while less acute patients (level 4)
undergo serial cardiac biomarker determination
Examples of chest pain units
and Tc-99 sestamibi radionuclide imaging from
Over the years, several manifestations of CPUs the ED.
have evolved. Each of these models has been The University of California at Davis protocol
successfully adopted by specic institutions for its uses a novel use of immediate exercise treadmill
patient populations. An individual institution testing without serial cardiac biomarker determi-
should carefully evaluate their patient population, nation. In a recent study of 1000 patients, 13%
physician expertise, physical structure, stang were positive and 64% were negative for ischemia.
model, and hospital environment to most ade- The remaining 23% of the patients had non-
quately determine their optimal CPU protocol. diagnostic tests. No adverse eects exist from
For instance, the University of Cincinnati exercise testing and no mortality occurred in
Center for Emergency Care Heart ER, estab- either of the patient evaluation groups at 30-day
lished in October 1991, was one of the rst ED- follow-up. The authors concluded that immediate
based CPUs. Patients at low to moderate risk for exercise testing of low-risk patients is safe and
ACS are admitted to the Heart ER for a protocol- accurate for the determination of which patients
driven evaluation of their chest discomfort. The can be safely discharged from the ED [6]. A care-
serial cardiac biomarkers CK-MB and troponin T ful history is required to ensure that only low- to
levels are drawn at 0, 3, and 6 hours after moderate-risk patients for ACS are exercised
presentation to the ED. The original protocol acutely in the ED.
used continuous ST-segment trend monitoring, The Mayo Clinic separates patients into low-,
which has since been discontinued. Graded exer- intermediate-, and high-risk categories according
cise testing or Tc-99 sestamibi radionuclide scan- to Agency for Health Care Policy Research
ning is now performed depending on patients (AHCPR) guidelines. Intermediate-risk patients
functional status and test availability. Patients are evaluated with CK-MB levels at 0, 2, and 4
having negative evaluations in the Heart ER are hours while undergoing continuous ST-segment
released to home with careful follow-up as an monitoring and 6-hour observation. If this eval-
outpatient [5]. uation is negative, an ECG exercise test, a nuclear
In the rst 2131 consecutive patients evaluated stress test, or an echocardiographic stress test is
over a 6-year period, 309 (14.5%) required ad- performed. Patients with positive or equivocal
mission and 1822 (85.5%) were released to home evaluations are admitted, whereas patients having
from the ED. Of admitted patients, 94 (30%) were negative evaluations are discharged to home with
found to have a cardiac cause for their chest pain. a 72-hour follow-up [19].
Follow-up of 1696 patients discharged from the Lastly, Brigham and Womens Hospital di-
Heart ER to home yielded 9 cardiac events vides patients into three groups: unstable angina
(0.53%, condence interval [CI], 0.24%1.01%; or AMI, possible ischemia, and nonischemic.
7 percutaneous transluminal coronary angioplasty, Patients with unstable angina or AMI are admit-
1 coronary artery bypass grafting, 1 death) [17]. ted, whereas nonischemic chest pain patients
These data suggest that the Heart ER program are discharged from the CPU. The intermediate,
provides a safe and eective means for evaluating or possible ischemia, group either undergoes
CHEST PAIN UNIT CONCEPT 415

exercise treadmill testing with a 6-hour period of 7%. Patients who have CK-MB levels lower
observation or a 12-hour period of observation. than 5% of the total CK without recurrent chest
At the end of the observation period, stable pain after 12 hours had a 0.5% missed AMI rate
patients are discharged to home. Nichol and while identifying 94% of AMI patients [22]. Far-
colleagues [20] evaluated the impact of this path- kouh and colleagues [19] demonstrated the use of
way approach in a retrospective cohort of 4585 a CPU protocol and CK-MB measurements for
patients and found that a 17% reduction in ad- patients identied as intermediate risk for ad-
missions and an 11% reduction in length of stay verse cardiac events. In this study, patients un-
would occur if even fewer than 50% of eligible pa- derwent 6 hours of observation followed by
tients for observation and exercise testing had provocative testing. This protocol identied all
participated. patients who experienced short and long term
Each institution has constructed its CPU pro- cardiac events while using fewer resources over
tocols to suit its patient population, physician a 6-month period [19].
expertise, and resource availability to optimize Symptom onset to patient presentation is
chest pain evaluation. Each hospital and patient a crucial factor in the use of cardiac biomarker
population represents a unique environment with protocols. Marker release kinetics vary with time
specic needs and resources that must be reected and as time to ED presentation may be as short as
in the ultimate CPU design and implementation. 90 minutes or as long as several days, no single bio-
Once the patient passes the CPU protocol marker determination is suitable for adequately
and apparently does not have ACS, the individual ruling-out myocardial necrosis [5,23,24]. In one
can be discharged safely to home and appropriate of the rst studies with CPU protocols, Gibler and
follow-up arranged. Adequate attention must be colleagues [5] used a 9-hour protocol with serial
given to the cause of the patients discomfort even CK-MB levels performed at 0, 3, 6, and 9 hours
if it is determined not to be cardiac in nature. along with continuous 12-lead ECG monitoring,
Testing for other noncardiac chest pain, including echocardiography, and graded exercise testing.
gastrointestinal or psychiatric disease, must then The authors found that serial cardiac biomarkers
be performed. Patients also require outpatient alone had a sensitivity and specicity for AMI of
provocative testing, if not received in the CPU 100% and 98%, respectively [5,23,24]. Serial car-
protocol, to further delineate their cardiac risk diac biomarker determinations are now recom-
caused by xed coronary artery lesions. These mended to increase sensitivity for detecting
tests must be followed and acted on by a primary necrosis rather than a single determination on
care physician. ED presentation.
Several other studies have examined the value
of cardiac biomarkers in the risk stratication of
heterogenous patients who present to the ED with
Ecacy
chest pain. In a multicenter study of over 5000
The protocols developed for CPUs must pro- patients in 53 EDs, the relative risk for ischemic
vide testing to evaluate every patient who presents complications and death for ED patients who
with potential ACS for three possibilities: myo- have positive CK-MB at 0 or 2 hours was 16.1 and
cardial necrosis, rest ischemia, and exercise- 25.4, respectively [25]. Serial CK-MB results also
induced ischemia. Numerous studies in several have proved to be sensitive in AMI detection
dierent hospital environments have validated the when collected at 0 and 3 hours after ED presen-
use of the CPU. Even early CPUs without the tation. Young and colleagues [26] found a 93%
benets of cardiac troponins and immediate nu- and 95% sensitivity and specicity when combin-
clear imaging were successful in safely evaluating ing 0, 3, and net change in CK-MB levels. As ex-
patients who are at low probability for AMI [21]. pected, the sensitivity improved with increased
Early myocardial necrosis rule-out proto- time from symptom onset [26]. Serial cardiac
cols challenged the traditional notion of a 24- biomarker measurements and comparison of
hour period required to detect AMI. Lee and biomarker elevation over 3 to 6 hours also im-
colleagues [22] multicenter trial validated a 12- proved sensitivity for MI [2729]. Even minor ele-
hour algorithm using CK and CK-MB in pa- vations of CK-MB as small as twice the upper
tients identied as low-risk through assessment limit of normal are associated with an increased
of clinical characteristics in the ED. Low-risk 6-month mortality when compared with those
was dened as the probability of AMI less than with normal levels [30].
416 BLOMKALNS & GIBLER

Serial CK/CK-MB protocols have become the period, was 97% [32]. As CPUs continue to
diagnostic standard for AMI in the CPU setting. evolve, aggressive and innovative biomarker strat-
Most all of studied protocols use a specic thresh- egies have been developed. McCord and col-
old above which is diagnostic for AMI. Fesmire leagues [35] found that AMI can be excluded 90
and colleagues [29] studied a promising new ap- minutes after patient presentation using point-
proach that included the change in CK-MB levels of-care myoglobin and cardiac troponin I.
within the normal range over the course of ED Data from protocols using myoglobin meas-
evaluation. In his population of 710 CPU pa- urements in patients who have a lower risk for
tients, a CK-MB increase or delta of 1.6 ng/mL AMI are sometimes conicting. In a study of 3075
over 2 hours was more sensitive for AMI than low-risk CPU patients with AMI prevalence of
a second CK-MB drawn 2 hours after patient ar- 1.4%, a 4-hour serial myoglobin protocol was
rival (93.8% versus 75.2%) [29]. Validation of reported as 100% sensitive for AMI [9]. Con-
these and other novel protocols will add further versely, in a study of 368 patients whose AMI
to the use of biomarkers in the CPU setting. prevalence was 11%, the sensitivity and specicity
The benecial attributes of myoglobin have of myoglobin at 0 and 2 to 3 hours was only 61%
made it a commonly used biomarker in CPU and 68%. Myoglobin change or increase did not
protocols. The diagnostic strength of myoglobin improve diagnostic performance either [28].
lies in its early release kinetics and sensitivity, Myoglobin in the CPU setting is probably best
whereas its primary weakness is a lack of speci- used in a serial fashion along with another bio-
city. Davis and colleagues [31] showed that serial marker of necrosis such as troponin which is
myoglobin levels were 93% sensitive and 79% highly specic for cardiac injury. It is most valu-
specic in detecting AMI in patients within 2 able when used in patients presenting early in the
hours of arrival. Similarly, Tucker and colleagues time course of symptoms and less so for remote
[32] showed a myoglobin sensitivity of 89% in pa- events, as the myoglobin elevation can be nor-
tients who had nondiagnostic ECGs within 2 malized within 12 hours after symptom onset
hours of ED presentation. Myoglobin seems to through renal clearance.
achieve maximal diagnostic accuracy within 5 TnI and TnT have revolutionized the risk
hours after symptom onset [33]. stratication of chest pain patients in CPU pro-
Therefore, it is reasonable and recommended tocols and now are a cornerstone of serial
that myoglobin should be combined with other evaluation. They have proved extremely valuable
more specic cardiac biomarkers when used in and sensitive in the diagnosis of myocardial
CPU protocols. Brogan and colleagues [34] found necrosis [36,37]. In addition to diagnosis of myo-
that a combination of carbonic anhydrase III and cardial necrosis in ACS, the troponins are valu-
serum myoglobin was more sensitive and equally able in risk stratication of low- and high-risk
specic as CK-MB in patients presenting early, patient populations. Troponin release kinetics
within 3 hours of symptom onset. In contrast, mimic CK-MB, but elevated levels persist for up
Kontos and colleagues [3] reported less encourag- to 12 days after AMI as a result of the breakdown
ing results from a study of 2093 patients combin- of the contractile apparatus over this period.
ing CK, CK-MB, and myoglobin obtained at 0, 3, The main two issues surrounding the cardiac
6, and 8 hours. A CK-MB level greater than 8.0 troponins are: (1) cuto values for TnI and TnT
ng/mL at three hours was 93% sensitive and and (2) appropriately dening the time of chest
98% specic for AMI; adding myoglobin de- pain onset in the context of the ED presentation.
creased the specicity to 86% with no signicant Most large studies and analyses have determined
increase in sensitivity. that TnI and TnT can identify patients at risk for
Much like the other cardiac biomarkers, myo- adverse cardiac events [3639].
globin levels become increasingly useful when Cardiac TnT is detected at slightly lower
drawn in a serial fashion. In a study of 133 serum levels than most TnI assays and has proved
consecutive admitted chest pain patients, myoglo- valuable in the emergency setting for early iden-
bin levels were obtained at 2, 3, 4, and 6 hours tication of myocardial necrosis. The GUSTO-II
after symptom onset. This regimen was found to Investigators compared TnI and TnT in short-
be 86% sensitive for AMI at 6 hours. The negative term risk stratication of ACS patients. This
predictive value (NPV) in patients who have model compared troponins collected within
negative myoglobin levels during 6 hours of eval- 3.5 hours of ischemic symptoms. Ohman and
uation, and without doubling over any 2-hour colleagues found that TnT showed a greater
CHEST PAIN UNIT CONCEPT 417

association with 30-day mortality (Chi2 18.0, CK-MB and myoglobin maintained better sensi-
P ! .0001) than TnI (Chi2 12.5, P .0002) tivity. The troponins were useful only when mea-
[40]. These authors concluded that TnT is sured 6 or more hours after arrival, exhibiting
a strong, independent predictor of short-term out- sensitivities and specicities of 82% and 97%
come in ACS patients, and serial levels were use- for TnI, and 89% and 84% for TnT [47]. Tropo-
ful in determining the risk for adverse cardiac nin use seems to be more benecial in later or de-
events [41]. layed patient presentations. In a study of 425
As with all new cardiac biomarkers, initial patients using serial TnI and CK-MB over 16
studies on troponin risk stratication were per- hours, Brogan and colleagues [48] showed no in-
formed initially on patients who have known crease in sensitivity or specicity between tropo-
ACS. Studies using TnI in ACS patients showed nin and CK-MB in patients with symptoms less
a statistically signicant increase in mortality than 24 hours. However, in patients presenting
among those patients who have levels greater with greater than 24 hours of symptoms, TnI
than 0.4 ng/mL [42]. TnT was an extremely eec- had a sensitivity of 100% compared with CK-
tive test for risk stratication of patients who have MB (56.5%) [48].
ACS in a large trial [43]. Stubbs and colleagues Sayre and colleagues showed that patients who
[44] showed that patients who have elevated base- exhibited a TnT level of 0.2ng/L or greater were
line TnT levels had up to 4 times higher mortality 3.5 times more likely to have a cardiac complica-
than ACS patients with normal values. tion within 60 days of ED presentation [49]. In
While the increased risk for troponin positive a CPU population, Newby and colleagues [50] de-
patients is well-established, the degree of risk termined that TnT positive patients had angio-
varies greatly between studies and patient popu- graphically signicant lesions (89% versus 49%)
lations. Meta-analyses have helped consolidate and positive stress testing (46% versus 14%)
conclusions and clinically useful parameters more frequently than TnT negative patients.
when using troponins for the evaluation of Long-term mortality was also higher in TnT pos-
patients. One such analysis in high-risk patients itive patients (27% versus 7%) [50]. Johnson and
performed by Wu [45] demonstrated a cumulative colleagues [37] studied a heterogeneous patient
odds ratio of a positive TnT for the development population admitted from an urban teaching hos-
of AMI or death from hospital discharge to 34 pital and found that TnT was elevated in 31% of
months was 4.3 (2.86.8 95% CI). The cumulative patients who presented without AMI and who
odds ratio of a positive TnT for predicting need had major short-term complications as compared
for cardiac revascularization within the same peri- with CK-MB activity and mass [37].
od was 4.4 (3.06.5 95% CI). Another analysis in- Other investigators have found that while
volving greater than 18,000 patients in 21 ACS patients with troponin positivity are at higher
studies found that troponin positive patients risk for adverse cardiac events, the test in isolation
had an odds ratio of 3.44 for death or MI at lacks sensitivity. Kontos and colleagues [51] found
30 days. Troponin positive patients with no ST- that while TnI positive patients were more likely
segment elevation and patients with unstable an- to have signicant complications (43% vs 12%),
gina carried odds ratios of 4.93 and 9.39 for the sensitivity for these end points was low
adverse cardiac outcomes [39]. (14%). Similarly, Polanczyk and colleagues
Benamer and colleagues [46] compared the [52] demonstrated that peak TnI greater than
prognostic value of TnI combined with C-reactive 0.4 ng/mL was associated with only a 47% sensi-
protein (CRP) in patients who present with unsta- tivity and 80% specicity for a major cardiac
ble angina. They found that while 23% of patients event within 72 hours of presentation.
with elevated TnI had major in-hospital cardiac The recent publication of the redenition of
events, there was no such prognostic signicance MI has brought troponin to the forefront of
associated with CRP [46]. diagnosis and risk stratication in this patient
Troponin applications in low- to moderate- population [53]. In a re-analysis of the data from
risk patients presenting to EDs have shown PARAGON B, GUSTO IIa, and CHECKMATE
similar encouraging results. Tucker and col- studies, patients with baseline troponin elevation
leagues [47] used a comprehensive marker strategy without CK-MB elevation were found to be at in-
including myoglobin, CK-MB, TnI, and TnT in creased risk for early and short term adverse out-
ED patients over 24 hours after arrival. As ex- comes [43,5456]. McCord and colleagues [57]
pected within the rst 2 hours of presentation, found that troponin and myoglobin measurement
418 BLOMKALNS & GIBLER

over a 9-hour period was most predictive of ad- inpatient admissions, making CPU protocols
verse events in their CPU population. more prevalent and important for eectively
CPUs likely will incorporate a multimarker managing the large number of patients presenting
approach that includes some combination of to the ED with chest discomfort. Rule-out MI
myoglobin, CK-MB, and troponin, and an in- is no longer an acceptable diagnosis to assign for
ammatory marker such as CRP [46,5860]. Ad- low and moderate risk patients with chest pain on
ditional biomarkers, such as brain natriuretic admission. Increased nancial pressures exerted
peptide (BNP), myeloperoxidase, and albumin co- by federal and private insurers on physicians and
balt binding assays, have been investigated and hospitals have indicated the need for more such
also show promising results [5964]. units. Emergency physicians, in particular, have
The evaluation of patients for rest ischemia is been challenged to develop more ecient and
also extremely important in the CPU setting. cost-eective strategies to evaluate these patients
Rapid perfusion imaging using radionuclide test- [70,71].
ing has been a revolutionary step for CPUs. Admissions for ultimately diagnosed noncar-
Perfusion imaging with agents, such as sestamibi, diac chest pain cost our society billions of dollars
has been used in various CPU settings and annually [72]. The CPU has allowed physicians to
protocols with great success [1115]. Tatum and condense a hospital admission into a 6- to 12-hour
colleagues [14] found that patients with normal evaluation, risk stratication, and observation pe-
imaging ndings had a 1-year event rate of 3% riod. Even CPUs of a decade ago proved to have
no MI or death compared with 42% with 11% ex- signicant economic advantages over hospital ad-
periencing MI and 8% cardiac death. Obtaining mission [9,73,74]. Roberts and colleagues [75] de-
multiple 12-lead ECGs can also be an eective termined that accelerated diagnostic protocols
method for detecting rest ischemia in patients for low-risk patients with chest pain saved total
who present with possible ACS [5,65]. Other im- hospital costs while reducing hospitalization rates
aging modalities, such as computed tomography and patient length of stay. Each CPU must be for-
and MRI, may also improve diagnostic perfor- mulated and evaluated based on the target patient
mance in the emergency setting [66,67]. In an population, hospital occupancy and resources,
ED in Rio de Janeiro, Brazil, a neural diagnostic and reimbursement patterns. Initial critics of
tree oered an innovative approach to the evalua- CPUs contend that this evaluation is expensive
tion of ACS in 566 patients, achieving sensitivities and cost prohibitive in some studies [76]. These
of 99% and 93% for ACS and AMI respectively, conclusions were largely based on previously im-
with a negative predictive value of 98% for both mature protocols. CPUs, by their innovative na-
[68]. ture, constantly must undergo a continuous
Finally, a CPU program must ultimately evolution toward optimal patient care and cost
evaluate patients for exercise-induced ischemia. savings.
Fixed coronary artery atherosclerotic lesions may In general, protocol and guideline driven
make a patient symptomatic only with exercise. medicine is eective for the evaluation of multiple
Standard graded exercise testing is extremely disease processes in the ED, chest discomfort
useful for these patients [57,9]. Many programs included. Continuous quality improvement
use treadmill or chemical stress testing with radio- should be an integral part of the continued
nuclide imaging or echocardiography to identify assessment of every CPU. The 2002 ACC/AHA
exercise-induced ischemia [3,1015]. guidelines for non-ST-segment elevation MI and
One frequently overlooked advantage of the unstable angina should be reected in every CPU
CPU, is the potential to educate patients about protocols diagnostic and treatment strategies [77].
atherosclerosis risk factors during their evalua- Finally, it should be mentioned that CPUs are
tion. For many patients, coming to the ED intended for the evaluation of low- to moderate-
represents a key interventional moment where risk patients, and not no-risk patients. Evalu-
education about coronary artery disease and risk ating all patients presenting with chest discomfort
modication may be particularly eective [69]. in a CPU would indeed be an inecient use of ED
and hospital resources. Continuous assessment of
alternative, more eective, and possibly less ex-
Cost eectiveness
pensive protocols will improve care standards for
Continued economic constraints and scarce these units over time. Already, the last decade has
inpatient resources discourage unnecessary seen a remarkable evolution of cardiac biomarker
CHEST PAIN UNIT CONCEPT 419

regimens and diagnostic and prognostic testing [13] Kontos MC, Anderson FP, Hanbury CM, et al. Use
for the evaluation of patients who present with of the combination of myoglobin and CK-MB mass
possible ACS. As more and more reports of for the rapid diagnosis of acute myocardial infarc-
successful CPUs are published and evidence tion. Am J Emerg Med 1997;15:149.
[14] Tatum JL, Jesse RL, Kontos MC, et al. Compre-
expands for cost-eectiveness, likely it will be an
hensive strategy for the evaluation and triage of
expectation, rather than an exception, that EDs the chest pain patient. Ann Emerg Med 1997;29:
and hospitals have such units. 11625.
[15] Abbott BG, Thwackers FJ. Emergency department
References chest pain units and the role of radionuclide imag-
ing. J Nucl Cardiol 1998;5:739.
[1] Pope JH, Aufderheide TP, Ruthazer R, et al. Missed [16] Lee-Lewandrowski E, Corboy D, Lewandrowski K,
diagnoses of acute cardiac ischemia in the emergency et al. Implementation of a point-of-care satellite lab-
department. N Engl J Med 2000;342:116370. oratory in the emergency department of an academic
[2] Nourjah P. National hospital ambulatory medical medical center. Impact on test turnaround time and
care survey: 1997 emergency department summary. patient emergency department length of stay. Arch
Advance data from vital and health statistics. Pathol Lab Med 2003;127:45660.
Hyattsville (MD): National Center for Health Statis- [17] Storrow AB, Gibler WB, Walsh RA. An emergency
tics; 2001. department chest pain rapid diagnosis and treatment
[3] Kontos MC, Anderson FP, Schmidt KA, et al. Early unit: results from a six year experience. Circulation
diagnosis of acute myocardial infarction in patients 1999;98:I-425.
without ST- segment elevation. Am J Cardiol 1999; [18] Jesse RL, Kontos MC. Evaluation of chest pain in
83:1558. the emergency department. Curr Probl Cardiol
[4] Gra LG, Dallara J, Ross MA, et al. Impact on the 1997;22:149236.
care of the emergency department chest pain patient [19] Farkouh ME, Smars PA, Reeder GS, et al. A clinical
from the chest pain evaluation registry (CHEPER) trial of a chest-pain observation unit for patients
study. Am J Cardiol 1997;80:5638. with unstable angina. Chest Pain Evaluation in the
[5] Gibler WB, Runyon JP, Levy RC, et al. A rapid di- Emergency Room (CHEER) Investigators. N Engl
agnostic and treatment center for patients with chest J Med 1998;339:18828.
pain in the emergency department. Ann Emerg Med [20] Nichol G, Walls R, Goldman L, et al. A critical
1995;25:18. pathway for management of patients with acute
[6] Amsterdam EA, Kirk JD, Diercks DB, et al. Imme- chest pain who are at low risk for myocardial ische-
diate exercise testing to evaluate low-risk patients mia: recommendations and potential impact. Ann
presenting to the emergency department with chest Intern Med 1997;127:9961005.
pain. J Am Coll Cardiol 2002;40:2516. [21] Gaspoz JM, Lee TH, Cook EF, et al. Outcome of
[7] Amsterdam EA, Kirk JD, Diercks DB. Early Exer- patients who were admitted to a new short-stay
cise testing for risk stratication of low-risk patients unit to rule-out myocardial infarction. Am J Car-
in chest pain centers. Crit Path Cardiol 2004;3: diol 1991;68:1459.
11420. [22] Lee TH, Juarez G, Cook EF, et al. Ruling out acute
[8] Sanchez CD, Newby LK, Hasselblad V, et al. Com- myocardial infarction. A prospective multicenter
parison of 30-day outcome, resource use, and coro- validation of a 12-hour strategy for patients at low
nary artery disease with and without diabetes risk. N Engl J Med 1991;324:123946.
mellitus assigned to chest pain units. Am J Cardiol [23] Lambrew CT, Bowlby LJ, Rogers WJ, et al. Factors
2003;91:122830. inuencing the time to thrombolysis in acute myocar-
[9] Mikhail MG, Smith FA, Gray M, et al. Cost-eec- dial infarction. Time to thrombolysis substudy of
tiveness of mandatory stress testing in chest pain the National Registry of Myocardial Infarction-1.
center patients. Ann Emerg Med 1997;29:8898. Arch Intern Med 1997;157:257782.
[10] Levitt MA, Promes SB, Bullock S, et al. Combined [24] Newby LK, Rutsch WR, Cali RM, et al. Time from
cardiac marker approach with adjunct two-dimen- symptom onset to treatment and outcomes after
sional echocardiography to diagnose acute myocar- thrombolytic therapy. GUSTO-1 Investigators.
dial infarction in the emergency department [see J Am Coll Cardiol 1996;27:164655.
comments]. Ann Emerg Med 1996;27:17. [25] Hoekstra JW, Hedges JR, Gibler WB, et al. Emer-
[11] Hilton TC, Thompson RC, Williams HJ, et al. Tech- gency department CK-MB: a predictor of ischemic
netium-99m sestamibi myocardial perfusion imag- complications. National cooperative CK-MB pro-
ing in the emergency room evaluation of chest ject group. Acad Emerg Med 1994;1:1727.
pain. J Am Coll Cardiol 1994;23:101622. [26] Young GP, Gibler WB, Hedges JR, et al. Serial
[12] Gersh BJ. Noninvasive imaging in acute coronary creatine kinase-MB results are a sensitive indica-
disease. A clinical perspective. Circulation 1991;84: tor of acute myocardial infarction in chest pain
I1407. patients with nondiagnostic electrocardiograms:
420 BLOMKALNS & GIBLER

the second Emergency Medicine Cardiac Research [39] Ottani F, Galvani M, Nicolini FA, et al. Elevated
Group Study. Acad Emerg Med 1997;4:86977. cardiac troponin levels predict the risk of adverse
[27] Gibler WB, Young GP, Hedges JR, et al. Acute outcome in patients with acute coronary syndromes.
myocardial infarction in chest pain patients with Am Heart J 2000;140:91727.
nondiagnostic ECGs: serial CK-MB sampling in [40] Christenson RH, Duh SH, Newby LK, et al. Cardiac
the emergency department. The Emergency Medi- troponin T and cardiac troponin I: relative values in
cine Cardiac Research Group. Ann Emerg Med short- term risk stratication of patients with acute
1992;21:50412. coronary syndromes. GUSTO-IIa Investigators.
[28] Polanczyk CA, Lee TH, Cook EF, et al. Value of ad- Clin Chem 1998;44:494501.
ditional two-hour myoglobin for the diagnosis of [41] Newby LK, Christenson RH, Ohman EM, et al.
myocardial infarction in the emergency department. Value of serial troponin T measures for early and
Am J Cardiol 1999;83:5259. late risk stratication in patients with acute coronary
[29] Fesmire FM, Percy RF, Bardoner JB, et al. Serial syndromes. The GUSTO-IIa Investigators. Circula-
creatine kinase (CK) MB testing during the emer- tion 1998;98:18539.
gency department evaluation of chest pain: utility [42] Antman EM, Tanasijevic MJ, Thompson B, et al.
of a 2-hour delta CK-MB of 1.6ng/ml. Am Heart Cardiac-specic troponin I levels to predict the risk
J 1998;136:23744. of mortality in patients with acute coronary syn-
[30] Alexander JH, Sparapani RA, Mahaey KW, et al. dromes. N Engl J Med 1996;335:13429.
Association between minor elevations of creatine ki- [43] Ohman EM, Armstrong PW, Christenson RH, et al.
nase-MB level and mortality in patients with acute Cardiac troponin T levels for risk stratication in
coronary syndromes without ST-segment elevation. acute myocardial ischemia. GUSTO IIA Investiga-
PURSUIT Steering Committee. Platelet glycopro- tors. N Engl J Med 1996;335:133341.
tein IIb/IIIa in unstable angina: receptor suppres- [44] Stubbs P, Collinson P, Moseley D, et al. Prognostic
sion using integrilin therapy. JAMA 2000;283: signicance of admission troponin T concentrations
34753. in patients with myocardial infarction. Circulation
[31] Davis CP, Barrett K, Torre P, et al. Serial myoglobin 1996;94:12917.
levels for patients with possible myocardial infarc- [45] Wu AH, Lane PL. Metaanalysis in clinical chemis-
tion. Acad Emerg Med 1996;3:5907. try: validation of cardiac troponin T as a marker
[32] Tucker JF, Collins RA, Anderson AJ, et al. Value of for ischemic heart diseases. Clin Chem 1995;41:
serial myoglobin levels in the early diagnosis of 122833.
patients admitted for acute myocardial infarction. [46] Benamer H, Steg PG, Benessiano J, et al. Compari-
Ann Emerg Med 1994;24:7048. son of the prognostic value of C-reactive protein and
[33] Sallach SM, Nowak R, Hudson MP, et al. A Change troponin I in patients with unstable angina pectoris.
in serum myoglobin to detect acute myocardial in- Am J Cardiol 1998;82:84550.
farction in patients with normal troponin I levels. [47] Tucker JF, Collins RA, Anderson AJ, et al. Early di-
Am J Cardiol 2004;94:8647. agnostic eciency of cardiac troponin I and Tropo-
[34] Brogan GX Jr, Vuori J, Friedman S, et al. Improved nin T for acute myocardial infarction. Acad Emerg
specicity of myoglobin plus carbonic anhydrase as- Med 1997;4:1321.
say versus that of creatine kinase-MB for early diag- [48] Brogan GX Jr, Hollander JE, McCuskey CF, et al.
nosis of acute myocardial infarction. Ann Emerg Evaluation of a new assay for cardiac troponin I vs
Med 1996;27:228. creatine kinase-MB for the diagnosis of acute myo-
[35] McCord J, Nowak RM, McCullough PA, et al. cardial infarction. Biochemical Markers for Acute
Ninety-minute exclusion of acute myocardial infarc- Myocardial Ischemia (BAMI) Study Group. Acad
tion by use of quantitative point-of-care testing of Emerg Med 1997;4:612.
myoglobin and troponin I. Circulation 2001;104: [49] Sayre MR, Kaufmann KH, Chen IW, et al. Mea-
14838. surement of cardiac troponin T is an eective
[36] Falahati A, Sharkey SW, Christensen D, et al. method for predicting complications among emer-
Implementation of serum cardiac troponin I as gency department patients with chest pain. Ann
marker for detection of acute myocardial infarction. Emerg Med 1998;31:53949.
Am Heart J 1999;137:3327. [50] Newby LK, Kaplan AL, Granger BB, et al. Compar-
[37] Johnson PA, Goldman L, Sacks DB, et al. Car- ison of cardiac troponin T versus creatine kinase-
diac troponin T as a marker for myocardial ische- MB for risk stratication in a chest pain evaluation
mia in patients seen at the emergency department unit. Am J Cardiol 2000;85:8015.
for acute chest pain. Am Heart J 1999;137: [51] Kontos MC, Anderson FP, Alimard R, et al. Ability
113744. of troponin I to predict cardiac events in patients ad-
[38] Ottani F, Galvani M, Ferrini D, et al. Direct com- mitted from the emergency department. J Am Coll
parison of early elevations of cardiac troponin T Cardiol 2000;36:181823.
and I in patients with clinical unstable angina. Am [52] Polanczyk CA, Lee TH, Cook EF, et al. Cardiac tro-
Heart J 1999;137:28491. ponin I as a predictor of major cardiac events in
CHEST PAIN UNIT CONCEPT 421

emergency department patients with acute chest [65] Decker WW, Prina CD, Smars PA, et al. Continuous
pain. J Am Coll Cardiol 1998;32:814. 12-lead electrocardiography monitoring in an emer-
[53] Myocardial infarction redeneda consensus docu- gency department chest pain unit: an assessment of
ment of The Joint European Society of Cardiology/ potential clinical eect. Ann Emerg Med 2003;41:
American College of Cardiology Committee for the 34251.
redenition of myocardial infarction. J Am Coll Car- [66] Achenback S, Ropers D, Pohle F-K, et al. The use of
diol 2000;36:95969. CTA in the chest pain center: a perspective. Crit Path
[54] Rao SV, Ohman EM, Granger CB, et al. Prognostic Cardiol 2004;3:8793.
value of isolated troponin elevation across the spec- [67] Kwong RY, Arai AE. Detecting patients with
trum of chest pain syndromes. Am J Cardiol 2003; acute coronary syndrome in the chest pain center
91:93640. of the emergency department with cardiac mag-
[55] Newby LK, Storrow AB, Gibler WB, et al. Bedside netic resonance imaging. Crit Path Cardiol 2004;
multimarker testing for risk stratication in chest 3:2531.
pain units: the Chest Pain Evaluation by Creatine [68] Bassan R, Pimenta L, Scofano M, et al. on behalf of
Kinase-MB, Myoglobin, and Troponin I (CHECK- the Chest Pain Project Investigators. Accuracy of
MATE) Study. Circulation 2001;103:18327. a neural diagnostic tree for the identication of acute
[56] Newby LK, Ohman EM, Christenson RH, et al. coronary syndrome in patients with chest pain and
Benet of glycoprotein IIb/IIIa inhibition in patients no ST-segment elevation. Crit Path Cardiol 2004;3:
with acute coronary syndromes and troponin t-pos- 728.
itive status: the paragon-B troponin T substudy. Cir- [69] Bahr RD. The changing paradigm of acute heart at-
culation 2001;103:28916. tack prevention in the emergency department: a fu-
[57] McCord J, Nowak RM, Hudson MP, et al. The turistic viewpoint? Ann Emerg Med 1995;25:956.
prognostic signicance of serial myoglobin, tropo- [70] National Heart Attack Alert Program issues report
nin I, and creatine kinase-MB measurements in on how to shorten intervention time. Am Fam Phy-
patients evaluated in the emergency department for sician 1994;50:156976.
acute coronary syndrome. Ann Emerg Med 2003; [71] Emergency department. rapid identication and
42:34350. treatment of patients with acute myocardial in-
[58] Pai JK, Pischon T, Ma J, et al. Inammatory farction. National Heart Attack Alert Program
markers and the risk of coronary heart disease in Coordinating Committee, 60 Minutes to Treat-
men and women. N Engl J Med 2004;351:2599610. ment Working Group. Ann Emerg Med 1994;23:
[59] Sabatine MS, Morrow DA, de Lemos JA, et al. Mul- 31129.
timarker approach to risk stratication in non-ST el- [72] Weingarten SR, Ermann B, Riedinger MS, et al.
evation acute coronary syndromes: simultaneous Selecting the best triage rule for patients hospitalized
assessment of troponin I, C-reactive protein, and with chest pain. Am J Med 1989;87:494500.
B-type natriuretic peptide. Circulation 2002;105: [73] Gomez MA, Anderson JL, Karagounis LA, et al. An
17603. emergency department-based protocol for rapidly
[60] Gibler WB, Blomkalns AL, Collins SP. Evaluation ruling out myocardial ischemia reduces hospital
of chest pain and heart failure in the emergency time and expense: results of a randomized study
department: impact of multimarker strategies and (ROMIO). J Am Coll Cardiol 1996;28:2533.
B-type natriuretic peptide. Rev Cardiovasc Med [74] Lee TH. Emergency department observation units:
2003;4(Suppl 4):S4755. has the time come? J Thromb Thrombolysis 1996;
[61] Brennan ML, Penn MS, Van Lente F, et al. Prognos- 3:25761.
tic value of myeloperoxidase in patients with chest [75] Roberts RR, Zalenski RJ, Mensah EK, et al. Costs
pain. N Engl J Med 2003;349:1595604. of an emergency department-based accelerated diag-
[62] Apple FS, Murakami MM, Pearee LA, et al. Multi- nostic protocol vs hospitalization in patients with
biomarker risk stratication of N-terminal pro-B- chest pain: a randomized controlled trial. JAMA
type natriuretic peptide, high-sensitivity C-reactive 1997;278:16706.
protein, and cardiac troponin T and I in end-stage [76] Shesser R, Smith M. The chest pain emergency de-
renal disease for all-cause death. Clin Chem 2004; partment and the outpatient chest pain evaluation
50:227985. center: revolution or evolution? Ann Emerg Med
[63] Baldus S, Heeschen C, Meinertz T, et al. Myeloper- 1994;23:33441.
oxidase serum levels predict risk in patients with [77] Braunwald E, Antman EM, Beasley JW, et al. ACC/
acute coronary syndromes. Circulation 2003;108: AHA 2002 guideline update for the management of
14405. patients with unstable angina and non-st-segment
[64] Bhagavan NV, Lai EM, Rios PA, et al. Evaluation elevation myocardial infarction: a report of the
of human serum albumin cobalt binding assay for american college of cardiology/american heart asso-
the assessment of myocardial ischemia and myocar- ciation task force on practice guidelines. J Am Coll
dial infarction. Clin Chem 2003;49:5815. Cardiol 2002;40(7):136674.
Cardiol Clin 23 (2005) 423451

Acute Coronary Syndromes in the Emergency


Department: Diagnostic Characteristics, Tests,
and Challenges
J. Hector Pope, MDa,b,c,*, Harry P. Selker, MD, MSPHb,c
a
Baystate Medical Center, 759 Chestnut Street, Springeld, MA 01199, USA
b
New England Medical Center, 750 Washington Street, #63 Boston, MA 02111, USA
c
Tufts University School of Medicine, Boston, MA 02111, USA

Failure to diagnose patients who have acute treating dysrhythmias and preventing or limiting
coronary syndromes (ACSs)deither acute myo- the size of AMI has grown, clinicians have tended
cardial infarction (AMI) or unstable angina to admit all patients who have even a low suspicion
pectoris (UAP)dwho present to the emergency of acute ischemia. As a result, clinicians have
department (ED) remains a serious public health generally admitted nearly all (92% to 98%)
issue. AMI is the leading cause of death for men patients presenting with AMI [49] and nearly
and women in the United States, with as many as 90% of those presenting with ACSs (including
1.2 million patients having myocardial infarctions those who have AMI and those who have UAP)
(MIs) annually [1] and about half of whom come [6,7,10]. The conscious strategy of maintaining
to EDs for their initial care. In addition, nearly a high diagnostic sensitivity (ie, that any error be
twice as many patients come to EDs having toward overdiagnosis) has the intended eect:
UAP. However, only 25% of patients who present among patients who have AMI who seek attention
to the ED with symptoms suggestive of an ACS, in EDs, the diagnosis is generally missed in 2% [3].
either AMI or UAP, will have a conrmed diag- High diagnostic sensitivity has been achieved at
nosis of the same on discharge [2]. The missed di- the cost of admitting many patients who do not
agnosis rate for AMI and UAP in this setting is have ACSs (low diagnostic specicity). Only 18%
about 2% each [3]. Thus clinicians have the unen- to 42% (typically about 30%) of the 1.5 million
viable task on the one hand of identifying, treat- patients admitted annually to CCUs [11] actually
ing, and hospitalizing (in the appropriate unit) experience AMI [7,1216], and only 50% to 60%
those patients who have a true ACS and on the have acute cardiac ischemia (ACI) [6,7,10,12].
other hand of avoid lling hospital teleme- Investigating the causes, progression, and
try, step-down units, and coronary care units treatment of ACSs continues to be a national
(CCUs) mostly with patients who have symptoms research priority, and this research continues to
suggestive of an ACS but who in fact do not have produce substantial progress in the areas of pre-
one of these syndromes. vention, diagnosis, and treatment of ACI and in
For many years, the diagnosis of ACSs was of advances in understanding its molecular and
more prognostic than therapeutic importance. cellular aspects. Over the past decade, there has
Over the past 3 decades, physicians diagnostic been a virtual revolution in our understanding of
and triage decisions for patients who have sus- the pathophysiology and management of coronary
pected cardiac ischemia have reected two tenden- artery disease (CAD) [17]. The conversion of a sta-
cies. First, as the number of acute interventions for ble atherosclerotic lesion into a ruptured plaque
with thrombosis has provided a unifying hypothe-
* Corresponding author. Tufts University School of sis for the cause of ACSs. The understanding of
Medicine, Boston, MA 02111. ACSs has evolved from this thesis. Thus, UAP
E-mail address: jameshpope@aol.com (J.H. Pope). (eg, rest angina, new-onset angina, increasing
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.012 cardiology.theclinics.com
424 POPE & SELKER

angina) and AMI are now well appreciated as cardiac imaging, myocardial performance indices,
parts of a continuum of myocardial ischemia. and computer-based decision aids remain under-
The overarching diagnosis of ACS has provided evaluated adjuncts to, but not substitutes for,
a framework for understanding the pathogenesis, clinicians judgment regarding diagnosis and tri-
clinical features, treatment, and outcome of pa- age of patients who have symptoms suggestive of
tients across the spectrum of myocardial ischemia. ACSs. However, despite 2 decades of focused
For emergency triage, the diagnosis of ACS better research and without compelling evidence for
identies patients for CCU or telemetry/step- a noncardiac cause, there remains no single way
down unit admission than does the diagnosis of to discriminate perfectly between those who
AMI alone. This fact is partly because of the di- should be admitted to exclude ACSs and those
culty in dierentiating unstable angina from in- who could be managed safely without admission.
farction, and partly because of intent, as it helps In essence, for the continuum of patient risk for
to reverse ischemia and prevent frank infarction. ACSs, the investigators have developed a matched
For patients who have ACSs and prolonged chest continuum of triage options for each patient, from
pain without infarction, the medium- and long- short-stay chest pain evaluation units to multiday
term mortality may be as poor or worse than for admissions for more intensive evaluations. This
those who actually have AMI [3,18]. In clinical article reviews the state of the art regarding the
medicine, much research has been focused on the diagnosis of ACSs in the emergency setting and
early diagnosis and treatment of ACSs. This re- suggests reasons why missed diagnosis continues
search has shown that early diagnosis and treat- to occur, albeit infrequently, and outlines strate-
ment of UAP is benecial and may prevent AMI. gies to deal with the remaining continuum of
For clinical reasons, to promote the optimal risk for patients who have suspected ACSs.
use of a limited resource, and to reduce unneces-
sary expenditure, research has focused on improv-
ing physicians diagnostic and triage accuracy. Methodologic Issues
However, there remains a need for improved
methods of diagnosis that can reduce unnecessary Consideration of the specic methods used in
hospitalization for patients incorrectly presumed studies of patients who have ACSs is vital when
to have acute ischemia without increasing the critically reviewing studies of the diagnosis/
number of patients who have acute ischemia who misdiagnosis and triage of patients in the ED
are sent home inappropriately [19]. To this end, who have suspected cardiac ischemia. The key
and as mandated by Congress, in 1991 the Na- methodologic issues for applicability of study re-
tional Heart, Lung, and Blood Institute of the Na- sults are:
tional Institutes of Health instituted the National
 Representative patient sample
Heart Attack Alert Program to focus on issues re-
 Representative prevalence of ischemic heart
lated to the rapid recognition and response to pa-
disease
tients who have symptoms and signs of ACSs in
 Broad patient inclusion criteria, not just chest
emergency settings, and made reports in 1997
pain
[20] and 2000 [21] on technologies for identifying
 Study setting includes a range of settings
ACI in such settings.
 Diagnostic end point includes unstable angina
For as many as 50% of women and a small
and acute infarction
percentage of men who have a suspected ACSs,
 Completeness of follow-up
other mechanisms of illness are at work, including
 Follow-up data appropriate and signicant
vasospastic angina, microvascular endothelial
 Validation of ndings in generalizable trials
dysfunction, transient left ventricular apical bal-
looning syndrome, and mitral valve prolapse. The Central to any study is whether the patient
accurate identication of these syndromes in the sample studied is representative of patients in the
ED or other acute care settings challenges the skill ED who are seen in actual practice. Also, the
of even the most seasoned clinician. Clinical positive predictive value (ie, the proportion of
features, patient demographics, physical examina- patients that actually has ACSs among all those
tion ndings, and a rm understanding the who have a positive test or attribute) of a symp-
strengths and limitations of the standard ECG tom, sign, or test result depends on the prevalence
can greatly assist with the identication of pa- of ischemic heart disease in the study population
tients who have ACSs. Biochemical markers, [22]. Thus, the proportion of patients who have
ACS IN THE ED 425

false-positive results will be higher (and positive is one of the most common and complex, ac-
predictive value lower) in a population that has counting for about 5.6 million ED visits annually
a low prevalence of ischemia (all patients in the [1]. It is important to keep in mind that many pa-
ED) compared with a population that has a high tients will not admit having chest pain, but will
prevalence (patients in the CCU). Even studies acknowledge the presence of chest discomfort
performed in EDs may not be comparable when because of their denition of pain. Published re-
ACSs prevalence is signicantly dierent. Inclu- ports suggest that up to 5% of visits to the ED in-
sion criteria can limit studies of patients in the volve complaints relating to chest discomfort [30].
ED if, for example, only patients who have chest The complaint of chest discomfort encompasses
pain are studied [2325], compared with the use many varying conditions, ranging from insigni-
of broad entry criteria, including multiple symp- cant to high-risk in terms of threat to the patients
toms that could be anginal equivalents such as life, including, but not limited to, acute coronary
any chest discomfort, epigastric pain, arm pain, syndromes (AMI and UAP), thromboembolic dis-
shortness of breath, dizziness, or palpitations ease (pulmonary embolism), aortic dissection,
[26]. Study setting (eg, urban versus rural hospital, pneumothorax, pneumonia, myocarditis, and
teaching versus community hospital) can also af- pericarditis. Chest discomfort may be perceived
fect the applicability of any ndings to various as pain with descriptions such as crushing, vice-
practice settings. like constriction; a feeling equivalent to an ele-
Aside from the study sample, other methodo- phant sitting on the chest; tightness; pressure;
logic issues warrant attention, including the ap- heartburn; or indigestion, or as discomfort most
propriateness of the measured diagnostic end noticeable for its radiation to an adjacent area
point. Some past ED studies have focused on of the body such as the neck, jaw, interscapular
identifying or predicting only AMI, but identify- area, upper extremities, or epigastrium. Elderly
ing UAP is also important for monitoring and patients or those who have diabetes may have al-
early therapy, especially considering that approx- tered ability to specically localize discomfort
imately 9% of patients admitted with new-onset [31]. Individuals of each gender and dierent cul-
angina or UAP progress to infarction [27,28]. tural groups vary in their expression of pain and
Completeness of follow-up must be considered. ability to communicate with health professionals,
Studies with substantial numbers of patients lost so that presentation may range from merely both-
to follow-up may have ascertainment bias, espe- ersome to cataclysmic for conditions that seem
cially when the participation rate among eligible nearly equivalent when objective criteria are
patients is not high. Also important is the type matched. The level of discomfort does not neces-
of follow-up data collection; for example, the oc- sarily correlate with the severity of illness, making
currence of AMI will be underestimated if fol- identication of potentially life-threatening condi-
low-up evaluation does not include biomarker tions dicult in certain patients. Because of the
determination results. serious nature of many conditions presenting
Finally, validation of the ndings of clinical with chest discomfort and the potential for signif-
studies is critical, especially for prediction rules icant reduction in morbidity and mortality with
and diagnostic aids; ndings may be center- or early diagnosis and treatment, clinical policies
data-dependent. The ideal validation study is a pro- have been developed to guide clinicians with their
spective trial of a ndings or prediction rules initial evaluation of chest discomfort, emphasizing
eects on patient care in diverse settings [29]. prompt triage, assessment, and initiation of ther-
apy [32]. This article only reviews the policies
that apply to ACSs.
Clinical presentation
It is sometimes dicult to distinguish cardiac
The clinically obtained history must be concise, from noncardiac chest discomfort, even though
yet detailed enough to establish the probability of chest pain is the hallmark of ACSs. Taking the
ACSs, and obtained expeditiously so as not to time to elicit the exact character of the sensation
delay implementation of therapy. (without prompting the patient, if possible) and
any pattern of radiation (if present) is most
helpful. Typically, the chest discomfort of acute
Chest discomfort
ischemia has a deep visceral character, preventing
Of all the symptoms for which patients seek the patient from localizing the discomfort to
emergency medical care, chest pain or discomfort a specic region of the chest. It is often described
426 POPE & SELKER

as a pressure-like or heavy weight on the chest, [37], but chest pain that radiates to the arms or
a tightness, a constriction about the throat, or an neck does increase the likelihood [3840]. In the
aching sensation that is not aected by respira- study by Sawe [37] that looked at admitted pa-
tion, position, or movement and that comes on tients who had AMI, 71% had pain radiation to
gradually, reaches its maximum intensity over arms or necks, whereas 39% of admitted patients
a period of 2 to 3 minutes, lasts longer than 30 who experienced pain radiation did not have
minutes rather than seconds, may wax and wane, AMI. Consistent with the classical description,
and may be remitting. It may be described as 33% of patients who proved to have infarction
indigestion and occasionally may be relieved by had radiation to both arms, 29% had pain radiat-
belching. In a large study of patients in the ED ing to the left arm only, and 2% to the right arm
who had suspected ACSs, Pope and colleagues [2] only [37].
found that the 76% of patients who presented Some investigators believe that a signicant
with the complaint of chest pain or discomfort (in- number of patients who have cardiac ischemia can
cluding arm, jaw, or equivalent discomfort) had present with abdominal pain as their chief com-
a 29% incidence of ACSs at nal diagnosis plaint [23,24]. However, Pope and colleagues [2]
(10% AMI, 19% UAP). In 69% of patients, chest found that 14% of study patients had this com-
pain or discomfort was the chief complaint, and plaint. This group had a 15% incidence of ACSs
this group had a 31% incidence of ACSs (10% at nal diagnosis (6% AMI, 9% UAP), but less
AMI, 21% UAP). In 21% of patients, it was the than 1% of these patients complained of abdom-
only complaint, and this group had a 32% inci- inal pain as their chief or only complaint and
dence of ACSs (9% AMI, 23% UAP). Further- had a 4% incidence of ACI (2% AMI, 2%
more, chest pain or discomfort as the chief UAP). Abdominal pain as a chief complaint or
complaint or presenting symptom was more fre- presenting symptom was associated with a higher
quently associated with a nal diagnosis of ACS incidence of a non-ACS nal diagnosis (0.6%
(88% ACS versus 62% non-ACS and 92% ACS non-ACS versus 0.1% ACS; 16% non-ACS ver-
versus 71% non-ACS, respectively; P .001). sus 9% ACS, respectively; P .001.002). Esoph-
Sharp, stabbing, or positional pain is less likely ageal reux and motility disorders are common
to represent ischemia [33] but does not exclude masqueraders of ACS. In a study of all patients
it. Lee and colleagues [34] found that among pa- discharged from a CCU with undetermined causes
tients in the ED who had sharp or stabbing of chest pain, over half had esophageal dysfunc-
pain, 22% had acute ischemia (5% AMI, 17% tion [41]. When these patients presenting com-
UAP). Among those who had partially pleuritic plaints were compared with those of patients
pain, 13% had acute ischemia (6% AMI, 7% who did not have ACSs, those who had esoph-
UAP), and among the group that had fully pleu- ageal disorders were more likely to complain of
ritic pain, none was shown to have acute ischemia. a lump in their throat, acid taste, overfullness af-
Of the patients whose pain was fully reproduced ter eating, a hacking cough, and chest pain that
by palpation, 7% nonetheless had acute ischemia caused awaking at night, and they were less likely
(5% AMI, 2% UAP), and 24% had pain partially to report eort-related chest pain, a history of
reproduced with palpation had ischemia (6% nitroglycerin use, or reliable chest pain relief
AMI, 18% UAP). Patients who describe their dis- with its use.
comfort as similar to previous episodes of cardiac
ischemia are more likely to have ACSs [16,35], but
Anginal pain equivalents
any chest discomfort carries a higher risk than no
discomfort [6,26,36]. Dyspnea, present in about one third of patients
Combinations of variables improved discrimi- who have infarction in some series [24,38,42], is
nation in these patients [24]. In patients who had the most important angina equivalent. In their
sharp or stabbing pain that was also pleuritic, posi- multicenter ED trial, Pope and colleagues [2]
tional, or reproducible by palpation, 3% had UAP found that 16% of patients who had suspected
and none had AMI. Furthermore, if these same pa- ACSs presented with a chief complaint of short-
tients had no history of ischemic heart disease, none ness of breath and had an 11% incidence of
had acute ischemia. The partially and fully ASCs at nal diagnosis (6% AMI, 5% UAP); in
groups were subjective and small in number. 8%, this was the only complaint, with a 10% inci-
Exact location of chest pain is not signicantly dence of ACS (5% AMI, 5% UAP). However, a -
dierent in patients who have or do not have AMI nal diagnosis of ACSs was not more frequent in
ACS IN THE ED 427

patients who had a presenting symptom of short- changes, 20% a change in appearance (ie,
ness of breath (56% ACS versus 56% non-ACS; looked pale), and 8% to 10% dizziness
P .5); as a chief complaint, shortness of breath [24,42]. Pope and colleagues [2] found that 28%
was more commonly associated with a nal diag- of patients who had suspected ACSs presented
nosis of non-ACS (18% non-ACS versus 7% to the ED with dizziness and had a 16% incidence
ACS; P .001), possibly reecting a high preva- of ACSs (5% AMI, 11% UAP). Dizziness was the
lence of patients who have lung disease in the primary complaint in 5% of study patients, with
study population. Yet, because 4% to 14% of pa- a 4% incidence of ACSs (2% AMI, 2% UAP),
tients who have AMI [23,24,26] and 5% of pa- and it was the only symptom in 1% of patients
tients who have UAP present only with sudden (2% AMI, 0% UAP). In the same study, dizziness
diculty breathing [2], ACSs should be consid- or fainting as a chief complaint were more com-
ered as a cause of unexplained shortness of breath. monly associated with a nal diagnosis of non-
Diaphoresis and vomiting, when associated ACS (7% non-ACI versus 1% ACS, P .001).
with chest pain, increase the likelihood of in- Similarly, dizziness or fainting as a presenting
farction [15,29,38]. Diaphoresis occurs in 20% to symptom was more frequently associated with -
50% of patients who have AMI [39,43]. One study nal diagnoses of non-ACS (31% non-ACS versus
showed that the presence of nausea without vom- 19% ACS; 8% non-ACS versus 2% ACS; respec-
iting did not discriminate, but vomiting was sig- tively; P .001). ECG evaluation is helpful in
nicantly more frequent in patients who ruled low-prevalence patients who have these vague
in [38]. Pope and colleagues [2] found nausea in complaints.
28% of patients who had suspected ACSs. Pa-
tients who had nausea as a presenting symptom
Atypical presentations
had a 26% incidence of ACSs at nal diagnosis
(10% AMI, 16% UAP); patients who had nausea Few studies address what proportion of pa-
or vomiting as chief complaint (2%) had a 15% tients in the ED who have ACSs present with
incidence of ACSs (11% AMI, 4% UAP); and atypical symptomsda group for which the di-
less than 1% of patients had nausea or vomiting agnostic/triage decision is often most problematic.
as their only symptom. The same study found Among hospitalized patients who have AMI, 13%
vomiting present in 10% of patients. Patients to 26% had no chest pain or had chief complaints
who had vomiting as a presenting symptom had other than chest pain (eg, dyspnea, extreme
a 23% incidence of ACSs (13% AMI, 10% fatigue, abdominal discomfort, nausea, syncope)
UAP) and patients who had vomiting as the chief [23,24]. In a large ED study of patients presenting
or only complaint had less than 1% incidence. with a wide range of clinical symptoms, Pope and
Furthermore, investigators showed that a chief colleagues [2] found that 31% of patients who had
complaint of nausea or vomiting was more suspected ACSs presented without chest pain,
frequently associated with a nal diagnosis of with a 26% incidence of ACSs at nal diagnosis
non-ACS (0.5% non-ACS versus 0.3% ACS; (18% infarction, 8% unstable angina), and had
P .15), yet a presenting complaint of nausea chief complaints other than chest pain (eg, short-
was more commonly associated with a nal diag- ness of breath, abdominal pain, nausea, vomiting,
nosis of ACSs (30% ACS versus 27% non-ACS; dizziness, fainting). Subanalysis of this same data
P .004); a presenting complaint of vomiting by Coronado and colleagues [31] found pain to be
did not show this association (10% ACS versus absent in 6.2% of patients who had acute ischemia
10% non-ACS; P .7). In a CCU study, 43% and 9.8% of patients who had AMI. These inves-
of patients who had Q-wave infarction and only tigators found that age and heart failure were in-
4% of patients who had nonQ-wave infarctions dependently associated with painless ACSs, in
or prolonged angina had vomiting [44]. addition to diabetes mellitus among those who
So-called soft clinical features, such as fa- had AMI, and that lack of pain predicted
tigue, weakness, malaise, dizziness, and clouding increased hospital mortality. Finally, in a large
of the mind, are surprisingly frequent, occurring group of patients on Medicare who were hospital-
in 11% to 40% of patients who have AMI ized with conrmed UAP, Canto and colleagues
[29,37,38,42]. Prodromal symptoms (those occur- [45] found over half of these patients had atypical
ring in the preceding days or weeks) are also fre- presentations. Independent predictors of atypical
quent: 40% report unusual fatigue or weakness, presentation for patients who had UAP were
20% to 39% dyspnea, 14% to 20% emotional older age, history of dementia, and absence of
428 POPE & SELKER

prior MI, hypercholesterolemia, or family history Society class 3 angina with new symptoms or
of heart disease. Patients who had atypical presen- symptoms that changed within 3 days before pre-
tation received aspirin, heparin, and beta-blocker sentation, inaccuracies in the clinician assessment
therapy less aggressively, but there was no dier- of the dynamic nature of anginal symptoms may
ence in mortality. have contributed to the failure to hospitalize pa-
Among patients in the ED, no single atypical tients who had UAP.
symptom is of overwhelming diagnostic impor-
tance, although combinations of symptoms can
Past medical history
identify high-risk patients who should be admitted
regardless of ECG ndings. Pope and colleagues In addition to the presenting clinical features,
[2] ranked atypical presenting symptoms in de- the presence of a CAD risk factor has tradition-
creasing order of association with ACI at nal di- ally been considered diagnostically helpful in the
agnosis as follows: nausea (26%), shortness of ED setting. In a large ED series [2] an association
breath (24%), vomiting (23%), dizziness (16%), was shown between patients having a past history
abdominal pain (15%), and fainting (65%). Canto of diabetes mellitus (31% ACS versus 18% non-
and colleagues [45], in their study of patients who ACS; P .001), MI (45% ACS versus 20%
had UAP, found the most frequent symptoms to non-ACS; P .001), or angina pectoris (63%
be dyspnea (69%), nausea (38%), diaphoresis ACS versus 29% non-ACS; P .001), and a nal
(25%), syncope (11%), and pain in the arms diagnosis of ACSs; however, these ndings re-
(12%), epigastrium (8%), shoulder (7%), or quire careful interpretation. From the Framing-
neck (6%). ham Study, it is well known that the risk for
Data from community-based epidemiologic developing ischemic heart disease is increased
studies [25,4648] suggest that 25% to 30% of over decades by male gender, advancing age,
all Q-wave infarctions go clinically unrecognized. a smoking habit, hypertension, hypercholesterol-
Of these, half were truly silent and half were asso- emia, glucose intolerance, ECG abnormalities,
ciated with atypical symptoms in retrospect a type A personality, a sedentary lifestyle, and
[25,46]. Because Q waves often resolve (in the a family history of early CAD [24,51,52]. Clini-
Framingham Study)d10% of patients were dis- cians customarily assess these factors when pro-
charged after anterior infarction and 25% of viding preventive care because they predict the
those discharged after inferior infarction lost their incidence of future coronary disease. However,
Q waves within 2 yearsdthe true incidence was coronary risk factors were established to provide
underestimated [49]. an estimate of risk over years. Thus, the Framing-
The rate of erroneous discharge from the ED ham Study showed that hypertension increases the
of patients who have AMI may be a marker for risk for ischemic heart disease twofold over 4
atypical cases, but such studies are limited by years [25], but only a very small portion of this
inclusion criteria, small numbers, and lack of risk applies to the few hours of acute illness that
complete follow-up. Rates of 2% [50], 4% [9], the patient in the ED experiences. A patients re-
and as high as 8% [8] have been reported. In port of coronary risk factors is also subject to
a large ED series [3], patients who had suspected biases and inaccuracies. This history is presum-
ACSs reported rates of erroneous discharge of ably less reliable than the methods used to assign
2% (2.1% AMI, 2.3% UAP). The early mortality risk in longitudinal studies.
(30-day) for these missed AMIs may be as high as In a multicenter study, Jayes and colleagues
10% to 33% [3,8,9]. [53] found that most of the classic coronary risk
Finally, in a large ED study of patients who had factors have little predictive value for ACSs
UAP, Pope and colleagues [3] found that 2.3% when used in the ED setting. Except for diabetes
were not hospitalized. Over three fourths of the pa- and a positive family history in men, no coronary
tients were evaluated by an attending physician, risk factor signicantly increased the likelihood
and more than one fourth by a consulting cardiol- that a patient had acute ischemia. Diabetes and
ogist. Although there was disagreement over the family history each confer only about a twofold
interpretation of 16% of the ECGs on subsequent relative risk for acute ischemia in men, whereas
review by an experienced cardiologist, this was not chest discomfort, ST-segment abnormalities, and
believed to be clinically signicant in any of the T-wave abnormalities confer relative risks of
cases. Given that most of the patients who were about 12-, 9-, and 5-fold, respectively. Because
not hospitalized had Canadian Cardiovascular these results run counter to the prevailing clinical
ACS IN THE ED 429

wisdom, it is possible that physicians who give In a large series of patients in the ED who had
risk factor history great weight may inappropri- suspected ACS, median rst and highest systolic
ately diagnose/triage patients in the ED, which blood pressures (SBPs) were higher in patients
is an issue that deserves further attention and who had a nal diagnosis of ACSs [2]. This nd-
investigation. ing suggests that the adrenergic excess associated
Finally, a past history of medication use for with ACSs might be greater than that associated
coronary disease increases the likelihood that the with non-ACS diagnoses. However, to use this hy-
current chest pain is an ACS. In the Boston City pothesis as a predictive factor, clinicians must
Hospital and the multicenter predictive instru- have some idea of their patients baseline blood
ment trials, a history of nitroglycerin use was pressure, which is not the case in most ED evalu-
found to be one of the most powerful predictors ations. Thus, the usefulness of this observation
of ACSs [6]. Although nitrates are an accepted may be limited.
mainstay in treating acute and chronic coronary In the same series [2], in addition to the eect
disease, the diagnostic and prognostic value of adrenergic release during acute ischemia, the
of chest pain relief with nitroglycerin has been higher initial and highest pulse pressures found
poorly studied. In the largest study to date, Hen- in patients who had a nal diagnosis of ACSs
rikson and colleagues [54] found that chest pain may also reect the lower compliance of the ische-
relief with nitroglycerin did not accurately predict mic left ventricle. Of relevance to those who are
active CAD or subsequent outcomes in a general candidates for thrombolytic therapy, excess pulse
population presenting to an ED. Nitrates can blood pressure (the extent to which a patients
cause dramatic relief of chest pain from esopha- pulse pressure exceeded 40 mmHg for patients
geal spasms [55], thus further questioning their di- who had an SBP of more than 120 mmHg) places
agnostic ability in ischemic heart disease. these patients at increased risk for thrombolysis-
related intracranial hemorrhage [29].
Pope and colleagues [3] found that median rst,
Physical examination
median highest, and median lowest SBPs of pa-
The physical examination is generally not tients who had AMI and were subsequently classi-
helpful in diagnosing ACSs when compared with ed as Killip class 4 (cardiogenic shock) were above
the value of historical data and ECG ndings, the threshold of this classication (SBP %90
except when it points to an alternate process. On mmHg) for these three blood pressure observa-
the other hand, clinicians must not be lulled into tions. This nding suggests that the adrenergic ex-
a sense of security by chest pain that is partially or cess associated with ACSs may be greater than that
fully reproduced by palpation, because 11% may associated with non-ACS diagnoses. Although the
have infarction or UAP [26]. Pope and colleagues number of such patients in this analysis was small,
[2] found the pulse rate to be lower in patients it did suggest that patients who have ACSs can
who had a nal diagnosis of ACSs versus those present with apparently normal blood pressures
who had a nal diagnosis of non-ACS (P .02), and can go on to develop cardiogenic shock.
but this dierence was not considered clinically Abnormal vital signs and certain combinations
signicant. of these have been shown to be critical observa-
Pulse rate observation in isolation appeared to tions in clinical outcome prediction. The reported
be generally not helpful in ACSs identication. probability of infarction decreases with a normal
First, the patients pulse rate could be slowed by respiratory rate [38] and increases with diaphore-
the presence of b-blockers as part of a prior sis [15], but other signs mainly help identify
treatment regime, coincident vagal stimulation high-risk patients who have infarction [56]. In
from ACSs (eg, reex bradycardia and vasode- the predictive instrument for AMI mortality
pressor eects associated with inferoposterior wall proposed by Selker and colleagues [36], blood
AMI), or diagnostic/therapeutic procedures in the pressure, pulse, and their interaction gured
ED (eg, phlebotomy, intravenous access). Second, prominently in three of the six clinical variables
the patients pulse may be increased by adrenergic used to develop the prediction instrument.
excess from just having to come to the ED and Finally, rales (of any degree), but not S3 gallops,
everything that accompanies such a visit, in ad- have been more frequently seen in patients who
dition to the adrenergic excess (eg, tachycardia have a nal diagnosis of ACSs [2]. This nding is not
and increased peripheral vascular resistance) as- surprising, as several clinical syndromes of pump
sociated with possible ongoing ACSs. failure can complicate ACSs. The investigators
430 POPE & SELKER

failure to nd association between an S3 gallop may poorly represent the patients ischemic
rhythm and ACSs at nal diagnosis is surprising, myocardium.
but it may have to do with a failure to document Second, 12-lead ECG is limited by its lack of
this nding consistently in the medical record on perfect detection [63]. Small areas of ischemia or
the part of the physicians in the EDs at study sites. infarction may not be detected; conventional leads
do not examine satisfactorily the right ventricle
[64] or posterior basal or lateral walls (eg, AMIs
ECG in the distribution of the circumex artery)
Standard 12-lead ECG [65,66].
Third, some ECG baseline patterns make
A complete summary of evidence related to the interpretation dicult or impossible, including
diagnostic usefulness of the standard ECG was prior Q waves, early repolarization variant, left
recently published [20,57], and this background ventricular hypertrophy, bundle branch block,
will not be repeated in this article. However, the and dysrhythmias [67]. Lee and colleagues [9]
National Heart Attack Alert Program (NHAAP) demonstrated that when the current ECG shows
Working Group on Evaluation of Technologies ischemic ndings, availability of a prior compari-
for Identifying Acute Cardiac Syndrome [57] son ECG improved triage.
found that most studies evaluate the accuracy of Fourth, ECG waveforms are frequently di-
the technologies and only a few evaluate the clini- cult to interpret, causing disagreement among
cal impact of routine use. Furthermore, the group readersdso-called missed ischemia. In a study
concluded that although the standard ECG is of patients who had AMI and were sent home,
a safe, readily available, and inexpensive technol- ECGs tended to show ischemia or infarction not
ogy with a high sensitivity for AMI, it is not highly known to be old, with 23% of the missed diagnoses
sensitive or specic for ACSs. However, the ECG owing to misread ECGs [8]. Jayes and colleagues
ndings remain an integral, if not the most impor- [53] compared ED physician readings of ECGs
tant, part of the evaluation of patients who have with formal interpretations by expert electrocar-
chest pain, and the Working Group recommended diographers, and calculated sensitivities of 0.59
that they remain the standard of care for evaluat- and 0.64 and specicities of 0.86 and 0.83 for ST-
ing patients who have chest pain in the ED. segment and T-wave abnormalities, respectively.
The ECG provides essential information when McCarthy and colleagues [18] and a review of liti-
the diagnosis is not obvious by symptoms alone gation in missed AMI cases [68] emphasized this
[58], despite one study noting that the results of factor of incorrect ECG interpretation. In the larg-
the ECG infrequently changed triage decision est study to date of ACSs in the ED, Pope and col-
based on initial clinical impressions [59]. The gen- leagues [3] found that although the rate of missed
erally dominant weights given to ECG variables in diagnoses of ACSs (2.1% AMI, 2.3% UAP) was
mathematic models for predicting ACSs substan- low, there was a small but important incidence of
tiate this impression [6,7,10,15,16]. Moreover, failure by the ED clinician to detect ST-segment el-
the initial ECG is increasingly important in intra- evations of 1 to 2 mm in the ECGs of patients who
hospital triage because of its value in predicting had MI (11%). Correct ECG interpretation by ED
complications of AMI [6062]. physicians is doubly important today because of
The fundamental limitations in the standard the need to use interventions such as thrombolytic
ECG include: agents and percutaneous coronary intervention
appropriately in ACSs.
 Single brief sample
Fifth, the implications of the ECG ndings
 Lack of perfect detection
must be interpreted in their clinical context,
 Confounding baseline patterns
a process performed intuitively by clinicians and
 Interpretation
formally stated in Bayesian analysis. When symp-
 Clinical context
toms alone strongly suggest ischemia, a normal or
 Imperfect sensitivity and specicity
minimally abnormal ECG will not substantially
First, it is a single brief sample of the whole decrease the probability of ischemia. Conversely,
picture of the changing supply-and-demand char- when the presentation is inconsistent with acute
acteristics of unstable ischemic syndromes. If ischemia, an abnormal ECG (unless diagnostic
a patient who has UAP is temporarily pain-free abnormalities are present) will only modestly
when the ECG is obtained, the resulting tracing increase the likelihood of ischemia. Bayes rule
ACS IN THE ED 431

tells us that the ECG will have the greatest impact segment elevation alone will have had an infarc-
when symptoms are equivocal [69]. tion. Other investigators found that if Q waves
Finally, the ECG suers from imperfect sensi- and ST-segment elevation were present, 82% to
tivity and specicity for ACS. When interpreted 94% actually sustained AMI [63]. However, it
according to liberal criteria for MI (ie, ECGs that must be remembered that ST-segment elevation
show any of the following as positive for AMI: can occur in the absence of ischemia (ie, early
nonspecic ST-segment or T-wave changes ab- repolarization variant, pericarditis, left ventricu-
normal but not diagnostic of ischemia; ischemia, lar hypertrophy, and previous infarction even in
strain, or infarction, but changes known to be old; the absence of a ventricular aneurysm) [76]. Con-
ischemia or strain not known to be old; and versely, Pope and colleagues [2] showed that
probable AMI), the ECG operates with high (but a large percentage of patients who have ACS
not perfect) sensitivity (99%) for AMI, at the cost (20% AMI, 37% UAP) can present with initial
of low specicity (23%; positive predictive value normal ECGs.
21%; negative predictive value 99%). Conversely, In their study of patients in the ED who had
when interpreted according to stringent criteria suspected ACS, Pope and colleagues [2] found that
for AMI (only ECGs that show probable AMI), ST-segment elevation of either 1 to 2 mm or more
sensitivity (61%) drops and specicity equals 95% than 2 mm was more frequently associated with a -
(positive predictive value 73%; negative predictive nal diagnosis of ACSs (9% ACS versus 7% non-
value 92%) [20]. ACS; 5% ACS versus 1% non-ACS, respectively;
Despite its usefulness, the ECG is insu- P .001). A full 30% of patients who have ST-
ciently sensitive to make the diagnosis of ACSs segment elevation of 1 mm or greater had a nal
consistently. The ECG should not be relied on to diagnosis of AMI. In addition, in a study of missed
make the diagnosis but rather should be included diagnosis of ACSs in the ED, Pope and colleagues
with history and physical examination character- [2] found a small but important incidence of failure
istics to identify patients who appear to have by the ED clinician to detect ST-segment elevations
a high risk for ACSs (ie, a supplement to, rather of 1 to 2 mm in the ECGs of patients who had AMI
than a substitute for, physician judgment). In (11%). This incidence represents an important and
rule-out AMI patients, a negative ECG carries potentially preventable contribution to the failure
an improved short-term prognosis [60,7073]. to admit such patients.
Providing the interpreter with old tracings would ST-segment depression usually indicates suben-
intuitively seem to be of value because baseline docardial ischemia. If these abnormalities are new,
abnormalities make current evaluation dicult. persistent, and marked, the likelihood of AMI
However, Rubenstein and Greeneld [74], in increases. About 50% to 67% of admitted patients
a study of 236 patients presenting to EDs with who have new or presumed new isolated ST-
the complaint of chest pain, found that only segment depression have infarctions [64,75]; even
a small proportion might have beneted from more patients have probable ischemia. Pope and
having a previous baseline ECG available (5% colleagues [2] found that all degrees of ST-segment
might have avoided unnecessary admission). Fur- depression (0.5, 1, 12, and R2 mm) were more
thermore, there was no patient for whom a base- commonly associated with a nal diagnosis of
line ECG would have aided in avoiding an ACSs (12% ACS versus 7% non-ACS; 8% ACS
inappropriate discharge. ECG sampling should versus 3% non-ACS; 2% ACS versus 0% non-
be periodic, not just static. The pitfalls of not or- ACS, respectively; P .001). A full 19% of patients
dering ECGs in younger, atypical patients, and who had ST-segment depression of at least 0.5 mm
of misinterpretation should be anticipated. Fi- or greater had a nal diagnosis of AMI. Quantita-
nally, clinicians should not be reluctant to obtain tive ST-segment depression and cardiac Troponin
a second opiniondby fax transmission if neces- T status have been found to be complementary in
sarydfor dicult tracings. assessing risk among ACS patients [77]. ST-seg-
ment depression may also occur in nonischemic set-
tings, including patients who are hyperventilating,
ST-segment and T-wave abnormalities
those taking digitalis, those who have hypokale-
ST-segment and T-wave abnormalities are the mia, and those who have left ventricular strain
sine quo non of ECG diagnosis of ACSs. (without voltage criteria) [76].
Numerous studies [63,73,75] have found that Inverted T waves may reect acute ischemia.
65% to 85% of CCU patients who have ST- One study showed that isolated T-wave inversion
432 POPE & SELKER

occurred in 10% of CCU admissions, of whom overdiagnosis (no comparison ECG available)
22% had AMI [78]. T-wave changes may reect and underdiagnosis (baseline abnormality obscu-
prior myocardial damage or left ventricular strain ration of ischemia) [82]. Lee and colleagues [34]
[76]. The study by Pope and colleagues [2] found found that emergency patients who had chest
that certain T-wave patterns (inverted 15 mm, in- pain and nondiagnostic ECG abnormalities had
verted R5 mm, or elevated) were more frequently a low risk for AMI but a signicant risk for
associated with a nal diagnosis of ACS (32% ACSs. Pope and colleagues [3] found that 53%
ACS versus 17% non-ACS; 1% ACS versus 0% of patients in the ED who had a missed diagnosis
non-ACS; 4% ACS versus 1% non-ACS, respec- of AMI had normal or nondiagnostic ECGs, as
tively; P .001). Flattened T waves did not did 62% of patients who had a missed diagnosis
have the same association with an ACS nal diag- of UAP.
nosis (18% ACS versus 20% non-ACS; P .001).
Furthermore, 39% of patients who had inverted T
waves of at least 1 mm or greater had a nal diag- Normal ECG
nosis of AMI.
Among patients in the ED who had normal
ECGs (ie, lacking Q waves; primary ST-segment
and T-wave abnormalities; and criteria for non-
Q waves
diagnostic abnormalities), 1% [34] to 6% [82] had
Q waves are diagnostic of MI, but what is the AMI. Among admitted patients who had normal
age of the Q wave? In the Multicenter Investiga- ECGs, 6% to 21% had AMI [12,78,8183]. Of pa-
tion of the Limitation of Infarct Size (MILIS) tients discharged home with a normal ECG, only
study of admitted patients in the CCU, isolated 1% had acute infarction [82]. Patients who have
new or presumed-new inferior or anterior Q waves a normal ECG and a suggestive clinical presenta-
were associated with acute infarction in 51% and tion still have a signicant risk for ACI, especially
77% of patients, respectively [63]. Other ndings if the ECG was obtained when the patient was
of the MILIS study that should be kept in mind pain free. On the other hand, a truly normal
are: 12% of healthy young men have inferior Q ECG in a patient unlikely to have acute ischemia
waves [76,78,79]; pathologic Q waves can be provides strong evidence against ACS [34].
from a previously unrecognized infarction and In their series, Pope and colleagues [2] found
can mask new same-territory ischemia; Q waves that patients who had normal ST-segment and T
alone do not identify ACSs and are rarely the waves and no Q waves more commonly had a nal
sole manifestation of AMI (6% in the MILIS diagnosis of non-ACS, yet 20% of these patients
study); and, nally, infarction can occur in the ab- had AMI and 37% had UAP at nal diagnosis.
sence of Q waves [80,81]. The investigators ED
study [2] showed that Q waves were more com-
monly associated with a nal diagnosis of ACS Prehospital 12-lead ECG
(25% ACS versus 11% non-ACS; P .001) and
The NHAAP Working Group on Evaluation
that 29% of patients who had Q waves present
of Technologies for Identifying ACSs [57] found
on their ECGs had a nal diagnosis of AMI.
that the diagnostic accuracy of prehospital ECG
for AMI and ACSs, as expected, is similar to
that of the standard 12-lead ECG, which is the
Nondiagnostic ECG patterns
standard of care in the management of patients
Nondiagnostic ST-segment and T-wave ab- suspected of having ACS (Tables 1 and 2). The ac-
normalities may be dened as follows: not having cumulation of evidence is substantial in the total
1 mm or greater (0.1 mV) ST-segment elevation or sample size and quality, and the data have been
depression in two contiguous leads, not having gathered from patient populations with few exclu-
new T-wave inversion in two contiguous leads, sion criteria. The evidence shows that obtaining
absence of signicant Q-waves (O1 mm deep and a prehospital ECG does not prolong time in the
0.3 s duration) in two contiguous leads, not eld or delay transport to the ED. In addition,
having second- or third-degree heart block, and prehospital ECG-guided thrombolytic therapy
not having a new conduction abnormality can be administered 45 minutes to 1 hour earlier
(eg, bundle branch block). These patterns are than hospital-based thrombolytics. Prehospital
the most dicult to interpret and can result in thrombolysis has a modest but signicant impact
Table 1
Summary of test performance studies of diagnostic technologies for acute cardiac ischemia in emergency departments
Condition Number of Population category Studies prevalence Sensitivityb Specicityb Diagnostic odds Overall quality
Technology studied studies (subjects) of studiesa range (%) (95% CI) (%) (95% CI) (%) ratiob (95% CI) of evidence
Prehospital 12-lead ECG ACI 5 (4311) I/II 4692 76 (5489) 88 (6796) 23 (6.385) B
AMI 10 (4481) I/II 1451 68 (5976) 97 (8992) 104 (48224) B
Continuous/serial ECG ACI 2 (1271) III/IV 440 2125c 9299c 3.845c C
AMI 1 (261) III/IV 11 39c 88c 4.9c B
Nonstandard lead ECG ACI 1 (52) IV 48 96c 41c 17c B
AMI 4 (897) IV 2265 5983c 7693c 1019c B
Exercise stress ECG ACI 2 (312) III 610 70100c 8293c 11320d B
CK (presentation) AMI 10 (2885) I/II/III 741 36 (2944) 88 (8093) 4.0 (2.66.2) C
CK (serial) AMI 2 (786) I 2643 6999c 6884c 12222c C
CK-MB (presentation) ACI 1 (1042) III 20 23c 96c 7.2c C
AMI 10 (2504) I/II/III 642 44 (3553) 96 (9497) 23 (1732) B
CK-MB (serial) ACI 1 (1042) III 20 31c 95c 8.5c C
AMI 7 (3381) I/II/III 153 87 (6795) 96 (9497) 171 (58505) B
Myoglobin (presentation) AMI 10 (1395) I/II/III 12413 49 (4157) 93 (8896) 13 (7.921) B

ACS IN THE ED
Myoglobin (serial) AMI 5 (831) I/II/III 2337 90 (7896) 90 (7896) 140 (66300) C
Troponin I (presentation) AMI 2 (874) II/III 639 2366c 8995c 5.714c A
Troponin I (serial) AMI 1 (773) III 6 100c 83c dd A
Troponin T (presentation) AMI 5 (1171) II/III 639 44 (3256) 92 (8895) 10 (5.918) B
Troponin T (serial) AMI 2 (1440) II/III 56 8093c 6590c 35120c A/C
P-selectin ACI 1 (263) II 33 35c 79c 2.0c B
AMI same study 8.4 45c 76c 2.6c
Rest echocardiography ACIe 2 (228) III 330 70 (4388) 87 (7294) 20 (948) C
AMI 3 (397) I/III 330 93 (8191) 66 (4383) 20 (762) B
Stress echocardiography AMI 1 (139) III 4 90c 89c 68c C
Sestamibi (rest) ACI 3 (702) III 917 81 (7487) 73 (5685) 18 (1129) B
AMI same studies 212 92 (7898) 67 (5279) 26 (6113)
ACI-TIPI ACI 4 (5496) I 1734 8695c,e 7892c,e 6169c,e A
Goldman chest pain protocol AMI 3 (5359) I (ACI 2730) 8891c 7074c 2023c A
1221
Algorithm/protocols No data from prospective studies
Computer-based decision aids AMI 6 (3606) I/II/III 742 5298c 5896c 4.49.04 A
Abbreviations: ACI-TIPI, acute cardiac ischemia time-insensitive predictive instrument; CK, creatine kinase; CK-MB, creatine kinase subunit.
a
Population categories: I: all patients who have symptoms suggestive of ACI; II: chest pain; III: chest pain with nondiagnostic ECG; IV: selected subpopulation.
b
Results from meta-analysis of several studies using random eects calculations unless otherwise indicated.
c
Point estimate from single study or a range of reported values; meta-analysis not performed.
d

433
Upper range cannot be estimated because of a study with 100% sensitivity.
e
ACI-TIPI is not intended to provide sensitivity and specicity. The results reported incorporate physicians triage decisions and are not reective of diagnostic test performance only.
434 POPE & SELKER

Table 2
Summary of clinical impact studies of diagnostic technologies for acute cardiac ischemia in emergency departments
Number
of Clinical Quality
Condition studies Population Prevalence outcomes Clinical of
Technologies studied (subjects) categorya (%) studied impact evidence
Prehospital ACI w10 studiesb I/II 46100 Time to A
12-lead ECG thrombolysis,
ejection
fraction,
mortality
AMI w8000 ptsb 1551
Continuous/ No study d d Not known d
serial ECG
Nonstandard No study d d Not known d
lead ECG
Exercise ACI 3 (272) III 06 Feasibility and Not known C
stress ECG safety
CK (single/serial) No study d d Not known d
CK-MB (single) No study d d Not known d
CK-MB (serial) ACI 1 (1042) III 20 Additional C
admissions or
discharges of
ACI and non-
ACI patients
Myoglobin No study d d Not known d
(single/serial)
Troponin I or T No study d d Not known d
Other biomarkers No study d d Not known d
Rest No study d d Not known d
echocardiography
Stress No study d d Not known d
echocardiography
Sestamibi No study d d Not known d
ACI-TIPI ACI 5 (14450) I 1759 CCU admission () A
rate,
inappropriate
discharge
Goldman chest AMI 1 (1921) III 6.6 Hospitalization () A
pain protocol rate,
length of stay,
estimated
costs
Algorithm/ ACI 2 (602) III 69 Length of stay, Not known B
protocols hospital
charges,
30-day and
150-day
mortality
Computer-based ACI 1 (977) III 48 (AMI 30) 30-day Not known A
decision aids mortality
Clinical impact scores range from low () to high ().
Quality of evidence scores range from low (C) to high (A).
Abbreviations: ACI-TIPI, acute cardiac ischemia time-insensitive predictive instrument; CK, creatine kinase; CK-
MB, creatine kinase subunit.
a
Population categories: I: all patients who have symptoms suggestive of ACI; II: chest pain; III: chest pain with non-
diagnostic ECG; IV: selected subpopulation.
b
Dierent outcomes analyzed involved dierent number of studies and patients.
ACS IN THE ED 435

on early mortality, with approximately 60 patients diagnostic performance is encouraging, it would


requiring prehospital treatment compared with be premature to make conclusions regarding this
hospital thrombolysis to save one additional life technology until additional high-quality studies
in the short-term. Short-term, benecial eects are conducted.
of thrombolysis on left ventricular ejection frac- There are also limited data on the clinical
tion have not been reported in randomized trials. impact of exercise testing for ACSs. Two studies
The long-term survival benets of prehospital [88,89] had no cardiac events and included very
thrombolysis remain uncertain. Although it has small sample sizes: 28 and 35, respectively. Adding
promise, the Working Group [57] believed that a third study [86], these investigations comprised
its best use would be in areas with long EMS only 272 subjects and are of low methodologic
transport times and perhaps in conjunction with quality; the clinical impact of this technology is
prehospital thrombolytic therapy. Its routine use unclear.
was not recommended.
Biochemical markers
Continuous ECG/Serial ECG
Creatine kinase, single and serial measurements
The Working Group [57] found that two stud-
ies evaluated the test performance of continuous/ There is a large amount of evidence on creatine
serial 12-lead ECG in the ED, but there was no kinase (CK) as a single test administered at
clinical impact study (see Tables 1 and 2). One presentation to patients in the ED (see Tables 1
study by Gibler and colleagues [84] included and 2) [57]. The evidence suggests that the sensi-
a large retrospective population of 1010 partici- tivity of a single CK reading for AMI is low
pating in a 9-hour protocol. The serial ECG con- (36%), and specicity is modest (88%). Limited
sisted of a 20-second interval between readings. evidence suggests that the sensitivity of the test de-
The second study [85] included patients from a vet- pends on the duration of the patients symptoms;
erans hospital in which two ECGs were taken 4 sensitivity increases with longer symptom dura-
hours apart. The prevalences of ACSs in these tion. Test performance across studies did not ap-
studies were very dierent (4% and 40%, respec- pear to vary by type of hospital, inclusion
tively) given the low-risk populations. The sensi- criteria, AMI prevalence, or test threshold.
tivity for ACSs was low (21% and 25%, Only two studies have evaluated serial CK
respectively) and the specicity was high (92% testing [90,91]. These studies used broad inclusion
and 99%, respectively). With the limitations and criteria but enrolled populations in which the
the varied source of data, a conclusion about the prevalence of AMI was moderate to high. Test
usefulness of this technology cannot be drawn. sensitivity was high (95%99%) in serial tests per-
formed over about 15 hours after presentation to
the ED (or from the onset of symptoms), but was
Nonstandard-lead ECG only modest (69%) in the one study that drew se-
The data on the diagnostic performance of rial samples for 4 hours. Test specicity was mod-
nonstandard-lead ECG from the four studies est in both studies (68% and 84%).
reported vary too much to draw any conclusion As a single test, CK is insensitive and only
[57]. The studies used 15, 18, 22, and 24 leads modestly specic for AMI. Serial testing appears
and were conducted with selected patients for to have higher sensitivity, although the specic-
admission (see Tables 1 and 2). The prevalence ity remains modest. However, the evidence is
is reective of this selective population: it ranged insucient to evaluate serial CK measurements
between 22% and 65% for AMI. There are no over a short time. Because high serum CK levels
clinical impact studies on non-standard ECGs. represent infarcted myocardium, CK has not been
evaluated for diagnosing ACSs in the ED. There
are no clinical impact studies for CK.
Exercise stress ECG
Creatine kinase subunit, single and serial
The data on the diagnostic performance of
measurements
exercise stress testing to detect ACSs in the ED are
limited to only two studies (see Tables 1 and 2) As is the case with CK, the total sample size
[86,87]. The overall data include a small sample and number of studies on a single CK subunit
size of a low-risk population. Although the (CK-MB) measurement at presentation to the ED
436 POPE & SELKER

are large (see Tables 1 and 2) [57]. The evidence for AMI. However, further research is necessary
suggests that the sensitivity of single CK-MB for to conclusively refute the possibility that the re-
AMI is low (47%), although specicity is high lease of cardiac troponins may also occur in the
(96%). Studies reported a broad range of sensitiv- setting of reversible myocyte injury resulting
ity for diagnosing AMI. Again, as is the case for from cellular ischemia [95]. Cardiac troponins of-
CK, limited evidence suggests that the sensitivity fer extremely high tissue specicity but do not dis-
of CK-MB depends on the duration of the pa- criminate between ischemic and nonischemic
tients symptoms; sensitivity increases with longer mechanisms of myocardial injury; thus, presently
symptom duration. In general, studies reported the clinician must assess whether a patients pre-
a narrow range (92%99%) of test specicity. senting symptoms are consistent with ACSs. In
Test performance across studies did not appear addition, renal failure does not appear to diminish
to vary by type of hospital, inclusion criteria, the prognostic value of troponins among patients
AMI prevalence, or test threshold. who have a high clinical probability of ACSs re-
The total sample size and number of studies of gardless of patients creatinine clearance [96].
serial tests for CK-MB in the ED setting are large. The evidence for the diagnostic performance of
Overall, serial testing has a modest sensitivity troponin T is substantial in diagnosing AMI but
(87%) and high specicity (96%) for AMI. rather limited in diagnosing ACSs (see Tables 1
However, test sensitivity is strongly related to and 2) [57]. Data for troponin I are limited, but
the timing of serial testing. All studies that its performance is similar to that of troponin T.
performed serial testing for at least 4 hours after The sensitivity of presentation troponin T for di-
presentation to the ED (or until at least 8 hours agnosing AMI in the ED is poor, but improves
after symptom onset) found test sensitivity to be substantially if serial measurements are obtained
greater than 90%. Conversely, all studies that for up to 6 hours after ED presentation. Most
performed serial testing to at most 3 hours found likely, the sensitivity is better for patients who
test sensitivity to be less than 90%. The pooled have had symptoms for longer periods of time.
sensitivity for serial testing to at least 4 hours is The specicity of troponin T for AMI is in the
96%; pooled sensitivity for serial testing until 3 range of 90%.
hours is only 81%. In general, test specicity was
in a narrow range across studies and was above
Myoglobin
90%. A recent report found that a 2-hour delta
CK-MB level (sensitivity 93%; specicity 94%) The diagnostic performance of myoglobin has
outperformed a myoglobin level in the early been well studied for diagnosing AMI, but not for
identication of AMI when troponins are used diagnosing ACSs (see Tables 1 and 2) [57]. The
as the criterion standard [92]. sensitivity of myoglobin for diagnosing AMI in
CK-MB as a single test is only modestly the ED is poor when a single initial measurement
sensitive and specic for AMI; however, serial is obtained, but sensitivity improves greatly if
testing performed over 4 to 9 hours is highly a second measurement is obtained 2 to 4 hours af-
sensitive and highly specic. Because serum CK- ter the rst one. However, the sensitivity for pa-
MB levels represent infarcted myocardium, CK- tients only recently symptomatic is poor, and
MB has not been tested for diagnosing ACS in a second measurement in 2 to 4 hours may still
the ED. There are no clinical impact studies for not be suciently sensitive to be useful. Specicity
CK-MB. is very good, but not excellent, depending on the
extent to which other reasons for elevated myo-
globin are excluded a priori. A doubling of myo-
Troponin T and troponin I
globin levels as soon as 1 to 2 hours after the
The cardiac troponins have expanded the initial measurement is almost perfectly sensitive
spectrum of detectable myocardial injury and for AMI.
enhanced the clinicians ability to identify patients The evidence suggests that a normal myoglo-
who have ACSs and are at higher risk for death or bin value 2 hours after presentation may be used
recurrent ischemic events even with low-level safely to rule out AMI. A doubling of myoglobin
elevation of cardiac troponin T or I [93,94]. The as early as 1 to 2 hours after the baseline
notion that any reliably detected troponin eleva- measurement establishes a diagnosis of AMI. A
tion results from myocyte necrosis serves as the small study [97] suggests that normal myoglobin
basis for the recent revision of diagnostic criteria and CK-MB values 2 hours after presentation
ACS IN THE ED 437

completely rule out AMI. The incremental value found that a low level of red-cell glutathione per-
of CK-MB compared with myoglobin alone can- oxidase 1 was independently associated with an
not be evaluated given the small sample sizes. In increased risk for cardiovascular events. Experi-
a much larger study, Kontos and colleagues [98] mental data suggest that PlGF, a member of the
found no advantage for myoglobin over baseline vascular endothelial growth factor family, acts
and 3-hour CK-MB values. In a subanalysis of as a primary inammatory instigator of athero-
patients who had nonST-elevation ACS, serum sclerotic plaque instability and thus may be useful
myoglobin above the MI detection thresholds as a risk-predicting biomarker in patients who
was associated with increased 6-month mortality have ACSs. In a subanalysis of data from the
and thus may be useful for early risk stratication. CAPTURE Trial, Heeschen and colleagues [110]
found PlGF to be an independent marker of ad-
verse outcome in patients who have suspected
Biomarkers of neurohumoral activation and
ACSs. Myeloperoxidase, an abundant leukocyte
inammation
enzyme, is elevated in culprit lesions that have s-
In addition to markers of myonecrosis, sured or ruptured in patients who had sudden
markers of neurohumoral activation and inam- death from cardiac causes. Numerous lines of ev-
mation may provide important prognostic and idence suggest mechanistic links between myelo-
possibly diagnostic information in ACS. Studies peroxidase and inammation and cardiovascular
on P-selectin [99,100], malondialdehyde-modied disease. In a prospective assessment of patients
low-density lipoprotein, high-sensitivity C-reac- in the ED who had suspected ACSs, Brennan
tive protein (hsCRP) [101104], B-type natriuretic and colleagues [108] found that a single initial
peptide [105], pregnancy-associated plasma pro- measurement of plasma myeloperoxidase inde-
tein A (PAPP-A) [106], serum amyloid A (SAA) pendently predicted the early risk for MI and
[107], soluble CD-40, myeloperoxidase [108], glu- the risk for major adverse cardiac events in the en-
tathione peroxidase 1 [109], placental growth factor suing 30-day and 6-month periods.
(PlGF) [110], matrix metalloproteinases [111], and In the future, tests for neurohormonal activa-
monocyte chemoattractant protein-1 (MCP-1) tion (B-type natriuretic peptide) and inammation
[112] are just beginning to appear. (C-reactive protein, PAPP-A, MCP-1, SAA) may
P-selectin in platelets and endothelial cells augment the ability to identify patients who have
mediates adhesion interaction with leukocytes to ACSs and are at risk for adverse events. The use
form thrombi. There is only one ED study of of these markers alone or in combination (a multi-
P-selectin that reported low sensitivity and low marker approach) could potentially augment the
specicity for AMI. Traditionally, hsCRP has ability to reserve the most expensive and aggres-
been thought of as a bystander marker of vascular sive therapies for patients who have the highest
inammation, without playing a direct role in the risk [113].
cardiovascular disease. More recently, accumulat-
ing evidence suggests that hsCRP may have direct
proinammatory eects, which is associated with Cardiac imaging
all stages of atherosclerosis, including plaque
Echocardiography
destabilization [104,107]. Laterza and colleagues
[106] studied ED patients who had suspected The total sample size and the number of
ACSs and found pregnancy-associated protein studies evaluating echocardiography for the di-
A, a potential proatherosclerotic metalloprotein- agnosis of ACSs are small (see Tables 1 and 2)
ase, to be only a modest predictor of adverse [57]. Limited evidence suggests that resting echo-
events, with a sensitivity of 67% and a specicity cardiography has high sensitivity (93%), although
of 51%. It is thought that cellular antioxidant en- only modest specicity (66%) for AMI. The avail-
zymes such as glutathione peroxidase 1 and super- ability of previous echocardiograms for compari-
oxide dismutase have a central role in the control son may improve the specicity [114]. But even
of reactive oxygen species. In vitro data and stud- if this improved specicity were veried with ad-
ies in animal models suggest that these enzymes ditional studies, the need for previous echocar-
may protect against atherosclerosis, but little is diography would limit its applicability in the
known about their relevance to human disease. general ED setting. In addition, the data pertain
In a prospective study of patients who have sus- mostly to patients who have normal or non-
pected ACSs, Blankenberg and colleagues [109] diagnostic ECGs. The data for stress dobutamine
438 POPE & SELKER

echocardiography are even more limited. One Computer-based decision aids


study suggests that it may be the next diagnostic
Acute cardiac ischemia time-insensitive predictive
step for patients who have a negative resting echo-
instrument
cardiogram, normal ECG, and normal enzyme
levels. There is no clinical impact study for this The acute cardiac ischemia time-insensitive
technology. predictive instrument (ACI-TIPI) [26] computes
a 0% to 100% probability that a given patient
has ACS (either acute MI or unstable angina pec-
Technetium 99m sestamibi myocardial perfusion
toris) (see Tables 1 and 2). Applicable to any pa-
imaging
tient in the ED presenting with any symptom
Data on the diagnostic accuracy of resting suggestive of ACS, it is based on a logistic regres-
technetium (Tc) 99m sestamibi imaging in the ED sion equation that uses presenting symptoms and
are limited (see Tables 1 and 2). The test has been ECG variables. Originally in hand-held calculator
used in selected patient populations that generally form, it is now incorporated into conventional
have a low-to-moderate risk for an ACS, no his- electrocardiographs so that the patients ACI-
tory of MI, and a presenting ECG nondiagnostic TIPI probability is printed with the standard
for an ACS. Thus, the generalizability of the cur- ECG header text. In large controlled interven-
rent evidence is limited, and the test should be tional trials in a wide range of hospitals, its use
reserved for these circumscribed populations. In by ED physicians has been shown to reduce un-
these patients, the test has excellent sensitivity necessary admissions of patients who do not
for AMI and very good, but not perfect, sensitiv- have ACSs and patients who have stable angina,
ity for coronary disease in general. Specicity is while not reducing appropriate hospitalization
modest for AMI, and although it may be a little for patients who have ACSs. It has also been
better for ACSs, it is still far from excellent. In shown to help the triage speed and accuracy of
a prospective ED trial, Udelson and colleagues less-trained and less-supervised residents. Al-
[115] showed that resting Tc 99m sestamibi imag- though this decision aid is a widely available soft-
ing improved ED triage decision making for pa- ware option oered by all the major ECG
tients who had symptoms suggesting ACSs but machines manufacturers in the United States, it
no obvious abnormalities on the initial ECG by has not been widely used in clinical practice. The
reducing unnecessary admissions without reduc- greater dissemination and use of ACI-TIPI could
ing appropriate admissions. result in signicant positive impact on the triage
of patients who have ACSs in the ED.

Myocardial performance index


Thrombolytic predictive instrument
Central aortic pressure
Correctly diagnosing STEMI for prompt use of
One index of myocardial contractility is the coronary reperfusion therapy can be lifesaving. In
rate of increase of intraventricular pressure during EDs this can be dicult, especially for less obvious
isovolumetric contraction (left ventricular dP/dt; candidates [121123]. Eorts by physician leaders,
arterial dP/dt). dP/dt represents the rate of change the NHAAP, health care organizations and moni-
of pressure during ejection [116119]. It has been toring entities, and pharmaceutical companies
shown previously that cardiac contractility and [123132], have helped increase use and prompt-
dP/dt decrease during acute cardiac ischemia ness of reperfusion therapy [133]. Further im-
[116,118]. Preliminary work by Gorenberg and provement is needed [133,134], however,
Marmor [120] suggests that a noninvasive device particularly for other than anterior STEMI, the
that quanties central aortic pressure through category for which thrombolytic therapy rst was
brachial artery sensors may be a technique that recognized as eective [121,122] and for women
could be used in selected patient populations who have received less coronary reperfusion ther-
that have a low-to-moderate risk for ACSs, no apy than men [134,135]. Also, a need remains for
history of MI, and a presenting ECG nondiagnos- ways to support prompt and accurate coronary re-
tic for ACSs. However, the generalizability of the perfusion therapy decisions in hospitals and pre-
current evidence is limited because of study size hospital EMS settings where consultation with
and design. o-site physicians is required.
ACS IN THE ED 439

Selker and colleagues [136] developed the who have developed inferior STEMI, 18% to
thrombolytic predictive instrument (TPI), a collec- 22% for women, and 30% to 34% for patients
tion of ve component predictive instruments de- who have an o-site physician [137]. Although the
signed to accurately assess the patient-specic TPIs eect was minimal on patients who had high
likely benets and risks from the use of thrombo- baseline coronary reperfusion therapy rates, such
lytic therapy for STEMI. The TPI helps clinicians as men who present with anterior STEMIs, for
identify patients for coronary reperfusion therapy the groups needing the most improvement in their
based on their probabilities of benets and com- rates of recognition for coronary reperfusion,
plications and facilitates earliest possible use of women and those with less obvious STEMIs,
reperfusion therapy [123,136]. In conventional and where involved physicians were o-site, the
computerized ECGs when signicant ST segment TPI increased recognition of STEMI and use
elevation of STEMI is detected, the TPI predic- and timeliness of coronary reperfusion therapy.
tions are automatically computed and printed on It is hoped that as TPI-capable ECGs become
the ECG header: probabilities for acute (30-day) more widely available in ED and EMS settings,
mortality either treated with thrombolytic therapy its use will facilitate the accuracy and speed of rec-
or untreated; 1-year mortality rates treated with ognition and treatment of these patients.
thrombolytic therapy; cardiac arrest treated with
thrombolytic therapy and untreated; thrombolytic Goldman Chest Pain Protocol
therapy-related stroke and major bleeding requir-
The Goldman Chest Pain Protocol [7] is based
ing transfusion.
on a computer-derived model using recursive-
To test whether the ECG-based TPI improves
partitioning analysis to predict MI in patients
ED selection of patients for coronary reperfusion
who have chest pain (see Tables 1 and 2). It has
therapy and promptness of treatment, the authors
good sensitivity (about 90%) for AMI, but was
ran a 22-month randomized controlled clinical
not developed to also detect UAP. In a clinical im-
eectiveness trial of its impact on the use of
pact study of low-intensity, non-intrusive inter-
thrombolytic therapy and overall coronary reper-
vention performed in the ED of a teaching
fusion therapy. The trial ran in EDs at 28 urban,
hospital [16], no dierences in hospitalization
suburban, and rural hospitals across the United
rate, length of stay, or estimated costs were dem-
States, from major cardiac centers to small
onstrated between the experimental group that
community hospitals. Study endpoints were per-
used the protocol and the control group. Gold-
centages of patients receiving (a) thrombolytic
man and colleagues [35] eventually switched to
therapy; (b) thrombolytic therapy within 1 hour of
predicting the need for intensive or other levels
ED presentation; and (c) all coronary reperfusion
of care. They found that clinicians who had higher
therapy, either by thrombolytic therapy or PTCA.
levels of training had a higher sensitivity for de-
At participating hospitals, software generating
tecting AMI, but at the expense of decreased spec-
TPI predictions was installed on their conven-
icity [138]. Reilly and colleagues [139] developed
tional computerized ECGs. When a signicant ST
a consensus to adapt the Goldman prediction rule
elevation characteristic of STEMI was automati-
and found a favorable impact on physicians hos-
cally detected, the ECG randomly assigned the
pital triage decisions solely by dierent triage de-
patient to the control or intervention group. If
cisions for very low-risk patients. Unfortunately,
assigned to the intervention group, the ECG
any algorithm that incorporates only clinical ele-
automatically prompted the user to enter infor-
ments and the ECG ndings at presentation is
mation needed to compute the TPI predictions:
likely to be suboptimal because of the substantial
age, sex, history of hypertension or of diabetes,
proportion of patients who present with atypical
blood pressure, and time since ischemic symptom
symptoms or with no or minimal ECG
onset. The remaining variables, based on ECG
abnormalities.
waveform measurements, were automatically ac-
quired by the ECG. Then the ECG was printed
Other computer-based decision aids
with TPI predictions on its header.
Among the 1197 patients who developed Several investigators have reported various
STEMI, the trial showed that the TPI increased computer-based decision aids to diagnose AMI
use of thrombolytic therapy, use of thrombolytic (see Tables 1 and 2). The articial neural network
therapy within 1 hour, and use of overall coronary by Baxt and Skora [140] had high sensitivity and
reperfusion therapy by 11% to 12% for patients high specicity for AMI in a prospective study,
440 POPE & SELKER

but the clinical impact has not been demonstrated. care resources [156158], partly because the patho-
A predictive model with automatic ECG interpre- physiology of ischemia in women is incompletely
tation has been shown to increase the use of bri- understood and gender-specic diagnostic and
nolytic therapy for acute ST-segment elevation treatment strategies are underdeveloped [159].
MI, especially in historically undertreated pa- Gender dierences with regard to ischemic heart
tients, such as those who have inferior AMI and disease appear to exist in several areas, including
women who have AMI [141]. established and novel risk factors; the metabolic
syndrome; the physiology of endogenous repro-
ductive hormones; the role of endothelial dys-
Women
function in producing obstructive macrovascular
It is becoming apparent that one of the major CAD, myocardial ischemia, chronic chest pain
contributing factors for the misdiagnosis of ische- syndromes, and ACS; genetic factors; proteomics;
mic heart disease in women is pathophysiologic the menstrual cycle and reproductive status; pain
dierences that appear to exist between men and threshold/perception; neurohumoral control; and
women who have these syndromes and between behavioral/psychosocial factors.
the various ischemic syndromes themselves. Ob- The diagnostic evaluation of women who have
structive CAD continues to be a major public suspected ACSs continues to be a major chal-
health problem in women and has represented the lenge. Knowing whether gender inuences the
leading cause of death and disability for more likelihood that a given patient in the ED is having
than a decade [1,142,143]. In the United States ACSs, and whether any specic presenting clinical
alone, more than a quarter of a million women features are dierentially associated with ACSs in
die of coronary heart disease (CHD) each year, women compared with men, can aid clinicians in
translating into 1 death every 2 minutes [1]. In- the accurate diagnosis of ACSs. The incidence of
creased awareness of these statistics and a recent AMI in the general population has been shown to
focus on womens health issues in general have re- be higher in men than women [160163], but until
sulted in renewed interest in trying to understand recently it has not been clear whether this gender
important gender dierences in patients who have dierence holds among symptomatic patients
ACSs. Although women have a lower prevalence who come to the ED. In addition, several stud-
of obstructive CAD compared with men who ies have looked at gender dierences in the presen-
have similar symptoms [144146], women have tation of patients who have AMI [157,164167].
a higher frequency of angina/chest pain than In a retrospective analysis of patients who have
men partly because of the higher prevalence of conrmed AMI, women had higher rates of atyp-
the less common causes of ischemia, such as vaso- ical presentations, such as abdominal pain, parox-
spastic angina and microvascular endothelial ysmal dyspnea, or congestive heart failure
dysfunction; transient left ventricular apical bal- [46,160,168170]. In a group of patients in the
looning syndrome (takotsubo cardiomyopathy) ED who had typical presentations such as chest
[147]; and syndromes of nonischemic chest pain, pain, the prevalence of AMI was lower in women
such as mitral valve prolapse [148]. Add to this [34,171]. However, in another study of patients in
the fact that young women who have obstructive the ED who had chest pain, when adjustments
CAD experience a signicantly worse outcome were made for other presenting clinical features
with regard to prognosis after MI compared (specically ECG), the gender dierence was no
with men [149], and that older women who have longer signicant [166]. From these results it is dif-
obstructive CAD often have greater comorbidities cult to assess whether the gender-specic dier-
that inuence their outcome adversely after AMI ences in AMI prevalence among symptomatic
or myocardial revascularization than do men patients in the ED were the result of gender-
[150153]. Furthermore, women presenting with specic biology or limitations in a particular
ACSs are less likely to receive eective acute diag- studys patient selection.
nostic and treatment strategies than men One reason for such challenges is that chest
[3,154,155]. When women develop obstructive pain in women is neither sensitive nor specic in
CAD, they have a greater functional expression predicting the presence of underlying CHD. The
of their disease and disability compared with highest sensitivity is found in women presenting
men. Most women who do not have obstructive with symptoms of typical angina pectoris, whereas
CAD at angiography continue to have symptom- the highest specicity is found in women present-
related disability and consume considerable health ing with nonspecic symptoms. In fact, although
ACS IN THE ED 441

women who have ACSs may present with symp- much more severe prognostic implications than in
tom patterns that dier from men (ie, atypical chronic stable angina, whereas although angina
symptoms), for most women, typical symptoms in patients who have normal coronary angiograms
are the strongest symptom predictors of ACSs is not associated with increased short-term risk of
[172174]. A further challenge is gender dierences infarction or sudden death, these patients may be
in reporting pain, including but not limited to crippled by pain. Inconsistent response to nitrates
chest pain. Gender dierences in endogenous and antianginal drugs indicates the need for re-
pain-modulating systems may contribute to dier- search on the various potential causes of coronary
ences in pain perception. Regarding biomarkers, vascular dysfunction that burden women dispro-
there is a suggestion that women who have ob- portionately, to facilitate development of rational
structive CHD may have a dierent pattern of forms of therapy.
presenting biomarkers compared with men. In
a TACTICS-TIMI 18 subanalysis examining pa-
Race
tients who have UA/NSTEMI, there was a dier-
ent pattern of presenting biomarkers between the Blacks have high levels of risk factors for
sexes: men were more likely to have elevated CAD, but how this nding inuences diagnosis
CK-MB and troponins, whereas women were in patients presenting to the ED with symptoms
more likely to have elevated C-reactive protein suggesting ACSs is not well understood [180,181].
and brain natriuretic peptide, suggesting that Studies that have included only patients who have
a multimarker approach to diagnosis may improve chest pain and not other symptoms suggestive of
triage in the ED. Lastly, the Womens Ischemic ACSs have found no signicant dierences in pre-
Syndrome Evaluation Study Group and others sentation, natural history, or nal diagnosis of
have suggested that those women who have chest AMI between black and white patients [182].
pain without ow-limiting lesions by angiography Evaluating chest pain and establishing the diagno-
may have associated microvascular endothelial sis of CAD in blacks is often dicult given the
dysfunction (cardiac syndrome X) [175] and im- presence of excess hypertension and left ventricu-
paired coronary ow reserve [176,177]. Prelimi- lar hypertrophy and the increased occurrence of
nary data suggest that coronary microvascular out-of-hospital cardiac arrest in blacks [183
dysfunction is associated with an increased rate 186]. Furthermore, the paradoxical nding of se-
of hospitalization for chest pain, poor quality of vere chest pain without signicant angiographic
life, and ongoing health care costs. Newer technol- CAD that can be seen in 30% to 40% of men
ogies, such as MR spectroscopy and gadolinium [187] complicates diagnosis and treatment of
cardiac MRI, may allow for the identication of blacks who have symptoms suggestive of ACSs
abnormalities in vascular function and structure [180,183]. In subanalysis of the ACI-TIPI Trial
not identiable by coronary angiogram that can data, Maynard and colleagues [188] found that
cause or contribute to development of myocardial black patients were 8 to 10 years younger and
ischemia, and may aid in the initial risk assessment that a higher percentage were women than was
of UA/NSTEMI, especially in women [178]. the case among white patients, which may partially
Finally, numerous studies have found that explain why physicians might be less inclined to
women have poorer outcomes than men after suspect the presence of ACSs in black patients. Fi-
a diagnosis of ACSs. Explanations have included nally, Pope and colleagues [3] found that among
gender dierences in pathophysiology and re- patients who have ACSs, the adjusted risk for be-
sponse to treatment; prehospital delays in symp- ing sent home was more than two times as high
tom recognition and action; and gender dierences among nonwhite as whites. Among those who
regarding evaluation and treatment in emergency had AMI, the risk was more than four times as
medical services. Recent reports of similar or better high among nonwhites as whites. In this study,
outcomes in women who have ACSs compared 5.8% of the black patients who had AMI were
with men suggest that pathophysiologic dierences not hospitalized, as compared with 1.2% of the
can be overcome with early, aggressive therapy white patients who had infarction.
[153,179]. The diagnosis of ischemia inuences
prognosis dierently in various clinical ischemic
Outcomes
syndromes. In unstable angina the detection of is-
chemia indicates the likelihood of persistence or re- Each year in the United States, over 6 million
currence of plaque instability and hence carries patients who have chest pain or other symptoms
442 POPE & SELKER

suggesting ACSs (eg, IMIR Study inclusion colleagues [6] in 1979 to 1981 (35% AMI, 65%
symptoms) [23] present to EDs [4]. These patients UAP). Finally, in an analysis of the ACI-TIPI
can have various clinical outcomes ranging from Trial data for failure to make the diagnosis of
discharge home to hospital admission after an ACS, the authors found that the missed diag-
thrombolytic therapy or percutaneous coronary nosis rate for ACSs was 2.2% (2.1% for AMI,
intervention. Table 3 shows the nal diagnosis 2.3% for UAP) [3].
for the ACI-TIPI Trial [2] control subjects by Recent approaches to patient safety emphasize
ED triage disposition. These data were employed systems thinking rather than focusing on in-
to develop a owchart (Fig. 1) to represent the di- dividual cognitive mistakes. The goal is to prevent
agnoses and triage dispositions of patients in the human errors, which are commonly made by
ED presenting with chest pain or other symptoms competent clinicians [189,190]. Suggested ap-
suggestive of ACSs. Although these data are not proaches include diagnostic protocols and path-
recent and the triage and intervention options ways, decision aids, novel approaches to stang,
have expanded over the last decade, the propor- chest-pain units, and other systems changes.
tions have remained remarkable similar.
The owchart (see Fig. 1) demonstrates that of
Missed diagnosis of ACS
all such patients, only 23% of patients (hospital
range 12%34%) had ACSs at nal diagnosis, Several factors, listed in Box 1, have been asso-
of which 94% were hospitalized and 6% were ciated with inappropriate discharge of patients
sent home. Conversely, 77% did not have ACSs who have symptoms suggestive of ACSs from
at nal diagnosis, of which 59% were hospitalized EDs and other acute care settings. These factors
and 41% were sent home. In the ACSs group of include younger age (!55 years), female gender,
patients, 36% of patients had AMI and 64% nonwhite ethnicity, atypical symptoms, no previ-
had UAP, which represented 8% and 15%, re- ous MI, normal or nondiagnostic ECG ndings,
spectively, of the overall group. In the AMI misinterpretation of the ECG (failure to recognize
group, 97% were hospitalized and 3% were sent minimal ST-segment elevation), and failure to ap-
home; in the UAP group, 92% were hospitalized preciate the dynamic nature of anginal symptoms,
and 8% were sent home. obtain multiple ECGs tracings over time, obtain
The authors work with Pozen and colleagues previous ECG tracings for comparison, appreciate
[6] from 1979 to 1981 at the same hospitals as the the dynamic nature of anginal symptoms, and
present report demonstrated a 7% ED discharge appreciate nonobstructive cardiac ischemic
rate for patients who had a nal diagnosis of syndromes [3].
ACSs; McCarthy and colleagues [18] found that Although women tend to develop obstructive
2% of these subjects had AMI at nal diagnosis. CAD 10 years later than men, Pope and col-
In the mid-1980s, Lee and colleagues [9] reported leagues [3] found that women under the age of
a 4% AMI discharge rate. The authors study 55 years were at the highest risk for not being
found a 6% discharge rate for ACSs and a 3% hospitalized with ACSs. Physicians must try not
AMI discharge rate, demonstrating stability of to pigeonhole patients into set categories, and
these gures over the decade. The proportions of remember that low risk for ACSs is not zero
AMI and UAP in the authors present study risk. Furthermore, because about 50% of women
(36% AMI, 64% UAP) were essentially identical and a small percentage of men who have symp-
to those from their work with Pozen and toms suggestive of ACSs have nonobstructive syn-
dromes, it is important that this subset of patients
not be overlooked as is has been previously, and
Table 3 that these patients are referred to cardiologists
Final diagnosis (%) for acute cardiac ischemia time-
for denitive diagnosis and therapy. In addition,
insensitive predictive instrument trial control subjects
by emergency department triage disposition
these researchers found that the risk for being
sent home was twice as high for nonwhites and
Triage AMI UAP non-ACS whites, citing younger age and higher percentage
disposition (n 496) (n 898) (n 4557)
of women as possible reasons for lowering clinical
Home 3 8 41 suspicions about ACSs. Although most patients,
Ward 1 2 6 including women who have ACSs, present in a typ-
Telemetry 31 61 43 ical fashion, many can present in an atypical fash-
CCU 66 29 10
ion with complaints of shortness of breath [191].
ACS IN THE ED 443

Fig. 1. Flowchart illustrating diagnoses and triage dispositions of patients presenting to the ED with chest pain or other
symptoms suggesting ACI. *Percentage of ED patients in the control group who have chest pain or symptoms consistent
with ACI.

Because about 25% of patients who are mistak- suggested that as many as 3% of patients who
enly discharged are sent home after an error in have suspected noncardiac chest pain will go on
interpretation in their ECG [192], improving the to experience a adverse cardiac event within 30
analysis and interpretation of ECGs could improve days of that assessment [193], and suggest that fea-
decision making. Finally, some investigators have tures such as male gender, hypercholesteremia, ad-
vanced age, CHD, heart failure, features that are
not typically associated with noncardiac chest pain,
Box 1. Factors contributing to the and a higher calculated ACI-TIPI score. However,
misdiagnosis of acute coronary safe noncardiac chest pain criteria have yet to be
syndromes derived and validated. Clinicians must at least con-
sider the possibility of the diagnosis of ACSs in pa-
 Younger age (<55 years of age) tients who have these factors, especially if a less
 Female sex serious cause for the patients symptoms is not
 Nonwhite ethnicity readily apparent.
 Atypical symptoms
 No previous MI
 Normal or nondiagnostic ECG findings Summary
 Misinterpretation of the initial ECG
 Failure to obtain multiple ECGs and Failure to diagnose patients who have ACSs,
obtain previous tracings either AMI or UAP, who present to the ED
 Failure to appreciate the dynamic remains a serious public health issue. Better un-
nature of anginal symptoms derstanding of the pathophysiology of CAD has
 Failure to appreciate nonobstructive allowed the adoption of a unifying hypothesis for
cardiac ischemic syndromes the cause of ACSs: the conversion of a stable
atherosclerotic lesion to a plaque rupture with
444 POPE & SELKER

thrombosis. Thus, physicians have come to ap- [4] Van der Does E, Lubson J, Pool J, et al. Acute cor-
preciate UAP and AMI as parts of a continuum onary events in a general practice: objectives and
of ACSs. However, for as many as 50% of women design of the Imminent Myocardial Infarction Rot-
who have a suspected ACSs, other mechanisms of terdam Study. Heart Bull 1976;7:91.
[5] McCaig L. National ambulatory care survey: 1992
illness are at work, including vasospastic angina,
emergency department summary. Adv Data 1994;
microvascular endothelial dysfunction, transient 245:112.
left ventricular apical ballooning syndrome, and [6] Pozen M, DAgostino R, Selker H, et al. A predic-
mitral value prolapse. The accurate identication tive instrument to improve coronary care unit ad-
of these syndromes in the ED or other acute care mission practices in acute ischemic heart disease:
settings challenges the skill of even the most a prospective multicenter clinical trial. N Engl J
seasoned clinician. Clinical features, patient de- Med 1984;310:12738.
mographics, physical examination ndings, and [7] Goldman L, Weinberg M. A computer-derived
a rm understanding the strengths and limitations protocol to aid in the diagnosis of emergency
of the standard ECG can greatly assist with the room patients with acute chest pain. N Engl J
Med 1982;307:58896.
identication of patients who have ACSs. Bio-
[8] Schor S, Behar S, Modan B, et al. Disposition of
chemical markers, cardiac imaging, myocardial presumed coronary patients from an emergency
performance indices, and computer-based deci- room; a follow-up study. JAMA 1976;236:9413.
sion aids remain underevaluated adjuncts to, but [9] Lee T, Rouan G, Weisberg M, et al. Clinical char-
not substitutes for, clinicians judgment regarding acteristics and natural history of patients with acute
diagnosis and triage of patients who have symp- myocardial infarction sent home from the emer-
toms suggestive of ACSs. Several factors have gency room. Am J Cardiol 1987;60:21924.
been associated with inappropriate discharge of [10] Pozen M, DAgostino R, Mitchell J, et al. The use-
patients who have symptoms suggestive of ACSs fulness of a predictive instrument to reduce inap-
from EDs and other acute care settings, including propriate admissions to the coronary care unit.
Ann Intern Med 1980;92:23842.
younger age (!55 years of age), female gender,
[11] Selker H, Pozen M, DAgostino R. Optimal identi-
nonwhite ethnicity, atypical symptoms, no pre- cation of the patient with acute myocardial ische-
vious MI, normal or nondiagnostic ECG ndings, mia in the emergency room. In: Calif R, Wagner G,
misinterpretation of the ECG, and failure to editors. Acute coronary care: principles and prac-
appreciate the dynamic nature of anginal symp- tice. Boston: Martinus Nijho; 1985. p. 28998.
toms. However, safe noncardiac chest pain criteria [12] Bloom B, Peterson O. End results, costs, and pro-
have yet to be derived and validated. Despite 2 ductivity of coronary care units. N Engl J Med
decades of focused research and without compel- 1973;288:728.
ling evidence for a noncardiac cause, there re- [13] Eisenberg J, Horowitz L, Busch R, et al. Diagnosis
mains no single way to discriminate perfectly of acute myocardial infarction in the emergency
room: a prospective assessment of clinical decision
between those who should be admitted to exclude
making and usefulness of immediate cardiac en-
ACSs and those who could be managed safely zyme determination. J Community Health 1979;4:
without admission. For the continuum of patient 1908.
risk for ACSs, there is a matched continuum of [14] Fuchs R, Scheidt S. Improved criteria for admission
triage options for each patient, from short-stay to coronary care units. JAMA 1981;246:203741.
chest pain evaluation units to multiday admis- [15] Tierney W, Roth B, Psaty B, et al. Predictors of
sions for more intensive evaluations. myocardial infarction in emergency room patients.
Crit Care Med 1985;13:52631.
References [16] Goldman L, Cook E, Brand D, et al. A computer
protocol to predict myocardial infarction in emer-
[1] American Heart Association. Heart Disease and gency department patients with chest pain. N
Stroke Statistics2005 Update. Dallas, TX: Ameri- Engl J Med 1988;318:707803.
can Heart Association; 2004. [17] Cannon CP. Management of coronary syndromes.
[2] Pope J, Ruthazer R, Beshansky J, et al. Clinical fea- In: Cannon CP, editor. Contemporary cardiology.
tures of emergency department patients presenting Totowa (NJ): Humana Press; 1999. Chapter 1, p. 8.
with symptoms of acute cardiac ischemia: a multi- [18] McCarthy B, Beshansky J, DAgostino R, et al.
center study. J Thromb Thrombolysis 1998;6: Missed diagnoses of acute myocardial infarction
6374. in the emergency department: results from a multi-
[3] Pope J, Aufderheide T, Ruthazer R, et al. Missed di- center study. Ann Emerg Med 1993;22:57982.
agnoses of acute cardiac ischemia in the emergency [19] McCarthy B, Wong J, Selker H. Detecting acute
department. N Engl J Med 2000;342:116370. cardiac ischemia in the emergency department: a
ACS IN THE ED 445

review of the literature. J Gen Intern Med 1990;5: [34] Lee T, Cook E, Weisberg M, et al. Acute chest pain
36573. in the emergency room: identication and examina-
[20] Selker HP, Zalenski RJ, Antman EM, et al. An tion of low-risk patients. Arch Intern Med 1985;
evaluation of technologies for identifying acute car- 145:659.
diac ischemia in the emergency department. Ann [35] Goldman L, Cook E, Johnson P, et al. Prediction of
Emerg Med 1997;29:187. the need for intensive care in patients who come to
[21] Agency for Healthcare Research and Quality the emergency department with acute chest pain.
(AHRQ) Evaluation of technologies for identifying N Engl J Med 1996;334:1498504.
acute cardiac ischemia in emergency departments. [36] Selker H, Grith J, DAgostino R. A time-insensi-
Technology assessment no. 26, prepared by Joseph tive predictive instrument for acute myocardial in-
Lau, MD. AHRQ Pub no. 01-E006, May 2001. farction mortality: a multicenter study. Med Care
1035. 1991;29:1196211.
[22] Rifkin R, Hood WJ. Bayesian analysis of electro- [37] Sawe U. Pain in acute myocardial infarction. A
cardiographic exercise stress testing. N Engl J study of 137 patients in a coronary care unit.
Med 1979;297:6816. Acta Med Scand 1971;190:7981.
[23] Uretsky B, Farquhar D, Berezin A, et al. Symptom- [38] Sawe U. Early diagnosis of acute myocardial in-
atic myocardial infarction without chest pain: prev- farction with special reference to the diagnosis of
alence and clinical course. Am J Cardiol 1977;40: the intermediate coronary syndrome: a clinical
498503. study. Acta Med Scand 1972;520(Suppl):176.
[24] Kinlen L. Incidence and presentation of myocar- [39] Levene D. Chest pain-prophet of doom or nagging
dial infarction in an English community. Br Heart necrosis? Acta Med Scand 1981;644(Suppl):113.
J 1973;35:61622. [40] Sievers J. Myocardial infarction. Clinical features
[25] Marglois J, Kannel W, Feinlieb M, et al. Clinical and outcome in three thousand thirty-six cases.
features of unrecognized myocardial infarction- Acta Med Scand 1964;406(Suppl 406):1120.
silent and symptomatic. Am J Cardiol 1973;32: [41] Areskog M, Tibbling L, Wranne B. Oesophageal
16. dysfunction in non-infarction coronary care unit
[26] Selker H, Beshansky J, Grith J, et al. Use of the patients. Acta Med Scand 1979;205:27982.
Acute Cardiac Ischemia Time-Insensitive Predic- [42] Alonzo A, Simon A, Feilieb M. Prodromata of
tive Instrument (ACI-TIPI) to assist with triage of myocardial infarction and sudden death. Circula-
patients with chest pain or other symptoms sugges- tion 1975;52:105662.
tive of acute cardiac ischemia. Ann Intern Med [43] Nattel S, Warnica J, Ogilivie R. Indications for
1998;129:84555. admission to a coronary care unit in patients
[27] Russell R. Unstable angina pectoris: National Co- with unstable angina. Can Med Assoc J 1980;122:
operative Study Group to compare medical and 1804.
surgical therapy: IV. Results in patients with left [44] Ingram D, Fulton R, Portal R, et al. Vomiting as
anterior descending coronary artery disease. Am J a diagnostic aid in acute ischemic cardiac pain.
Cardiol 1981;48:51724. BMJ 1980;281:6367.
[28] Krauss KR, Hutter AM Jr, DeSanctis RW. Acute [45] Canto J, Fincher C, Kiefe C, et al. Atypical presen-
coronary insuciency. Course and follow-up. Cir- tations among Medicare beneciaries with unstable
culation 1972;45(Suppl 1):I6671. angina pectoris. Am J Cardiol 2003;91:1189.
[29] Wasson J, Sox H, Ne R, et al. Clinical prediction [46] Kannel W, Abbott R. Incidence and prognosis of
rules: applications and methodological standards. unrecognized myocardial infarction: an update on
N Engl J Med 1985;313:7939. the Framingham Study. N Engl J Med 1984;311:
[30] McCaig L, Burt C. National Hospital Ambulatory 11447.
Medical Care Survey: 2002 emergency department [47] Rosenman R, Friedman M, Jenkins C, et al. Clini-
summary. Adv Data 2004;340:134. cally unrecognized myocardial infarction in the
[31] Coronado BE, Pope JH, Grith JL, et al. Clinical Western Collaborative Group Study. Am J Cardiol
features, triage, and outcome of patients presenting 1967;19:77682.
to the ED with suspected acute coronary syndrome [48] Grimm R, Tillinghast S, Daniels K, et al. Unrecog-
but without pain: a multicenter study. Am J Emerg nized myocardial infarction; experience in the Mul-
Med 2004;22:56874. tiple Risk Factor Intervention Trial (MRFIT).
[32] American College of Emergency Physicians. Clini- Circulation 1987;75(Suppl 2):1168.
cal policy for the initial approach to adults present- [49] Kannel WB. Silent myocardial ischemia and infarc-
ing with a chief complaint of chest pain with no tion: insights from the Framingham study. Cardiol
history of trauma. Ann Emerg Med 1995;25: Clin 1986;4:58391.
27499. [50] McCarthy B, Beshansky J, DAgostino R, et al.
[33] Short D. Diagnosis of slight and subacute coronary Can missed diagnoses of acute myocardial infarc-
attacks in the community. Br Heart J 1981;45: tion in the emergency room be reduced? Clin Res
299310. 1989;37:779A.
446 POPE & SELKER

[51] Gordon T, Sorlie P, Kannel W. Coronary heart dis- [65] Wrenn K. Protocols in the emergency room evalu-
ease, atherothrombotic brain infarction, intermit- ation of chest pain: do they fail to diagnose lateral
tent claudicationa multivariate analysis of some wall myocardial infarction? J Gen Intern Med
factors related to their incidence: Framingham 1987;2:667.
Study, 16-year follow-up. Washington (DC): US [66] Nestico P, Hakki A, Iskandrian A, et al. Electro-
Government Printing Oce; 1971. cardiographic diagnosis of posterior myocardial
[52] Truett J, Corneld J, Kannel W. A multivariate infarction revisited. J Electrocardiol 1986;19:
analysis of the risk of coronary artery disease in 3340.
Framingham. J Chronic Dis 1967;20:51124. [67] Fisch C. Electrocardiography, exercise stress test-
[53] Jayes R, Larsen G, Beshansky J, et al. Physician ing, and ambulatory monitoring. In: Kelly W, edi-
electrocardiogram reading in the emergency de- tor. Textbook of internal medicine. Philadelphia:
partment: accuracy and eect on triage decisions: Lippincott; 1989. p. 30516.
ndings from a multicenter study. J Gen Intern [68] Rusnak R, Stair T, Hansen K, et al. Litigation
Med 1992;7:38792. against the emergency physician: common features
[54] Henrikson C, Howell E, Bush D, et al. Chest pain in cases of missed myocardial infarction. Ann
relief by nitroglycerin does not predict active coro- Emerg Med 1989;18:102934.
nary artery disease. Ann Intern Med 2003;139: [69] Griner P, Mayewski R, Mushlin A, et al. Selection
97986. and interpretation of diagnostic tests and proce-
[55] Orlando R, Bozymski E. Clinical and manometric dures: principles and applications. Ann Intern
eects of nitroglycerin in diuse esophageal spasm. Med 1981;94:55792.
N Engl J Med 1973;289:235. [70] Bell M, Montarello J, Steele P. Does the emergency
[56] Killip T, Kimball J. Treatment of myocardial in- room electrocardiogram identify patients with sus-
farction in a coronary care unit. A two year experi- pected myocardial infarction who are at low risk of
ence with 250 patients. Am J Cardiol 1967;20: acute complications? Aust N Z J Med 1990;20:
45764. 5649.
[57] Ornato JP, Selker HP, Zalenski RJ. NIH National [71] Zalenski R, Sloan E, Chen E, et al. The emergency
Heart Attack Alert Program Working Group on department ECG and immediate life-threatening
Evaluation of Technologies for Identifying Acute complications in initially uncomplicated suspected
Cardiac Ischemia in Emergency Departments Re- myocardial ischemia. Ann Emerg Med 1988;17:
port. Ann Emerg Med 2001;37:45094. 2216.
[58] Selker H. Electrocardiograms and decision aids in [72] Cohen M, Hawkins L, Geeenburg S, et al. Useful-
coronary care triage: the truth but not the whole ness of ST-segment changes in R 2 leads on the
truth. J Gen Intern Med 1987;2:6770. emergency room electrocardiogram in either unsta-
[59] Homan J, Igarashi E. Inuence of electrocar- ble angina pectoris or non-Q-wave myocardial in-
diographic ndings on admission decisions in pa- farction in predicting outcome. Am J Cardiol
tients with acute chest pain. Am J Med 1985;79: 1991;67:136873.
699707. [73] Fesmire F, Percy RF, Wears R, et al. Initial ECG in
[60] Brush J, Brand D, Acampora D, et al. Use of the Q wave and non-Q wave myocardial infarction.
initial electrocardiogram to predict in-hospital Ann Emerg Med 1989;18:7416.
complications of acute myocardial infarction. N [74] Rubenstein L, Greeneld S. The baseline ECG in
Engl J Med 1985;312:113741. the evaluation of acute cardiac complaints.
[61] Slater D, Hlatky M, Mark D, et al. Outcome in sus- JAMA 1980;244:25369.
pected acute myocardial infarction with normal [75] Miller DH, Kligeld P, Schreiber TL, et al. Rela-
or minimally abnormal admission electrocardio- tionship of prior myocardial infarction to false-
graphic ndings. Am J Cardiol 1987;60:76670. positive electrocardiographic diagnosis of acute
[62] Stark M, Vacek J. The initial electrocardiogram injury in patients with chest pain. Arch Intern
during admission for myocardial infarction; use Med 1987;147:25761.
as a predictor of clinical course and facility utiliza- [76] Goldberger A. Myocardial infarction electrocar-
tion. Arch Intern Med 1987;147:8436. diographic dierential diagnosis. St. Louis (MO:
[63] Rude R, Poole W, Muller J, et al. Electrocardio- CV Mosby; 1979.
graphic and clinical criteria for recognition of acute [77] Kaul P, Newby L, Fu Y, et al. Troponin T and
myocardial infarction based on analysis of 3,697 quantitative ST-depression oer complementary
patients. Am J Cardiol 1983;52:93642. prognostic information in the risk stratication of
[64] Lopez-Sendon J, Coma-Canella I, Alcasena S, et al. acute coronary syndrome patients. J Am Coll Car-
Electrocardiographic ndings in acute right ven- diol 2003;41:37180.
tricular infarction: sensitivity and specicity of elec- [78] Granborg J, Grande P, Pederson A. Diagnostic and
trocardiographic alterations in right precordial prognostic signicance of transient isolated nega-
leads V4R,V5R,V1,V2,V3. J Am Coll Cardiol tive T-waves in suspected acute myocardial infarc-
1985;19:12739. tion. Am J Cardiol 1986;57:2037.
ACS IN THE ED 447

[79] Fisch C. Abnormal ECG in clinically normal indi- [94] Rao S, Ohman E, Granger C, et al. Prognostic
viduals. JAMA 1983;250:13213. value of isolated troponin elevation across the spec-
[80] DeWood M, Stifer W, Simpson C, et al. Coronary trum of chest pain syndromes. Am J Cardiol 2003;
arteriographic ndings soon after non-Q-wave 91:93640.
myocardial infarction. N Engl J Med 1986;315: [95] Morrow D. Troponins in patients with acute coro-
41723. nary syndromes: biologic, diagnostic, and thera-
[81] Kennedy J. Non-Q-wave myocardial infarction. peutic implications. Cardiovasc Toxicol 2001;1:
N Engl J Med 1977;315:4513. 10510.
[82] Behar S, Schor S, Kariv I, et al. Evaluation of elec- [96] Aviles R, Askari A, Lindahl B, et al. Troponin t lev-
trocardiogram in emergency room as a decision- els in patients with acute coronary syndromes, with
making tool. Chest 1977;71:48691. or without renal dysfunction. N Engl J Med 2002;
[83] McGuinness J, Begg T, Semple T. First electrocar- 346:204752.
diogram in recent myocardial infarction. BMJ [97] Montague C, Kircher T. Myoglobin in the early
1976;2:44951. evaluation of acute chest pain. Am J Clin Pathol
[84] Gibler W, Runyon J, Levy R, et al. A rapid diag- 1995;104:4726.
nostic and treatment center for patients with chest [98] Kontos M, Anderson F, Schmidt K, et al. Early di-
pain in the emergency department. Ann Emerg agnosis of acute myocardial infarction in patients
Med 1995;25:18. without ST-segment elevation. Am J Cardiol
[85] Hedges J, Young G, Henkel G, et al. Serial ECGs 1999;83:1558.
are less accurate than serial CK-MB results for [99] Ikeda H, Takajo Y, Ichiki K, et al. Increased solu-
emergency department diagnosis of myocardial ble form of P-selectin in patients with unstable an-
infarction. Ann Emerg Med 1992;21:144550. gina. Circulation 1995;92:16936.
[86] Kirk J, Turnipseed S, Lewis W, et al. Evaluation of [100] Itoh T, Nakai K, Ono M, et al. Can the risk for
chest pain in low-risk patients presenting to the acute cardiac events inacute coronary syndromes
emergency department: the role of immediate exer- be indicated by platelet membrane activation
cise testing. Ann Emerg Med 1998;32:17. marker P-selectin? Coron Artery Dis 1995;6:
[87] Lewis W, Amsterdam E, Turnipseed S. Immediate 64550.
exercise testing of low-risk patients with known [101] Bassuk S, Rifai N, Ridker P. High-sensitivity C-
coronary artery disease presenting to the emer- reactive protein: clinical importance. Curr Probl
gency department with chest pain. J Am Coll Car- Cardiol 2004;29:43993.
diol 1999;33:18437. [102] Oltrona L, Ottani F, Galvani M. Clinical signi-
[88] Tsakonis J, Shesser R, Rosenthal R, et al. Safety of cance of a single measurement of troponin-I and
immediate treadmill testing in selected emergency C-reactive protein at admission in 1773 consecutive
department patients with chest pain: a preliminary patients with acute coronary syndromes. Am Heart
report. Am J Emerg Med 1991;9:5579. J 2004;148:40515.
[89] Kerns J, Shaub T, Fontanarosa P. Emergency car- [103] Avanzas P, Arroyo-Espliguero R, Cosin-Sales J,
diac stress testing in the evaluation of emergency et al. Markers of inammation and multiple
department patients with atypical chest pain. Ann complex stenoses (pancoronary plaque vulnera-
Emerg Med 1993;22:7948. bility) in patients with non-ST segment elevation
[90] Gerhardt W, Waldenstrom J, Horder M, et al. Cre- acute coronary syndromes. Heart 2004;90:
atine kinase and creatine kinase B-subunit activity 84752.
in serum in cases of suspected myocardial infarc- [104] Pai J, Pischon T, Ma J, et al. Inammatory markers
tion. Clin Chem 1982;28:27783. and risk of coronary heart disease in men and
[91] Roxin L, Cullhed I, Groth T, et al. The value of se- women. N Engl J Med 2004;351:2599610.
rum myoglobin determinations in the early diagno- [105] de Lemos J, Morrow D, Bentley J, et al. The prog-
sis of acute myocardial infarction. Acta Med Scand nostic value of B-type natriuretic peptide in
1984;215:41725. patients with acute coronary syndromes. N Engl J
[92] Fesmire F, Christenson R, Feintuch T. Delta creat- Med 2001;345:101421.
inine kinase-MB outperforms myoglobin at two [106] Laterza O, Cameron S, Chappell D, et al. Eval-
hours during emergency department identication uation of pregnancy-associated plasma protein A
and exclusion troponin positive non-ST-segment as a prognostic indicator in acute coronary
elevation acute coronary syndromes. Ann Emerg syndrome patients. Clin Chim Acta 2004;348:
Med 2004;44:129. 1639.
[93] Morrow D, Cannon C, Rifai N, et al. Ability of mi- [107] Johnson B, Kip K, Marroquin O, et al. Serum am-
nor elevations of troponins I and T to predict ben- yloid A as a predictor of coronary artery disease
et from an early invasive strategy in patients with and cardiovascular outcome in women: the Na-
unstable angina or non-ST elevation myocardial in- tional Heart, Lung, and Blood Institute-Sponsored
farction: results from a randomized trial. JAMA Womens ischemia Syndrome Evaluation (WISE).
2001;286:240512. Circulation 2004;109:72632.
448 POPE & SELKER

[108] Brennan M, Penn M, Van Lente F, et al. Prognostic [121] Sarasin FP, Reymond JM, Grith JL, et al. Impact
value of myeloperoxidase in patients with chest of the acute cardiac ischemia time-insensitive pre-
pain. N Engl J Med 2003;349:1595604. dictive instrument (ACI-TIPI) on the speed of tri-
[109] Blankenberg S, Rupprecht H, Bickel C, et al. Glu- age decision making for emergency department
tathione peroxidase 1 activity and cardiovascular patients presenting with chest pain. J Gen Intern
events in patients with coronary artery disease. Med 1994;9:18794.
N Engl J Med 2003;349:160513. [122] Gruppo Italiano per lo Studio Della Streptochinasi
[110] Heeschen C, Dimmeler S, Fichtlscherer S, et al. NellInfarcto Miocardico (GISSI). Eectiveness of
Prognostic value of placental growth factor in intravenous thrombolytic therapy in acute myocar-
patients with acute chest pain. JAMA 2004;291: dial infarction. Lancet 1986;1:397401.
43541. [123] Fibrinolytic Therapy Trialists (FTT) Collabora-
[111] Nomoto K, Oguchi S, Watanabe I, et al. Involve- tive Group. Indications for brinolytic therapy
ment of inammation in acute coronary syndromes in suspected acute myocardial infarction: collab-
assessed by levels of high-sensitivity C-reactive orative overview of early mortality and major
protein, matrix metalloproteinase-9 and soluble morbidity results from all randomized trials
vascular-cell adhesion molecule. J Cardiol 2003; of more than 1000 patients. Lancet 1994;343:
42:2016. 31122.
[112] de Lemos J, Morrow D, Sabatine M, et al. Associ- [124] Selker HP, Grith JL, Beshansky JR, et al. The
ation between plasma levels of monocyte chemoat- thrombolytic predictive instrument project: com-
tractant protein-1 and long-term clinical outcomes bining clinical study databases to take medical eec-
in patients with acute coronary syndromes. Circu- tiveness research to the streets. Washington D.C.:
lation 2003;107:6905. AHCPR, DHHS; 1992:931.
[113] Sabatine MS, Morrow DA, de Lemos JA, et al. [125] Cannon CP, Antman EM, Walls R, Braunwald E.
Multimarker approach to risk stratication in Time as an adjunctive agent to thrombolytic ther-
non-ST elevation acute coronary syndromes: si- apy. J Thromb Thrombolysis 1994;1:2734.
multaneous assessment of troponin I, C-reactive [126] National Heart Attack Alert Program Coordinat-
protein, and B-type natriuretic peptide. Circulation ing Committee. 60 Minutes to Treatment Working
2002;105:17603. Group. Emergency department: rapid identica-
[114] Mohler ER III, Ryan T, Segar D, et al. Clinical util- tion and treatment of patients with acute myocar-
ity of troponin T levels and echocardiography in dial infarction. Ann Emerg Med 1994;23:31129.
the emergency department. Am Heart J 1998;135: [127] National Heart Attack Alert Program Coordinat-
25360. ing Committee Access to Care Subcommittee.
[115] Udelson J, Beshansky J, Ballin D, et al. Myocardial Emergency medical dispatching: rapid identica-
perfusion imaging for evaluation and triage of tion and treatment of acute myocardial infarction.
patients with suspected acute cardiac ischemia. Am J Emerg Med 1995;13:6773.
JAMA 2002;288:2693700. [128] National Heart Attack Alert Program Coordinat-
[116] Mohr R, Rath S, Meir O, et al. Changes in systemic ing Committee Access to Care Subcommittee.
vascular resistance detected by the arterial resis- 91-1: rapid identication and treatment of acute
tometer: preliminary report of a new method tested myocardial infarction. Am J Emerg Med 1995;13:
during percutaneous transluminal coronary angio- 18895.
plasty. Circulation 1986;74:7805. [129] Rogers WJ, Bowlby LJ, Chandra NC, et al. Treat-
[117] Mohr R, Meir O, Smolinsky A, et al. A method for ment of myocardial infarction in the United States
continuous on-line monitoring of systemic vascular (1990 to 1993) observations from the national reg-
resistance (COMS) after open heart procedures. istry of myocardial infarction. Circulation 1994;90:
J Cardiovasc Surg (Torino) 1987;28:55865. 210314.
[118] Mohr R, Dinbar I, Bar-El Y, et al. Correlation be- [130] National Heart Attack Alert Program Coordinat-
tween myocardial ischemia and changes in arterial ing Committee Access to Care Subcommittee. Ac-
resistance during coronary bypass surgery. J Cardi- cess to timely and optimal care of patients with
othorac Vasc Anesth 1992;6:3341. acute coronary syndromesdcommunity planning
[119] Kantartzis M, Sunderdiek U, Bircks W, et al. Car- considerations: a report by the National Heart At-
diac eciency during coronary occlusion and dur- tack Alert Program. J Thromb Thrombolysis 1998;
ing reperfusion after emergency revascularization 6:1936.
under cardioprotection. Thorac Cardiovasc Surg [131] National Heart Attack Alert Program Coordinat-
1996;44:206. ing Committee Working Group on Educational
[120] Gorenberg M, Marmor A, Rotstein H. Detection Strategies to Prevent Prehospital Delay in Patients
of chest pain of non-cardiac origin at the emer- at High Risk for Acute Myocardial Infarction.
gency room by a new non-invasive device avoid- Educational strategies to prevent prehospital de-
ing unnecessary admission to hospital. Emerg lay in patients at high risk for acute myocardial in-
Med J 2005;22:4869. farction: a report by the National Heart Attack
ACS IN THE ED 449

Alert Program. J Thromb Thrombolysis 1998;6: 19711987. MMWR Morb Mort Wkly Rep 1992;
4761. 41:5269.
[132] Sims RJ, Topol EJ, Holmes DR, et al. Link be- [143] Misra D. The womens health data book: a prole
tween the angiographic substudy and mortality of womens health in the United States. 3rd edition.
outcomes in a large randomized trial of myocardial Washington (DC): Jacobs Institute of Womens
reperfusion: importance of early and complete Health. The Henry J. Kaiser Family Foundation;
infarct artery reperfusion. Circulation 1995;91: 2001. p. 6973.
19238. [144] Diamond G, Stanilo H, Forrester J, et al. Com-
[133] Ryan TJ, Antman EM, Brooks NH, et al. 1999 puter-assisted diagnosis in the noninvasive evalua-
update: ACC/AHA guidelines for the manage- tion of patients with suspected coronary artery
ment of patients with acute myocardial infarction: disease. J Am Coll Cardiol 1983;1:44455.
executive summary and recommendations: a re- [145] Merz C, Kelsey S, Pepine C, et al. The Womens
port of the American College of Cardiology/ Ischemia Syndrome Evaluation (WISE) study:
American Heart Association Task Force on Prac- protocol design, methodology and feasibility re-
tice Guidelines (Committee on Management of port. J Am Coll Cardiol 1999;33:145361.
Acute Myocardial Infarction). Circulation 1999; [146] Kennedy J, Killip T, Fisher L, et al. The clinical
100:101630. spectrum of coronary artery disease and its surgical
[134] Rogers WJ, Canto JG, Lambrew CT, et al, for the and medical management, 19741979. The Coro-
Investigators in the National Registry of Myocar- nary Artery Surgery study. Circulation 1982;66:
dial Infarction 1, 2, and 3. Temporal trends in the III1623.
treatment of over 1.5 million patients with myocar- [147] Bybee K, Kara T, Prasad A, et al. Systemic re-
dial infarction in the US from 1990 through 1999. view: transient left ventricular apical ballooning:
J Am Coll Cardiol 2000;36:205663. a syndrome that mimics ST-segment elevation
[135] Berger AK, Radford MJ, Krumholtz HM. Factors myocardial infarction. Ann Intern Med 2004;
associated with delay in reperfusion therapy in el- 141:85865.
derly patients with acute myocardial infarction: [148] Douglas P, Ginsburg G. The evaluation of
analysis of the cooperative cardiovascular project. chest pain in women. N Engl J Med 1996;334:
Am Heart J 2000;139:98592. 13115.
[136] Selker HP, Beshansky JR, Grith JL, for the TPI [149] Coronado B, Grith J, Beshansky J, et al. Hospital
Trial Investigators. Use of the electrocardiograph- mortality in women and men with acute cardiac is-
based thrombolytic predictive instrument to assist chemia: a prospective multicenter study. J Am Coll
thrombolytic and reperfusion therapy for acute Cardiol 1997;29:14906.
myocardial infarction: a multicenter randomized [150] Vaccarino V, Parsons L, Every N, et al. Sex-based
clinical eectiveness trial. Ann Intern Med 2002; dierences in early mortality after myocardial in-
137:8795. farction. National Registry of Myocardial Infarc-
[137] Barron HV, Bowlby LJ, Breen T, et al, for the Na- tion 2 participants. N Engl J Med 1999;341:21725.
tional Registry of Myocardial Infarction 2 Investi- [151] Jacobs A, Kelsey S, Brooks M, et al. Better out-
gators. Use of reperfusion therapy for acute come for women compared with men undergoing
myocardial infarction in the United States. Data coronary revascularization: a report from the By-
from the National Registry of Myocardial Infarc- pass Angioplasty Revascularization Investigation
tion 2. Circulation 1998;97:11506. (BARI). Circulation 1998;98:127985.
[138] Ting H, Lee T, Soukup J, et al. Impact of physician [152] Edwards F, Carey J, Grover F, et al. Impact of gen-
experience on triage of emergency room patients der on coronary bypass operative mortality. Ann
with acute chest pain at three teaching hospitals. Thorac Surg 1998;66:12531.
Am J Med 1991;91:4018. [153] Mehilli J, Kastrati A, Dirschinger J, et al. Dieren-
[139] Reilly B, Evans A, Schaider J, et al. Impact of a clin- ces in prognostic factors and outcomes between
ical decision rule on hospital triage of patients with women and men undergoing coronary artery stent-
suspected acute cardiac ischemia in the emergency ing. JAMA 2000;284:1799805.
department. JAMA 2002;288:34250. [154] Maynard C, Beshansky J, Grith J, et al. Inuence
[140] Baxt W, Skora J. Prospective validation of articial of sex on the use of cardiac procedures in patients
neural network trained to identify acute myocar- presenting to the emergency department: a prospec-
dial infarction. Lancet 1996;347:125. tive multicenter study. Circulation 1996;94:II938.
[141] Selker H, Beshansky J, Grith J. Use of the elec- [155] Ayanian J, Epstein A. Dierences in the use of pro-
trocardiograph-based thrombolytic predictive in- cedures between women and men hospitalized for
strument to assist thrombolytic and reperfusion coronary heart disease. N Engl J Med 1991;325:
therapy for acute myocardial infarction. Ann In- 2215.
tern Med 2002;137:8795. [156] Romeo F, Rosano G, Martuscelli E, et al. Long-
[142] Centers for Disease Control (CDC). Coronary term follow-up of patients initially diagnosed as
heart disease incidence, by sexdUnited States, syndrome X. Am J Cardiol 1993;71:66973.
450 POPE & SELKER

[157] Sullivan A, Holdright D, Wright C, et al. Chest chest discomfort: the Framingham Heart Study.
pain in women: clinical, investigative, and prognos- Am J Med 1990;89:297302.
tic features. BMJ 1994;308:8836. [172] Milner K, Funk M, Arnold A, et al. Typical symp-
[158] Kaski J, Rosano G, Collins P, et al. Cardiac syn- toms are predictive of acute coronary syndromes in
drome X: clinical characteristics and left ventricu- women. Am Heart J 2002;143:2838.
lar function. Long-term follow-up study. J Am [173] Milner K, Funk M, Richards S, et al. Gender dier-
Coll Cardiol 1995;25:80714. ences in symptom presentation associated with cor-
[159] Bairey Merz N, Bonow RO, Sopko G, et al. Wom- onary heart disease. Am J Cardiol 1999;84:3969.
ens Ischemic Syndrome Evaluation: current status [174] Goldberg R, Go D, Cooper L, et al. Age and sex
and future research directions: report of the Na- dierences in presentation of symptoms among
tional heart, Lung and Blood Institute workshop. patients with acute coronary disease: the REACT
Circulation 2002;109:8057. trial. Coron Artery Dis 2000;11:399407.
[160] Lerner D, Kannel W. Patterns of coronary heart [175] Maseri A, Crea F, Kaski J, et al. Mechanisms of an-
disease morbidity and mortality in the sexes: a 26- gina pectoris in syndrome X. J Am Coll Cardiol
year follow-up of the Framingham population. 1991;17:499506.
Am Heart J 1986;111:38390. [176] Panting J, Gatehouse P, Yang G, et al. Abnormal
[161] Smith W, Kenicer M, Tunstall-Pedoe H, et al. subendocardial perfusion in cardiac syndrome X
Prevalence of coronary heart disease in Scotland: detected by cardiovascular magnetic resonance im-
Scottish Heart Health Study. Br Heart J 1990; aging. N Engl J Med 2002;346:194853.
64:2958. [177] Quyyumi A. Endothelial function in health and dis-
[162] Elveback L, Connolly D. Coronary heart disease in ease: new insights into the genesis of cardiovascular
residents of Rochester, Minnesota, V: prognosis of disease. Am J Med 1998;105:32S9S.
patients with CAD based on initial manifestation. [178] Wiviott S, Cannon C, Morrow D, et al. Dierential
Mayo Clin Proc 1985;60:30531. expression of cardiac biomarkers by gender in
[163] Seeman T, Mendes de Leon C, Berkman L, et al. patients with unstable angina/non-ST-elevation
Risk factors for coronary heart disease among myocardial infarction: a TACTICS-TIMI 18 sub-
older men and women: a prospective study of com- study. Circulation 2004;109:5806.
munity-dwelling elderly. Am J Epidemiol 1993;138: [179] Mueller C, Neumann F, Roskamm H, et al. Women
103749. do have an improved long-term outcome after non-
[164] Maynard C, Weaver W. Treatment of women with ST-elevation acute coronary syndromes treated
acute MI: new ndings from the MITI Registry. very early and predominantly with percutaneous
J Myocard Ischemia 1992;4:2737. coronary intervention: a prospective study in
[165] Sharpe P, Clark N, Janz N. Dierences in the im- 1,450 consecutive patients. J Am Coll Cardiol
pact and management of heart disease between 2002;40:24550.
older women and men. Women Health 1991;17: [180] Maynard C, Fisher L, Passamani E, et al. Blacks in
2534. the Coronary Artery Surgery Study (CASS): risk
[166] Cunningham M, Lee T, Cook E, et al. The eect of factors and coronary artery disease. Circulation
gender on the probability of myocardial infarction 1986;74:6471.
among emergency department patients with acute [181] Cooper R, Ford E. Comparability of risk factors
chest pain. J Gen Intern Med 1989;4:3928. for coronary artery disease among black and whites
[167] Liao Y, Lui K, Dyer A, et al. Sex dierential in the in the NHANES-I epidemiologic follow-up study.
relationship of electrocardiographic ST-T abnor- Ann Epidemiol 1992;2:63745.
malities to risk of coronary death: 11.5 year fol- [182] Johnson P, Lee T, Cook E, et al. Eect of race on the
low-up ndings of the Chicago heart association presentation and management of patients with acute
detection project in industry. Circulation 1987;75: chest pain. Ann Intern Med 1993;118:593601.
34752. [183] Curry C, Lewis J. Cardiac anatomy and function in
[168] Lusiani L, Perrone A, Pesavento R, et al. Preva- hypertensive blacks. In: Hall W, Sanders E,
lence, clinical features, and acute course of atypical Shulman N, editors. Hypertension in blacks. Chicago
myocardial infarction. Angiology 1994;45:4955. (IL): Year Book Medical Publishers; 1985. p. 617.
[169] Fiebach N, Viscoli C, Horwitz R. Dierences be- [184] Lenfant C. Report of the NHLBI working group
tween women and men in survival after myocardial on research in coronary artery disease in blacks.
infarction: biology or methodology? JAMA 1990; Circulation 1994;90:161323.
263:10926. [185] Becker L, Han B, Meyer P, et al. Racial dierences
[170] Dittrich H, Gilpin E, Nicod P, et al. Acute myocar- in the incidence of cardiac arrest and subsequent
dial infarction in women: inuence of gender on survival. N Engl J Med 1993;329:6006.
mortality and prognostic variables. Am J Cardiol [186] Cowie M, Fahrenbruch C, Cobb L, et al. Out-
1988;62:17. of-hospital cardiac arrest: racial dierences in
[171] Murabito JM, Anderson KM, Kannel WB, et al. outcome in Seattle. Am J Public Health 1993;
Risk of coronary heart disease in subjects with 83:9559.
ACS IN THE ED 451

[187] Maseri A. Womens Ischemic Syndrome Evalua- [191] Zucker D, Grith J, Beshansky JR, et al. Pre-
tion: current status and future research directions. sentations of acute myocardial infarction in
Circulation 2004;109:e623. men and women. J Gen Intern Med 1997;12:
[188] Maynard C, Beshansky J, Grith J, et al. Causes 7987.
of chest pain and symptoms suggestive of acute [192] Jayes R, Beshansky J, DAgostino R, et al. Do
cardiac ischemia in African-American patients patients coronary risk factor reports predict acute
presenting to the emergency department: a multi- cardiac ischemia in the emergency department?
center study. J Natl Med Assoc 1997;89:66571. A multicenter study. J Clin Epidemiol 1992;45:
[189] Kohn L, Corrigan J, Donaldson N, editors. To err 6216.
is human: building a safer health system. Washing- [193] Miller C, Lindsell C, Khandelwal S, et al. Is the ini-
ton (DC): National Academy Press; 2000. tial diagnostic impression of noncardiac chest
[190] Leale L. Error in medicine. JAMA 1994;272: pain adequate to exclude cardiac disease? Ann
18517. Emerg Med 2004;44:56574.
Cardiol Clin 23 (2005) 453465

Use of Biomarkers in the Emergency


Department and Chest Pain Unit
Allan S. Jae, MD
Consultant in Cardiology and Laboratory Medicine Mayo Clinic and Mayo Medical School,
200 First Street SW, Rochester, MN 55905, USA

The use of biomarkers of cardiac injury in the interaction of actin and myosin [1]. Tissue-specic
emergency department (ED) and observation unit isoforms of each troponin exist; however, the car-
settings has several nuances that are dierent and, diac form of troponin C is shared by smooth mus-
therefore, worthy of its own set of use guidelines. cle, so it lacks cardiac specicity [2,3]. For cardiac
The markers that are used, however, are the same. troponin I (cTnI), the cardiac isoform is structur-
The primary marker of choice continues to be ally dierent due to the presence of a 32-amino
cardiac troponin (Tn). Other markers that have acid posttranslational tail [4] and sequence dissim-
been advocated for use in the ED because of the ilarity with other isoforms [5]. Various monoclo-
need for rapid triage have been myoglobin and nal antibodies have been developed that are
fatty acid binding protein. In addition, some highly specic and have no cross-reactivity with
centers still prefer less sensitive and less specic other forms [6]. Since cTnI has not been found
markers, such as creatine kinase myocardial band in any tissue outside of the heart, even in response
(CK-MB). More recently, a push has occurred to to tissue injury, it has unique specicity for the
develop markers of ischemia, such as ischemia heart [7,8].
modied albumin (IMA), to determine which Three dierent genes control the expression of
patients have ischemia, even in the absence of cardiac troponin T. These genes and alternative
cardiac injury. As troponin assays become more mRNA splicing produce a series of isoforms with
sensitive and how to use them becomes better un- varying sequences [9]. Cardiac muscle contains
derstood, the use of other markers are being rele- four of these isoforms, but only one is character-
gated to lesser and lesser roles. Markers of istic of the normal adult heart, and that is the
ischemia would be useful but at present, despite one against which antibodies have been made
some enthusiasm, they are not ready for routine [10]. Some of the other cardiac troponin T (cTnT)
use. Before describing the recommendations for isoforms are expressed in other tissues, including
clinical use of biomarkers in the ED, a basic under- skeletal muscle in response to injury [11], and
standing of some of the science and measurement the initial assay for cTnT detected some of these
issues related to these analytes is helpful. and cross-reacting skeletal muscle TnT [12].
However, it has been established with immuno-
Background histochemistry and PCR, that the antibodies used
in the present iteration of the assay do not de-
Troponin tect these forms [13,14]. Accordingly, cTnI and
The Tns are a complex of three proteins (I, C, cTnT have high specicity for the heart.
and T) that regulate the calcium mediated In addition to having high specicity, these
proteins have high sensitivity. Most of the tropo-
Dr. Jae receives research support and is a consul- nin is complexed to the contractile apparatus. A
tant for Dade, Beckman, and Roche. He has consulted small amount (3% for cTnI and 6% for cTnT)
for most major diagnostic companies over the years. exists which is not structurally bound [15,16]. This
E-mail address: Jae.Allan@Mayo.edu small amount has been termed the cytosolic
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.013 cardiology.theclinics.com
454 JAFFE

pool, although its localization was not proved detection of the 99th percentile of a normal refer-
denitively. The relative amount of troponin in ence population with that level of precision
this pool is similar to that of CK-MB [15]. This [21,25]. At present, only one assay meets that
pool is released acutely. Because it is of the same specication. In general, the 10% CV value is
magnitude as CK-MB, troponin might be ex- close to the 99th percentile but how close they
pected to have similar sensitively, but the sensitiv- are varies assay to assay (Table 1). Thus, clinicians
ity of the troponins is substantially better because must know the characteristics of the assay that
a greater percentage of what is lost from the heart they rely on. In addition, other analytic issues ex-
eventually resides in blood (ie, the so-called re- ist that can confound assays. Good laboratories
lease ratio is higher) [17]. Subsequently, pro- should be able to help trouble-shoot potential
longed elevations of troponin occur as a result problems. The most common problem is brin in-
of degradation of the contractile pool in the area terference, which may on occasion require addi-
that has been injured. The cytosolic pool permits tional centrifugation of the sample to remove
the early kinetics of release similar to that of cross-reacting brin.
CK-MB or even earlier with more sensitive assays. In addition, it should be appreciated that
The persistence of elevation is due to release from troponin elevations indicate cardiac injury but
the structural pool, since the half life of troponin do not dene the nature of the cardiac injury. In
in the circulation is short [18]. The degree of per- patients who present with chest pain, most tropo-
sistence varies tremendously and is somewhat lon- nin elevations likely are related to coronary artery
ger for cTnT [19] than for some of the cTnI disease. The astute clinician must be aware,
assays. For cTnI, the degree of persistence in however, that elevations in troponin can occur
part depends on which forms are measured and from acute pulmonary embolism, myocarditis,
which epitopes the antibodies used in the assay and congestive heart failure to name just a few
are targeted toward. A 7- to 10-day time window common entities that can confound the diagnosis
during which elevations may be present, however, of acute myocardial infarction [26]. Thus, the idea
is a reasonable estimate. Accordingly, if the initial that an elevated troponin is synonymous with
troponin in a patient who presents with chest pain
is elevated, it may be from a prior event or from
an acute event. In that situation, a rising pattern Table 1
is helpful in determining if the event is acute. Concentrations corresponding to 10% CV imprecision
The heterogeneity of assays indicated above is and 99th percentile reference limit for the evaluated tro-
of particular importance. Dierent assays detect ponin assays
dierent forms, and various complex formations Ratio of
and protein modications exist that can alter the 10% CV
relative sensitivity and specicity of one assay 99th 10% total concentration
compared with another. The International Feder- percentile CV to 99th
ation of Clinical Chemistry and Laboratory limit* concentration percentile
Medicine (IFCC) group has published quality Platform (mg/L) (mg/L) limit
specications for troponin assays and an analysis AxSYM 0.30 1.22 4.1
of the presently available assays [20,21]. Clinicians Centaur 0.10 0.33 3.3
must understand which assay is being used in their Access 0.04 0.06 1.5
center to understand how to use the data opti- Vidas 0.10 0.36 3.6
mally [22]. As lower values are used, such as the Dimension 0.07 0.26 3.7
RxL
recommended value of the 99th percentile as rec-
Stratus CS 0.07 0.10 1.4
ommended by the ESC/ACC Committee on the
Alpha Dx 0.15 ND
Redenition of MI [23,24], the assays invariably E170 0.01 0.04 4.0
perform with less precision. This puts clinicians AIA21 0.06 0.09 1.5
and patients at risk for false-positive and false-
Abbreviations: CV, coecient of variability; ND, not
negative results. Accordingly, a recommendation
determined.
has been made that the 99th percentile be mea- * Data obtained from manufacturers package insert
sured with a less than 10% coecient of variabil- or through personal communications with manufacturers.
ity (CV). This measurement will eliminate many Data from Panteghini M, Pagani F, Yeo KT, et al.
false-positive and false-negative values and has Evaluation of imprecision for cardiac troponin assays
been recommended until all assays can achieve at low-range concentrations. Clin Chem 2004;50:32732.
BIOMARKERS IN THE ED AND CPU 455

coronary artery disease can lead to missed diagno- troponin cuto values in their studies, which am-
ses and suboptimal patient care. plies the value of the testing [29,30]. As better
troponin assays have been developed, the amount
of time saved by a nonspecic marker like myo-
Creatine kinase myocardial band globin has progressively diminished [31].
CK-MB [27] served us well for many years, and
facile assays exist that are reasonably sensitive.
Fatty acid binding protein
The percentage of CK-MB depleted from the
heart that ends up in the blood, however, is mod- Fatty acid binding protein [27] is similar to myo-
est (15% in the absence of reperfusion and 30% globin in its release kinetics, but a much larger
with reperfusion), limiting its sensitivity compared percentage exists in the heart relative to skeletal
with troponin. In addition, because there is CK- muscle. This increased percentage has led some to
MB in skeletal muscle, individuals have a circulat- claim an increased degree of specicity for eleva-
ing constitutive level and CK-MB elevations can tions. Again, with more sensitive assays for tropo-
be diagnosed only when that broad normal range nin, it is unclear how much time measuring fatty
is exceeded. This further limits sensitivity, but it acid binding protein might save.
also has implications for specicity. Elevations
can occur as a result of skeletal muscle damage
in patients with renal failure and abnormalities Markers of ischemia
in clearance associated with hypothyroidism. In
The only test presently approved for detection
addition, re-expression of the B-chain gene in
of ischemia by the Federal Drug Administration
skeletal muscle when damage to muscle occurs
(FDA) is IMA. Based on experimental data, it
can increase the percentage of CK-MB and con-
has been argued that ischemia injures the amino
found diagnosis. The use of the relative index,
terminal end of albumin, which inhibits it from
which relies on the percentage of CK-MB with re-
binding cobalt [32]. This fact has been taken ad-
spect to total CK, improves specicity, but be-
vantage of in an assay that looks at the percentage
cause there is so much CK in skeletal muscle, it
of delivered cobalt that becomes bound. The argu-
diminishes sensitivity in individuals who may
ment is that this test can provide information con-
have coronary and skeletal muscle diseases [28].
cerning whether ischemia has occurred in the
In addition, a multiplicity of assay issues for
absence of necrosis. In response to ischemia, ele-
CK-MB exist that assay issues often are more
vations of IMA do occur. They are short lived
complex than those related to troponin. The role
(roughly four hours), however, and then diminish
of CK-MB in 2005 is unclear beyond its role in
[33]. These elevations, it is argued, can be used to
helping to evaluate the timing of events in patients
dene if ischemia has occurred and, thus, could be
who present with chest discomfort.
helpful in the triage of patients in the ED setting,
especially in excluding ischemia. The problem
with these studies is the lack of a clear gold stan-
Myoglobin
dard. In addition, low levels of IMA seem protec-
Myoglobin [27] is a 17.8 kd protein that is re- tive and high levels seem to have some prognostic
leased from all tissues ubiquitously. It is released signicance for eventual troponin elevations [34].
rapidly and has a short persistence in the circula- In the high pretest probability high-risk group,
tion (! 20 minutes in studies in which it has been however, it is unclear whether this test is necessary
injected) and, therefore, a pattern of increases and to further dene risk. In the low-risk group, a sim-
reductions are common. This protein responds ilar question can be posed. Most of the patients in
rapidly and quickly to insults, and if the insults whom clinicians are unsure, however, have values
are sustained, elevations will persist. Because it in the middle range and it is unclear how well they
has cleared renally, abnormalities in renal func- separate those with ischemia from those without
tion can cause substantial elevations, and because ischemia. Furthermore, ischemia in other tissues
it is ubiquitous, it lacks specicity for the heart. also stimulates abnormalities in albumin binding
Myoglobin can and has been used as a nonspecic also and assay issues exist related to the eect of
marker to screen for disease in patients who often lactate and pH [35]. These issues are why the
have been poorly triaged. In addition, many stud- FDA-approval was predicated on the fact that
ies have used high, rather then conventional, a negative test excludes ischemia. The argument,
456 JAFFE

has been made, therefore, that if a test is per-


formed and fails to show an elevation, that those
patients are probably safe to be sent home without
further testing. Various studies, some of which are
commented on in later discussion, have attempted
to prove that postulate. On the other hand, what
should be done with a positive test is less clear.
Obviously, if markers of cardiac injury subse-
quently show that necrosis exists, then the patient
is triaged easily. Suppose, however, that no evi- Fig. 1. Frequency of successful direct PCI when tropo-
nin is elevated (cTnI O .4 ng/ml) compared with when
dence of necrosis exists. Is there a reason, based
it is normal (! 0.4ng/ml). From Matetzky S, Sharir
on the fact that an ischemic bed exists somewhere,
T, Domingo M, et al. Elevated troponin I level on ad-
to do more work on that particular patient? Per- mission is associated with adverse outcome of primary
haps this is the group that requires stress testing? angioplasty in acute myocardial infarction. Circulation
Do these patients require evaluation of other or- 2000;102:16116; with permission).
gan systems also? These questions require answers
before advocacy for the widespread use of the test.
These markers of cardiac injury need to be elevated troponins do poorly. In analyses where
intertwined within the overall risk assessment of one attempts to correct for the time from the onset
patients who present to the ED. of infarction to presentation, however, a major
adverse eect of having an elevated troponin still
exists (Fig. 2) [38]. The ability to determine when
Clinical use of troponin and other biomarkers in the onset of symptoms has occurred may be inaccu-
specic patient subsets seen in the emergency rate or it may be that other pathophysiology exists
department or chest pain unit related to how infarction is initiated, which medi-
ates this eect. Nonetheless, knowing that the tro-
The use of biomarkers is predicated in large ponin is elevated in patients who present with
part on the underlying nature of the patients being STEMI predicts a lower likelihood of recanaliza-
evaluated. The following suggestions are devel- tion, a lower TIMI ow grade with reperfusion
oped predicated on risk stratication. [39], and an adverse short- and long-term progno-
sis. The therapeutic implications of this phenome-
Patients with ST elevation and myocardial non are unclear at this time, but in one study,
infarction stenting seemed to have a positive impact [40]. No
Patients who present with ST elevation myo-
cardial infarction (STEMI) do not need biomarker
measurements before the initiation of therapy.
Therapy should be started before the return of
biomarker data to ensure the most rapid initiation
of therapy, which is critical in this setting. Recent
data suggests that troponin elevations identify
patients at high risk for adverse events. Patients
who have elevated troponins have a lower rate of
coronary recanalization with thrombolysis [36] or
in response to direct percutaneous coronary inter-
vention (PCI) (Fig. 1 ) [37]. The largest part of
this eect probably is related to the fact that it takes Fig. 2. Relationship between the time of onset of symp-
time for elevations of troponin to develop, and toms, troponin and mortality. Note at any given time
from the onset of symptoms to randomization there is
therefore, patients who present with elevated tro-
an increased risk associated with an elevated troponin.
ponin, are patients who present later than those
From Ohman EM, Armstrong PW, White HD, et al.
who do not manifest elevations. Time from the Risk stratication with a point-of-care cardiac troponin
onset of symptoms to treatment is the most critical T test in acute myocardial infarction. GUSTO III Inves-
determinant of the success or failure of recanaliza- tigators. Global use of strategies to open occluded cor-
tion and limitation of infarction size, which is likely onary arteries. Am J Cardiol 1999;84:12816; with
the largest reason why patients who present with permission.
BIOMARKERS IN THE ED AND CPU 457

data exist about this issue with biomarkers other exclusively in the recurrent event category. The ef-
than troponin. fects of IIB/IIIA agents are particularly strong in
those patients who undergo percutaneous inter-
vention [48]. Their use, then, is the present stan-
High-risk patients with acute coronary syndromes
dard of care for patients who present with acute
The high-risk patient group that has been well coronary syndromes and elevated troponins [49].
represented in most of the major intervention Patients who present without elevated troponins
trials. These patients are often elderly, have had do not seem to benet from the adjunctive use
chest discomfort at rest, have transient ST seg- of low molecular weight heparin or IIB/IIIA
ment changes or bundle branch block, an increase agents in the setting of acute coronary syndromes.
in the tempo of their symptoms or signs, and The recent accelerating use of clopidogrel has
can have evidence of hemodynamic instability or raised questions about whether the results of these
arrhythmias. This group often has a high fre- previous studies would be the same if clopidogrel
quency of elevated troponins in the range of 50% had been used adjunctively. Clopidogrel recapitu-
to 60% [4144]. If so, they fulll the denition of lates the benets of IIB/IIIA agents, it could be
non-STEMI per the ESC/ACC [23,24]. argued. Alternatively, it is possible that synergism
In this patient subset, as with STEMI, an exists between clopidogrel and IIB/IIIA agents
elevated troponin often is unnecessary for deci- and that patients would benet even more from
sions concerning admission. Information suggests, the combination. Until ongoing trials answer
however, that elevated troponin values dene this key question, the standards and guidelines
a subset at high risk [4145] and provide impor- proposed by the American Heart Association/
tant information with which to guide therapy. Pa- American Cardiac College would recommend
tients who present with elevated troponins have that patients who present with an acute coronary
adverse coronary anatomy as assessed angio- syndrome and an elevated troponin receive IIB/
graphically with more thrombus, reduced TIMI IIIA agents [49]. In the CURE trial, clopidogrel
grades of perfusion, more complex lesions, and seemed to benet patients who presented with
more extensive disease (Fig. 3) [46]. Thus, they re- and without troponin elevations [50].
quire and are more apt to benet from the use of As might be expected, because short-term
more sophisticated therapies. These patients have benet is associated with the use of these agents
fewer ischemic events if treated with low molecu- so is long term benet [51]. In addition, recent
lar weight heparin rather than unfractionated hep- data has substantiated that the use of early inva-
arin [47]. These studies do not show a mortality sive therapy, which in the United States means
benet. IIB/IIIA agents also have been found to cardiac catheterization within the next 12 to 24
markedly reduce death and recurrent MI in these hours and interventional therapy (with stenting
patients, but, again, the benet is almost and a PCI or coronary artery bypass grafting),

Fig. 3. Angiographic relationships with troponin in patients with acute coronary syndromes. From Scirica BM, Morrow
DA. Troponins in acute coronary syndromes. Prog Cardiovasc Dis 2004;47:17788; with permission.
458 JAFFE

leads to substantial benet in patients who present approach is that it is costly to admit a large number
with acute coronary syndromes and elevated of patients solely because of this concern.
troponins [52,53]. This benet is not generally ob- In this setting, troponin markers are again the
served in patients who present without elevations. markers of choice. In a landmark study of 733
The one exception recently published is that consecutive patients who presented with chest
women who more often present with normal val- discomfort, Hamm and colleagues [55] demon-
ues of troponin but elevated BNP or CRP may, strated that almost all patients at short-term risk
nonetheless, benet from an early aggressive inva- could be identied using troponin (Fig. 5). Several
sive strategy (Fig. 4) [54]. Unfortunately, the data caveats existed that were important in that study.
were based not on serial troponin data but on only The rst caveat was that at least one troponin
the initial sample. Few other data exist at present value needed to have been obtained at least 6
to substantiate this claim. hours after the onset of symptoms to make sure
that there had been time for the marker to rise.
Unless the onset of symptoms is totally clear, it
Intermediate and low risk for the patient subsets
is wise to start the clock at the time of presenta-
The intermediate and low-risk subsets of tion. A second important caveat was the troponin
patients are the groups that usually are problem- value used was the limit of detectablity (ie, a highly
atic in the ED or chest pain unit. The intermediate sensitive standard). Although the assays involved
risk group often does not have rest pain or ongo- were less sensitive than contemporary assays, the
ing pain and lacks ECG changes, although this concept probably is still correct. With modern as-
group may have a high pretest probability of says, the use of the so-called 10% CV value is ad-
having coronary artery disease. In contrast, the vocated [21,25]. When there is a high pretest
low-risk group usually has normal ECGs or only probability of disease, as in the patients with overt
minimal T changes and often may not have im- acute coronary syndromes, the data are clear that
pressive risk factors. Nonetheless, because of the using the lowest level of troponin detectable, (ie,
need to diagnose as many patients as possible at the 99th percentile), provides greater predictive
risk for acute infarction given its morbidity and accuracy [56]. The pretest probability of a given
potential mortality, these patients need to be tri- elevation, being a true positive rather than a false
aged and evaluated aggressively. In addition, this positive, is high in the group with severe disease.
group is where a high percentage of malpractice When one deals with a patient population where
dollars are lost. For these reasons, a conservative the pretest probability of disease is somewhat
approach often is mandated. The downside of this less, severe disease is less likely. In that

Fig. 4. Outcomes of therapy by gender. Note patients with any positive marker (troponin, BNP or CRP) beneted.
Some suggestion of detriment exists in an invasive therapy in patients without an elevations in any marker. From Wiviott
SD, Cannon CP, Morrow DA, et al. Dierential expression of cardiac biomarkers by gender in patients with unstable
angina/non-ST-elevation myocardial infarction: a TACTICS-TIMI 18 (Treat Angina with Aggrastat and determine Cost
of Therapy with an Invasive or Conservative Strategy-Thrombolysis In Myocardial Infarction 18) substudy. Circulation
2004;109:5806; with permission.
BIOMARKERS IN THE ED AND CPU 459

Fig. 6. Events in patients with various troponin values.


Fig. 5. Survival of patients evaluated to include or ex-
Note even what are called in this gure marginal eleva-
clude acute myocardial infarction. From Hamm CW,
tions have prognostic signicance. From Henrikson
Goldmann BU, Heeschen C, et al. Emergency room tri-
CA, Howell EE, Bush DE. et al. Prognostic usefulness
age of patients with acute chest pain by means of rapid
of marginal troponin T elevation. Am J Cardiol
testing for TnT or troponin I. N Engl J Med
2004;93:2759; with permission.
1997;337:164853; with permission.

population, if one uses values where the assays are Intermediate or low-risk patients who present
imprecise, the percentage of false-positive eleva- with an elevated troponin may or may not have
tions is far greater. For that reason, in this popu- acute ischemic heart disease. Many etiologies exist
lation, it is important to use what is known as the for troponin elevations, and many of them may
10% CV value rather than the 99th percentile, need to be evaluated as part of an evaluation in
which is a level where few, if any, false-positive el- the ED or chest pain center. Two of the more
evations are caused by imprecision. Most would common ones are pulmonary embolism [58,59]
advocate using the 10% CV level (which is usually and congestive heart failure [60,61]; these can
close to the 99th percentile) in all patients with the mimic the presentation of an acute coronary syn-
recognition that in patients who are at high risk drome and both are associated with an adverse
clinically even lower levels may have a prognostic prognosis when troponin elevations occur. In gen-
and therapeutic signicance. In the intermediate eral, the troponin elevations observed are modest.
and lower-risk groups, however, if one uses the Elevations associated with pulmonary embolism
lower level (ie, the 99th percentile), an unaccept- resolve rapidly (by 40 hours) [62]. Elevations in
ably high percentage of patients is apt to have patients who present with congestive heart failure
false positives. The 99th percentile and the 10% persist for a longer period and are found in pa-
CV values for present assays are shown in Table tients with and without coronary artery disease.
1 [21]. Caution is advised because the assays them- Chest pain and elevated troponins in patients
selves and, thus, the values, presented change may be caused by some of these other etiologies
frequently. for elevated troponins, but no overt coronary ar-
Many of the studies that have been done have tery disease angiographically were at accentuated
used high cuto values for troponin in evaluating risk in the Tactics-TIMI-18 trial [63]. The coro-
ED patients. Recently, a study from Johns Hop- nary angiogram may not be, in every instance, to-
kins [57] evaluated patients who had low levels and tally reliable for the detection of acute coronary
patients who had intermediate levels between the artery disease, especially when a delay exists be-
99th percentile and what is known as the ROC de- tween the onset of symptoms and implementation
termined cuto value for MI, a level initially used of the diagnostic procedure. Nonetheless, these
when the assays were novel to match the sensitivity data suggest that the due diligence of any good cli-
of CK-MB. As with almost every other study that nician seeing a patient who presents with an ele-
has been done, a substantial short- and long-term vated troponin in these intermediate and lower-
adverse prognostic eect was observed with even risk groups is not only to determine whether or
minor elevations of troponin (Fig. 6). More of not the patient is at risk for acute ischemic heart
these studies are needed, but they were consonant disease. The physician must also make sure that
with the vast body of clinical information that sug- the elevations are not a result of other causes
gests the elevated troponin values usually have im- that may require dierent types of therapy (see
portant prognostic signicance. later discussion).
460 JAFFE

Extensive evaluations of low risk patients have Patients with renal failure
not been common. In an important study, however,
The most common cause of death in patients
the group from Galveston [64] evaluated a cohort
who present with renal failure is cardiovascular
of over 400 patients who presented with a low-
[65]. Coronary artery disease is common and
risk history and normal or near normal ECG.
events tend to occur during the period when pa-
They selected their inclusion criteria to be associat-
tients have a longer interval between dialysis visits
ed with a 5% to 7% incidence of MI. They then
[65,66]. In patients who present with acute coro-
evaluated the frequency of coronary artery disease
nary syndromes, the interpretation of troponin el-
by oering angiography to all those patients with
evations should not be altered by the presence of
an elevated troponin and enrolling a control group
renal failure [67].
that was selected from similar patients with chest
Other dialysis patients at risk who do not have
pain but without elevated troponins as a compari-
acute coronary syndromes can often also be
son group. In that study, which used a higher tro-
identied by having an elevated troponin [68,69].
ponin than has been advocated by this author,
These patients invariably have pathologic evi-
patients who presented with an elevated troponin
dence of cardiac injury [70]. This group manifests
had a 90% frequency of angiographic coronary ar-
dierences in the frequency of elevations of cTnT
tery disease. In two thirds, it was double or triple
and cTnI [68]. Many more elevations of cTnT
vessel disease. In addition, during follow up, these
occur than cTnI, and this has led some people
individuals had a 33% incidence of ischemic events
to believe that the elevations of cTnT are false
caused by coronary heart disease (Fig. 7). These
positives. With the initial iteration of the assay
data demonstrate the importance of an elevated
this might have been the case but is no longer
troponin in patients who present with chest pain,
a credible explanation. The present iteration of
even those at low risk. Had the investigators used
the assay has been shown not to cross-react with
a lower cuto value, the 10% CV value for exam-
noncardiac isoforms of cTnT [1014], nor does it
ple, it is possible there may well have been a larger
detect the isoforms of cardiac cTnT that can be
number of patients with increases attributable to
re-expressed in response to skeletal muscle injury.
diseases other than acute ischemic heart disease.
The epitopes for detection of cTnI may be lost in
However, it is possible that the frequency of coro-
the milieu of renal failure. A substantial percent-
nary artery disease seen in the control group of
age of patients on dialysis, therefore, have eleva-
23% would have been substantially diminished.
tions of troponin, especially cTnT, posing
Thus, it seems likely that the use of the low cuto
a dicult problem in the ED because the initial
values may help to identify patients who are at
samples in these patients may indicate an eleva-
risk when they present with chest discomfort.
tion. Having a baseline cTnT is helpful. The
No comparable data exist with any other
FDA has approved the use of cTnT for risk strat-
markers in either risk group.
ication in patients who are on dialysis, and it is
recommended that baseline values be available
for comparison. If baseline values are not avail-
able, looking for a change in serial values is help-
ful. Patients who experience rising elevations are
far more likely to have acute events than those
whose values stay the same. This information
does not imply that the latter are false positives
only that they are less apt to be associated with
an acute event. In a group, such as dialysis pa-
tients, where elevations in the absence of acute is-
chemic heart disease are common, this is a helpful
approach to determine which patients require ad-
Fig. 7. Prognostic signicance of troponin elevations in mission or urgent care and which do not. This
patients with normal or near normal electrocardiograms. same principle seems to be the case in patients
From deFilippi CR, Tocchi M, Parmar RJ, et al. Cardiac who seem to be at risk for re-infarction [71]. Initial
troponin T in chest pain unit patients without ischemic guidelines suggested the use of other markers to
electrocardiographic changes: angiographic correlates
help in this area, but it is clear clinically, although
and long-term clinical outcomes. J Am Coll Cardiol
extensive collaborating data have yet to be
2000;35:182734; with permission.
BIOMARKERS IN THE ED AND CPU 461

developed in the literature, that using rising values


are an eective strategy. Thus, dialysis patients Box 1. Elevations of troponins without
who present with rising values are those in need overt ischemic heart disease
of aggressive care. Patients who present with
 Trauma (including contusion; ablation;
more chronic elevations can be evaluated more
pacing; ICD firings, such as atrial
electively assuming no other clinical signs exist
defibrillators, cardioversion,
pointing to acute disease. This evaluation can in-
endomyocardial biopsy, cardiac
clude considerations of all other entities capable
surgery, after-interventionalclosure
of causing elevations in troponin, such as pulmo-
of ASDs)
nary embolism, congestive heart failure, and the
 Congestive heart failuredacute and
like.
chronic
These caveats apply only to patients who are
 Aortic valve disease and HOCM with
on dialysis. Although there are an increased
significant LVH
number of elevations of troponin in patients
 Hypertension
who present with signicant renal dysfunction,
 Hypotension, often with arrhythmias
most seem to be caused by concomitant co-
 Postoperative noncardiac surgery
morbidities and are not synonymous with the
patients who seem to do well
chronic elevations seen in dialysis patients. Ac-
 Renal failure
cordingly, the caveats above should be applied to
 Critically ill patients, especially with
dialysis patients only. The concept that elevations
diabetes, respiratory failure
in troponin with milder forms of renal dysfunc-
 Drug toxicity (eg, adriamycin, 5 FU,
tion can be attributed to this chronic process is at
herceptin, snake venoms)
present unsupported by data. The potential etiol-
 Hypothyroidism
ogies of troponin elevations in dialysis patients
 Coronary vasospasm, including apical
likely are related to the metabolic milieu of
ballooning syndrome
dialysis, concomitant endothelial dysfunction, se-
 Inflammatory diseases (eg,
vere hypertension with LVH, acute ventricular
myocarditis, including Parvovirus
stretch, and various other mechanisms that occur
B19, Kawasaki disease, sarcoid,
in patients who present with renal dysfunction.
smallpox vaccination, or myocardial
Delayed clearance of the troponin by way of the
extension of BE)
kidneys is an unlikely cause of the troponin
 Post-PCI patients who seem to be
elevations.
uncomplicated
Other etiologies for elevated troponins  Pulmonary embolism, severe
pulmonary hypertension
As indicated in earlier discussion and in Box 1,  Sepsis
acute ischemic heart disease is only one reason for  Burns, especially if TBSA greater than
elevations. Many other reasons exist that need to 30%
be considered [26]:  Infiltrative diseases, including
amyloidosis, hemachromatosis,
1. Many of the entities listed involve some
sarcoidosis, and scleroderma
degree of ischemia but it is not ischemia
 Acute neurologic disease, including
necessarily caused by acute coronary
CVA, subarchnoid bleeds
abnormalities. For example, patients who
 Rhabdomyolysis with cardiac injury
present with left ventricular hypertrophy,
 Transplant vasculopathy
whether caused by hypertension, aortic ste-
 Vital exhaustion
nosis, or hypertrophic cardiomyopathy, and
so forth, are known to have an underper-
fused subendocardium. Thus, patients may
be at risk if rapid heart rates, hypotension, which pulmonary embolism albeit in the
or hypertension for elevated troponins as a re- right rather than the left ventricle, induces
sult of injury to the sub-endocardium. When injury.
patients have arrhythmias, a similar mecha- 2. Cardiac toxicity as a result of drugs, snake
nism could be involved. This mechanism is bites, or other toxins in the environment
probably the mechanism for elevations by are certainly capable of causing troponin
462 JAFFE

elevations. Heat shock protein or TNF are or chest pain clinic [27,29,30,72]. The value of this
believed to participate in cardiac injury seen approach is predominantly to exclude infarction
with sepsis. at an earlier point in time because these analytes
3. Myocarditis, acute or chronic, can cause tro- lack cardiac specicity and often are falsely posi-
ponin elevations and a series of patients who tive. Thus, what has been relied on is their nega-
might be confused with those suering with tive predictive value, which is fairly high. A
STEMI has been published. The diagnosis critical issue in this area has been that many of
should be considered in patients initially sus- these studies done evaluating these analytes have
pected of having acute coronary syndromes used insensitive troponin assays or high cuto val-
who have normal coronary arteries. ues despite the use of sensitive assays. Recent data
4. Cardiomyopathies, congestive or inltrative, suggest that as one begins to use cuto values,
and congestive heart failure likely caused by such as the 99th percentile and 10% CV that the
wall stress and reduced subendocardial per- dierence in timing associated with these other
fusion are also capable of leading to elevated laboratory tests becomes more modest than previ-
troponins and have signicant prognostic ously suggested (Fig. 8) [31]. A similar approach
signicance. relying on the negative predictive value of nonele-
5. Critically ill patients. A large body of infor- vated values has been taken with IMA [3234],
mation suggests that elevations in troponin which is a marker of ischemia rather than necro-
are common and highly prognostic in pa- sis. The inuence of more sensitive troponin as-
tients who are critically ill whether in the hos- says on the timing of elevations can also
pital or ED. inuence the amount of time saved with this ana-
lyte. The time saved which is now more modestly
worthwhile compared with the cost and the con-
The use of other markers
founds associated with the frequent false posi-
Many studies previously done suggest that tives. The studies done with IMA have used
because increases in short-acting markers, such more contemporary troponin assays and cuto
as myoglobin or fatty acid binding protein, or in values [3234] but continuing issues exist concern-
the past, isoforms of CK increase more rapidly ing the proper cuto values for the test, analytic
than troponin that they may save time in the ED issues related to pH and lactate [35] and concern

Fig. 8. Sensitivity of cTnI, CK-MB and myoglobin over time using modern prognostic cuto values. Note that 0.7 ng/ml
is the 99th percentile for the assay used. From Eggers KM, Oldgren J, Nordenskjold A, et al. Diagnostic value of serial
measurement of cardiac markers in patients with chest pain: limited value of adding myoglobin to troponin I for exclu-
sion of myocardial infarction. Am Heart J 2004;148:57481;with permission.
BIOMARKERS IN THE ED AND CPU 463

about what should be done if a patient has an el- [13] Ricchiuti V, Voss EM, Ney A, et al. Cardiac tro-
evated IMA without an elevated troponin. Ac- ponin T isoforms expressed in renal diseased skel-
cordingly, a much simpler, more cost-eective etal muscle will not cause false-positive results by
paradigm would be to measure troponins on two the second generation cardiac troponin T assay
by Boehringer Mannheim. Clin Chem 1998;44:
occasions: (1) when the patient rst presents and
191924.
(2) 6 hours later or earlier if the onset of symp- [14] Haller C, Zehelein J, Remppis A, et al. Cardiac tro-
toms is totally clear. Additional data likely will ponin T in patients with end-stage renal disease: ab-
shorten this time somewhat also. This paradigm sence of expression in truncal skeletal muscle. Clin
would be a highly sensitive, highly specic, and Chem 1998;44:9308.
cost-eective way of triage but requires thoughtful [15] Adams JE III, Schechtman KB, Landt Y, et al. Com-
consideration of the etiologies of the potential parable detection of acute myocardial infarction by
troponin elevations, some clinical judgment, and creatine kinase MB isoenzyme and cardiac troponin
a knowledge of the assays that are used locally. I. Clin Chem 1994;40:12915.
[16] Katus HA, Remppis A, Scheold T, et al. Intracellu-
lar compartmentation of cardiac troponin T and its
References release kinetics in patients with reperfused and non-
reperfused myocardial infarction. Am J Cardiol
[1] Takeda S, Yamashita A, Maeda K, et al. Structure 1991;67:13607.
of the core domain of human cardiac troponin in [17] Tanaka H, Abe S, Yamashita T, et al. Serum levels
the Ca(2 )-saturated form. Nature 2003;424:3541. of cardiac troponin I and troponin T in estimating
[2] Dhoot GK, Gell PG, Perry SV. The localization of myocardial infarct size soon after reperfusion.
the dierent forms of troponin I in skeletal and car- Coron Artery Dis 1997;8:4339.
diac muscle cells. Exp Cell Res 1978;117:35770. [18] Jae AS, Landt Y, Parvin CA, et al. Comparative
[3] Dhoot GK, Perry SV. Distribution of polymorphic sensitivity of cardiac troponin I and lactate dehydro-
forms of troponin components and tropomyosin in genase isoenzymes for diagnosing acute myocardial
skeletal muscle. Nature 1979;278:7148. infarction. Clin Chem 1996;42:17706.
[4] Perry SV. Troponin I: inhibitor or facilitator. Mol [19] Katus HA, Remppis A, Looser S, et al. Enzyme
Cell Biochem 1999;190:932. linked immuno assay of cardiac troponin T for the
[5] Katrukha A. Antibody selection strategies in cardiac detection of acute myocardial infarction in patients.
troponin assay. In: Wu AHB, editor. Cardiac J Mol Cell Cardiol 1989;21:134953.
markers. 2nd edition. Totowa (NJ): Humana Press [20] Panteghini M, Gerhardt W, Apple FS, et al. Quality
Inc.; 2003. p. 17385. specications for cardiac troponin assays. Clin
[6] Larue C, Defacque-Lacquement H, Calzolari C, Chem Lab Med 2001;39:1759.
et al. New monoclonal antibodies as probes for hu- [21] Panteghini M, Pagani F, Yeo KT, et al. Evaluation
man cardiac troponin I: epitopic analysis with syn- of imprecision for cardiac troponin assays at low-
thetic peptides. Mol Immunol 1992;29:2718. range concentrations. Clin Chem 2004;50:32732.
[7] Toyota N, Shimada Y. Dierentiation of troponin [22] Jae A. Caveat emptor. Am J Med 2003;115:2414.
in cardiac and skeletal muscles in chicken embryos [23] Alpert JS, Thygesen K, Antman E, et al. Myocardial
as studied by immunouorescence microscopy. infarction redeneda consensus document of The
J Cell Biol 1981;91:497504. Joint European Society of Cardiology/American
[8] Bodor GS, Portereld D, Voss EM, et al. Cardiac College of Cardiology Committee for the redeni-
troponin-I is not expressed in fetal and healthy or tion of myocardial infarction. J Am Coll Cardiol
diseased adult human skeletal muscle tissue. Clin 2000;36:95969.
Chem 1995;41:17105. [24] Jae AS, Ravkilde J, Roberts R, et al. Its time for
[9] Perry SV. Troponin T genetics, properties and func- a change to a troponin standard. Circulation 2000;
tion. J Muscle Res Cell Motil 1998;19:575602. 102:121620.
[10] Muller-Bardor M, Hallermayer K, Schroder A, [25] Apple FS, Wu AH, Jae AS. European Society of
et al. Improved troponin T ELISA specic for car- Cardiology and American College of Cardiology
diac troponin T isoform: assay development and guidelines for redenition of myocardial infarction:
analytical and clinical validation. Clin Chem 1997; how to use existing assays clinically and for clinical
43:45866. trials. Am Heart J 2002;144:9816.
[11] Bodor GS, Survant L, Voss EM, et al. Cardiac tro- [26] Jae AS. Elevations in cardiac troponin measure-
ponin T composition in normal and regenerating hu- ments: false false-positives. Cardiovasc Toxicol
man skeletal muscle. Clin Chem 1997;43:47684. 2001;1(2):8792.
[12] Katus HA, Looser S, Hallermayer K, et al. Develop- [27] Adams JE III, Abendschein DR, Jae AS. Biochem-
ment and in vitro characterization of a new immuno- ical markers of myocardial injury. In: MB creatine
assay of cardiac troponin T. Clin Chem 1992;38: kinase the choice for the 1990s [review]? Circulation
38693. 1993;88:75063.
464 JAFFE

[28] Adams JE III, Bodor GS, Davila-Roman VG, et al. [41] Hamm CW, Ravkilde J, Gerhardt W, et al. The
Cardiac troponin I. A marker with high specicity prognostic value of serum troponin T in unstable
for cardiac injury. Circulation 1993;88:1016. angina. N Engl J Med 1992;327:14650.
[29] Newby LK, Storrow AB, Gibler WB, et al. Bedside [42] Antman EM, Tanasijevic MJ, Thompson B, et al.
multimarker testing for risk stratication in chest Cardiac-specic troponin I levels to predict the risk
pain units: the chest pain evaluation by creatine of mortality in patients with acute coronary syn-
kinase-MB, myoglobin, and troponin I (CHECK- dromes. N Engl J Med 1996;335:13429.
MATE) study. Circulation 2001;103:18327. [43] Lindahl B, Venge P, Wallentin L. Relation between
[30] McCord J, Nowak RM, McCullough PA, et al. troponin T and the risk of subsequent cardiac events
Ninety-minute exclusion of acute myocardial infarc- in unstable coronary artery disease. The FRISC
tion by use of quantitative point-of-care testing of study group. Circulation 1996;93:16517.
myoglobin and troponin I. Circulation 2001;104: [44] Galvani M, Ottani F, Ferrini D, et al. Prognostic in-
14838. uence of elevated values of cardiac troponin I in
[31] Eggers KM, Oldgren J, Nordenskjold A, et al. patients with unstable angina. Circulation 1997;95:
Diagnostic value of serial measurement of cardiac 20539.
markers in patients with chest pain: limited value [45] Ottani F, Galvani M, Nicolini FA, et al. Elevated
of adding myoglobin to troponin I for exclusion cardiac troponin levels predict the risk of adverse
of myocardial infarction. Am Heart J 2004;148: outcome in patients with acute coronary syndromes.
57481. Am Heart J 2000;140:91727.
[32] Bar-Or D, Lau E, Winkler JV. Novel assay for [46] Heeschen C, van Den Brand MJ, Hamm CW, et al.
cobalt-albumin binding and its potential as a mar- Angiographic ndings in patients with refractory
ker for myocardial ischemia-a preliminary report. unstable angina according to troponin T status. Cir-
J Emerg Med 2000;19:3115. culation 1999;100:150914.
[33] Sinha MK, Gaze DC, Tippins JR, et al. Ischemia [47] Morrow DA, Antman EM, Tanasijevic M, et al.
modied albumin is a sensitive marker of myocardial Cardiac troponin I for stratication of early out-
ischemia after percutaneous coronary intervention. comes and the ecacy of enoxaparin in unstable
Circulation 2003;107:24035. angina: a TIMI-11B substudy. J Am Coll Cardiol
[34] Christenson RH, Duh SH, Sanhai WR, et al. Char- 2000;36:18127.
acteristics of an albumin cobalt binding test for [48] Newby LK, Ohman EM, Christenson RH, et al.
assessment of acute coronary syndrome patients: Benet of glycoprotein IIb/IIIa inhibition in patients
a multicentric study. Clin Chem 2001;47:46470. with acute coronary syndromes and troponin t-
[35] Zapico-Muniz E, Santalo-Bel M, Merce-Muntanola positive status: the paragon-B troponin T substudy.
J, et al. Ischemia-modied albumin during skeletal Circulation 2001;103:28916.
muscle ischemia. Clin Chem 2004;50:10635. [49] Braunwald E, Antman EM, Beasley JW, et al. Amer-
[36] Ohman EM, Armstrong PW, Christenson RH, et al. ican College of Cardiology. American Heart Associ-
Cardiac troponin T levels for risk stratication in ation. Committee on the Management of Patients
acute myocardial ischemia. GUSTO IIA Investiga- With Unstable Angina. ACC/AHA 2002 guideline
tors. N Engl J Med 1996;335:133341. update for the management of patients with unstable
[37] Matetzky S, Sharir T, Domingo M, et al. Elevated angina and non-ST-segment elevation myocardial
troponin I level on admission is associated with ad- infarction summary article: a report of the American
verse outcome of primary angioplasty in acute myo- College of Cardiology/American Heart Association
cardial infarction. Circulation 2000;102:16116. task force on practice guidelines (Committee on the
[38] Ohman EM, Armstrong PW, White HD, et al. Risk Management of Patients With Unstable Angina).
stratication with a point-of-care cardiac troponin T J Am Coll Cardiol 2002;40:136674.
test in acute myocardial infarction. GUSTO III [50] Yusuf S, Zhao F, Mehta SR, et al. Eects of clopi-
Investigators. Global use of strategies to open oc- dogrel in addition to aspirin in patients with acute
cluded coronary arteries. Am J Cardiol 1999;84: coronary syndromes without ST-segment elevation.
12816. N Engl J Med 2001;345:494502.
[39] Giannitsis E, Muller-Bardor M, Lehrke S, et al. [51] Aviles RJ, Wright RS, Aviles JM, et al. Long-term
Admission troponin T level predicts clinical out- prognosis of patients with clinical unstable angina
comes, TIMI ow, and myocardial tissue perfusion pectoris without elevation of creatine kinase but
after primary percutaneous intervention for acute with elevation of cardiac troponin I levels. Am J Car-
ST-segment elevation myocardial infarction. Circu- diol 2002;90:8758.
lation 2001;104:6305. [52] Invasive compared with non-invasive treatment in
[40] Giannitsis E, Lehrke S, Wiegand UK, et al. Risk unstable coronary-artery disease: FRISC II prospec-
stratication in patients with inferior acute myocar- tive randomised multicentre study. FRagmin and
dial infarction treated by percutaneous coronary Fast Revascularisation during InStability in Coro-
interventions: the role of admission troponin T. Cir- nary artery disease Investigators. Lancet 1999;354:
culation 2000;102:203844. 70815.
BIOMARKERS IN THE ED AND CPU 465

[53] Morrow DA, Cannon CP, Rifai N, et al. Ability of survivors of conrmed severe pulmonary embolism.
minor elevations of troponins I and T to predict ben- Clin Chem 2002;48:6735.
et from an early invasive strategy in patients with [63] Dokainish H, Pillai M, Murphy SA, et al. TAC-
unstable angina and non-ST elevation myocardial TICS-TIMI-18 Investigators. Prognostic impli-
infarction: results from a randomized trial. JAMA cations of elevated troponin in patients with
2001;286:240512. suspected acute coronary syndrome but no critical
[54] Wiviott SD, Cannon CP, Morrow DA, et al. Dier- epicardial coronary disease: a TACTICS-TIMI-18
ential expression of cardiac biomarkers by gender in substudy. J Am Coll Cardiol 2005;45:1924.
patients with unstable angina/non-ST-elevation [64] deFilippi CR, Tocchi M, Parmar RJ, et al. Cardiac
myocardial infarction: a TACTICS-TIMI 18 (Treat troponin T in chest pain unit patients without ische-
Angina with Aggrastat and determine Cost of Ther- mic electrocardiographic changes: angiographic cor-
apy with an Invasive or Conservative Strategy- relates and long-term clinical outcomes. J Am Coll
Thrombolysis In Myocardial Infarction 18) Cardiol 2000;35:182734.
substudy. Circulation 2004;109:5806. [65] Herzog CA, Ma JZ, Collins AJ. Poor long-term sur-
[55] Hamm CW, Goldmann BU, Heeschen C, et al. vival after acute myocardial infarction among
Emergency room triage of patients with acute chest patients on long-term dialysis. N Engl J Med 1998;
pain by means of rapid testing for cardiac troponin 339:799805.
T or troponin I. N Engl J Med 1997;337:164853. [66] Bleyer AJ, Russell GB, Satko SG. Sudden and car-
[56] James S, Armstrong P, Cali R, et al. Troponin T diac death rates in hemodialysis patients. Kidney
levels and risk of 30-day outcomes in patients with Int 1999;55:15539.
the acute coronary syndrome: prospective verica- [67] Aviles RJ, Askari AT, Lindahl B, et al. Troponin T
tion in the GUSTO-IV trial. Am J Med 2003;115: levels in patients with acute coronary syndromes,
17884. with or without renal dysfunction. N Engl J Med
[57] Henrikson CA, Howell EE, Bush DE, et al. Prognos- 2002;346:204752.
tic usefulness of marginal troponin T elevation. Am [68] Apple FS, Murakami MM, Pearce LA, et al. Predic-
J Cardiol 2004;93:2759. tive value of cardiac troponin I and T for subsequent
[58] Konstantinides S, Geibel A, Olschewski M, et al. death in end-stage renal disease. Circulation 2002;
Importance of cardiac troponins I and T in risk 106:29415.
stratication of patients with acute pulmonary [69] deFilippi C, Wasserman S, Rosanio S, et al. Cardiac
embolism. Circulation 2002;106:12638. troponin T and C-reactive protein for predicting
[59] La Vecchia L, Ottani F, Favero L, et al. Increased prognosis, coronary atherosclerosis, and cardiomy-
cardiac troponin I on admission predicts in-hospital opathy in patients undergoing long-term hemodialy-
mortality in acute pulmonary embolism. Heart 2004; sis. JAMA 2003;290:3539.
90:6337. [70] Ooi DS, Isotalo PA, Veinot JP. Correlation of ante-
[60] Missov E, Calzolari C, Pau B. Circulating cardiac mortem serum creatine kinase, creatine kinase-MB,
troponin I in severe congestive heart failure. Circula- troponin I, and troponin T with cardiac pathology.
tion 1997;96:29538. Clin Chem 2000;46:33844.
[61] Horwich TB, Patel J, MacLellan WR, et al. Cardiac [71] Apple FS, Murakami MM. Cardiac troponin and
troponin I is associated with impaired hemodynam- creatine kinase MB monitoring during in-hospital
ics, progressive left ventricular dysfunction, and in- myocardial reinfarction. Clin Chem 2005;61:4603.
creased mortality rates in advanced heart failure. [72] Tanaka T, Hirota Y, Sohmiya K-I, et al. Serum and
Circulation 2003;108:8338. urinary human heart fatty acid binding protein in
[62] Muller-Bardor M, Weidtmann B, Giannitsis E, acute myocardial infarction. Clin Biochem 1991;24:
et al. Release kinetics of cardiac troponin T in 195201.
Cardiol Clin 23 (2005) 467490

Point-of-Care Testing and Cardiac Biomarkers:


The Standard of Care and Vision
for Chest Pain Centers
Gerald J. Kost, MD, PhD, MS, FACBa,b,*, Nam K. Tran, BSa
a
Point-of-Care Testing Center for Teaching and Research, Department of Pathology and Laboratory Medicine,
UCD Health System, School of Medicine, University of California, Davis, CA, USA
b
Aliate Faculty, Chulalongkorn University, Bangkok, Thailand

Point-of-care testing (POCT) is dened as article focuses on cardiac biomarker POCT for
testing at or near the site of patient care [1]. chest pain centers and emergency medicine.
POCT decreases therapeutic turnaround time POC test clusters encompass electrolytes (eg,
(TTAT), increases clinical eciency, and im- K, Ca), metabolites (glucose, lactate), func-
proves medical and economic outcomes [2,3]. tion monitors (acid-base, hemostasis), and several
TTAT (Fig. 1) represents the time from test order- biomarkers [2]. Hospital penetration of POC car-
ing to patient treatment [2]. POC technologies diac biomarker testing increased from 4% in
have become ubiquitous in the United States, 2001 to 12% in 2004, a threefold change [13,14].
and, therefore, so has the potential for speed, con- Now, with a strong annual growth rate, cardiac
venience, and satisfaction, strong advantages for biomarker POCT represents one of the most
physicians, nurses, and patients in chest pain cen- rapidly expanding areas in clinical diagnostics.
ters. POCT is applied most benecially through The growth rate, estimated as 10% to 20%
the collaborative teamwork of clinicians and labo- [13,15], could double use quickly. Sales are ex-
ratorians who use integrative strategies, perfor- pected to grow from about $200 million in 2003
mance maps, clinical algorithms, and care paths to around $400 to $500 million in 2008, an increase
(critical pathways) [25]. For example, clinical in- of over 100% in ve years. The current market
vestigators [69] have shown that on-site inte- trend is driven by the following: (a) an evolving
gration of testing for cardiac injury markers standard of care for the timeliness of evaluation
(myoglobin, creatinine kinase MB isoenzyme and treatment in chest pain centers and EDs; (b)
[CK-MB], and cardiac troponin I [cTnI]) in accel- motivation to produce cost-eective medical solu-
erated diagnostic algorithms produces eective tions for acute coronary syndromes (ACS) and
screening, less hospitalization, and substantial other cardiovascular conditions, such as congestive
savings. Chest pain centers [10], which now total heart failure; (c) need to reduce medicolegal risk in
over 150 accredited in the United States [11], in- emergency department (ED) settings where acute
corporate similar types of protocol-driven perfor- myocardial infarction (AMI) is missed too fre-
mance enhancements. This optimization allows quently; (d) physician and nurse demands for rapid
chest pain centers to improve patient evaluation, response testing and consistently fast test results,
treatment, survival, and discharge [12]. This day and night; (e) emphasis on the total value of di-
agnostic information (versus piecemeal costs of in-
dividual tests); (f) smaller, smarter, faster, and
cheaper portable devices with improved technolo-
* Corresponding author. 506 Citadel Drive, Davis, gies for POCT, such as critical care proling, faster
CA 95616. immunotesting, and integrated multimarker index-
E-mail address: gjkost@ucdavis.edu (G.J. Kost). ing; (g) a continuing paradigm shift to the hybrid
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.005 cardiology.theclinics.com
468 KOST & TRAN

biomarker POCT will help maximize future


benets.

Status of cardiac biomarker point-of-care testing


In 2004, an Enterprise Analysis Corporation
[13] telephone interview survey of 493 United
States hospitals with bed size of at least 150
showed that cardiac biomarkers were evaluated
for use more frequently but rejected more com-
monly than other types of POCT, which implies
Fig. 1. Therapeutic turn around time (TTAT) is time that devices oered fall short of meeting needs
from test ordering to appropriate treatment (Rx). If [14]. Of hospitals deciding not to install a new
blood specimens are obtained rst or before verbal or- POCT system, 75% (82 of 110) evaluated cardiac
ders are documented after, as may occur in critical situa- biomarkers [13]. Reasons for not installing in-
tions, TTAT starts with collection. POCT and whole- cluded: (a) lack of improved turnaround time
blood analysis (WBA) shorten TTAT. The interval (TAT), (b) high costs, (c) inaccuracy, (d) not
from Rx to outcome typically also decreases, thereby en- user friendly for nonlaboratorians, and (e) labora-
hancing eciency of patient care and usually improving
tory unwillingness to relinquish test control.
medical and economic outcomes.
Nonetheless, as noted above, between 2001 and
2004, hospital penetration of POC cardiac marker
testing by discipline trebled with smaller hospitals
laboratory [2] where laboratorians work more (150 to 299 beds) most frequently adopting
closely on-site with clinicians who perform bedside POCT. Sites where hospitals used cardiac bio-
testing; (h) increasing penetration of POC cardiac marker POCT included: (a) ED (56%), (b) core
biomarkers in a niche hospital market previously laboratory (52%), (c) outpatient area (8%), (d)
characterized by slow growth; and (i) physician stat laboratory (6%), (e) cardiac catheterization
expectations, patient empowerment, health con- laboratory (3%), (f) intensive care unit (2%),
sciousness, and early detection extending even to and (g) operating room (2%)dsum of 129% be-
the home because too many patients succumb to cause of multiple sites. Overall, Biosite Diagnos-
AMI before reaching the hospital. tics (San Diego, CA) led the market in cardiac
An important goal of this article is to help biomarkers with B-type natruiretic peptide
optimize the guidelines- and evidence-based clin- (BNP) on the Triage Cardiac Panel platform.
ical practice of POCT [16] in chest pain centers Note that EDs holding their own Clinical Labora-
through systematic planning and implementing tory Improvements Act licenses were not con-
of POC cardiac biomarker testing. Themes in- tacted in the survey; therefore, the frequency of
clude: (a) POCT should be applied broadly to distributed testing in EDs may exceed 56%.
reduce medical risk; (b) POCT for cardiac bio- According to the United States survey [13], car-
markers is evolving into the standard of care; diac biomarker tests performed included: (a) car-
and (c) whole-blood analysis (WBA) time ulti- diac troponin (cTn) alone (24% of hospitals
mately determines the minimum time interval for using cardiac biomarker POCT); (b) cTn, CK-
rapid response testing. This article delineates MB, and myoglobin (67%); and (c) BNP (74%).
POC cardiac test clusters; describes clinical prior- Forty-ve percent reported they would switch
ities for their application; summarizes handheld, BNP to the main laboratory if possible. Average
portable, and transportable instruments and POC cTn test volume (5227) represented about
test kits; and identies future needs that must be twice that of BNP (2766), but all types constituted
fullled to serve acute care patients eciently 20% or less of total cardiac biomarker test volume
and ecaciously. Current heavy investment in in 58% of institutions. Overall, hospitals stream-
commercial development raises unique opportuni- lined the number of POCT systems (average about
ties for creative interdigitation of POCT, bedside three) under pressures for training, quality con-
decision making, and evidence-based therapy in trol, maintenance, integration, and oversight. Of
chest pain centers. Active leadership participation those hospitals projecting rapid or very rapid
during this dynamic phase of expansion of cardiac growth, speed and improved technology were cited
CARDIAC BIOMARKER POCT 469

most frequently. Thus, the survey showed that Several areas for potential improvement in the
high accuracy, speed, improved technology, user POC cardiac biomarker repertoire are identied
friendliness, broad test menu, cost-eectiveness, in Table 1. Miniaturized bedside immunochemis-
quality control, and information integration, as try has just passed its infancy and integrated mul-
well as faster quantitative analysis, represent key timarker test clusters have appeared only recently
design specications to be fullled by cardiac bio- on portable and handheld devices. Despite over
marker POC devices. Other important features one decade of development [44], WBA times for
include built-in data management and quality immunochemistry remain prolonged, in the order
control, plus error alerting [17]. Use of POCT of 15 minutes. Following the precedent set in
does not excuse inaccuracy or errors. Clinicians other clinical applications, such as biosensor-based
and laboratorians should insist on accuracy bedside glucose testing, analysis time should be re-
and reliability irrespective of where testing is duced to just a few minutes while analytic sensitiv-
performed. ity (minimum detection level) should be improved.
These potentially conicting goals may justify the
use of novel nano- or microuidics, microarray,
Technologies for cardiac biomarker or optical approaches heretofore not commonly
point-of-care testing present on POC devices. Complex manufacturing
and funding hurdles stall eorts to combine car-
Instruments, decision levels, and applications
diac biomarkers with other critical care analytes,
Handheld, portable, and transportable instru- such as electrolytes (eg, K and Ca), blood
ments and devices available for cardiac biomarker gases, and pH. Currently, these analytes can be
POCT are summarized in Table 1 [1838]. Test clus- measured with exchangeable cartridge-based sys-
ters, methods, analysis times, sample types, and tems (eg, i-STAT) but should be fully integrated
other pertinent information are shown. Disposable on other instruments for the evaluation, diagnosis,
qualitative test kits are shown also. For additional and monitoring of critically ill patients.
details of POC analytic principles, see Tang and col- The inconsistency in qualitative and quantita-
leagues [39] and for an earlier analysis of technol- tive decision levels is emphasized in Table 2. In
ogy, see the University HealthSystem Consortium this phase of rapid growth, standardization of car-
report by Cummings and colleagues [40]. Note diac biomarker measurements and decision levels
that several of the devices and test kits in Table 1 should occupy a position of high priority. Stan-
may be used in near-patient laboratories or at the dardization, which demands consideration of sub-
bedside. Before implementation, operators should ject reference levels and assay characteristics (eg,
check with manufacturers regarding regulatory epitopes, antibodies, and matrices, all producing
constraints, decision thresholds, and test classica- in the case of cTnI, varying specicities for dier-
tions. Currently, none of the tests listed qualify for ent forms released), improves the consistency of
so-called waived status under federal statutes gov- diagnostic interpretation, reduces the potential
erning diagnostic testing [41]. Thus, testing must for errors, and allows physicians to follow trends
be performed by licensed, certied, and validated if dierent assays are used for testing after patients
personnel in accredited facilities with attention are admitted. Inadequate computer interfacing of
to required daily quality control, and should be devices and lack of critical results alerting repre-
supervised by POC coordinators who assure sent two additional weaknesses. Full integration
high-quality, satisfy inspection requirements, and with hospital computerized systems is required
incorporate prociency testing (objective external for settings using an electronic medical record.
performance review). Nurses and POCT coordina- As noted in later discussion of risk, missed diagno-
tors and, often, experienced medical technologists, ses resulting from poor communication and other
have become indispensable partners stitching these root causes can result in signicant nancial pen-
aspects of POCT together and maintaining the fab- alties. Examples of special situations where car-
ric of excellence in testing irrespective of where it is diac biomarker POCT is being developed or can
performed [42,43]. Chest pain centers should not at- help facilitate unusual clinical problem solving
tempt to implement and manage POCT without the are presented in Table 3 [4558]. Generally, per
assistance of an experienced POCT coordinator test cost for POCT exceeds that on larger main-
who provides valuable liaison with laboratory med- frame chemistry analyzers found in clinical labo-
icine and essential continuity of POCT throughout ratories. Despite high costs, POCT can improve
the hospital. overall cost-eectiveness [3,69], discussed below.
470
Table 1
Cardiac Biomarker POCT
Test/Device name Biomarker(s) Device characteristics Reference notes
Cardiac Reader
 Cardiac T cTnT Sample: 150 mL heparinized or Cardiac Reader interprets Cardiac T/M/D rapid assay strips.
 Cardiac M Myoglobin EDTA WB Comparison between the rapid test and a established
 Cardiac D D-dimer Analysis time: 12 min, TnT; 8 min, laboratory-based method showed sucient agreement of
 NT-proBNP (Portable) NT-proBNPa myoglobin; and 8 min, D-dimer results with a correlation of r 0.89 for troponin T and

KOST & TRAN


Roche Diagnostics Method: Immunoassay and r 0.912 for myoglobin [18]. The Cardiac T exhibited
photosensor Sn 100% and Sp 76.2% for MI [19]. See also [20].
TropT sensitive cTnT (QUAL) Sample: 150 mL heparinized Out of 34 patients who present with AMI, the TropT sensitive
(Handheld, disposable) or EDTA WB test showed positive quantitative results for 20 patients versus
Roche Diagnostics Analysis time: 15 min 29 patients when the Cardiac Reader was used [21]. At
Indianapolis, IN Method: Immunoassay 3 hours after onset of symptoms for AMI, Sn 50%,
www.roche.com Sp 96.3%, PPV 80%, and NPV 86.7% for AMI [22].
Cardiac STATus Myo cTnI Sample: 150 mL heparinized WB, Qualitative results from CK-MB and myoglobin test cards were
(Handheld, disposable) Myo cTnI CK-MB heparinized plasma or serum comparable diagnostically to quantitative results from the
Spectral Diagnostics (test clusters) Analysis time: 15 min Ciba Corning ACS-180 and Dade Stratus IIntellect [23]. cTnI
Toronto, Ontario, CAN cTnI Method: Lateral ow based tests showed 98% concordance with ELISA [24]. In
www.spectraldx.com CK-MB immunoassay comparison with Roche TropT, the Cardiac STATus proved
more eective in early diagnosis of AMI [25]. The Cardiac
STATus had a sensitivity of 96% for the detection of AMI
within 3 hours of presentation [26].
i-Lynx NA The system consists of two interfaced The i-Lynx is a connectivity solution with full QC functionality
(Handheld, portable) components; a handheld electronic in open architecture, including system maintenance, user ID,
Spectral Diagnostics device to capture and analyze an image QC log, lockout, passcode protection, universal barcode, and
Toronto, Ontario, CAN from the current Cardiac STATus rapid connectivity by way of TCP-IP for POC Coordinators (Chris
www.spectraldx.com test strip, and a local area network Wayne, Spectral Diagnostics, e-mail communication.)
compatible docking station to
automate data capture communication
Cholestech LDX CRP Sample: 50 mL WB or WB and plasma samples showed a concordance rate of 94%
(Portable) 4050 mL plasma and 90%, respectively, with results from Dade Behring
Cholestech Corp. Analysis time: 7 min BN100 [27].
Hayward, CA Other analytes: Total cholesterol, HDL,
www.cholestech.com triglycerides, ALT, and glucose.
Method: Reectance photometry,
immunoassay
i-STAT 1 and PCA cTnI Sample: 16 mL heparnized WB or plasma cTnI test performance proved insensitive to hematocrits
(Handheld) Analysis time: 10 min, cTnI between 0 to 65%, but higher hematocrits produced
i-STAT Other analytes: Na, K, Cl, Ca2, glucose, imprecision [28].
East Windsor, NJ creatinine, lactate, BUN, Hct, Hb, pH,
www.i-stat.com pCO2, pO2, ACT, and PT
Method: ELISA
LifeSign MI cTnI myo Sample: 120 mL WB, plasma, or serum Each test cluster is equipped with a built in control for quality
(Handheld, disposable) CK-MB myo Analysis time: 15 min assurance [29].
PBM CK-MB myo cTnI Method: Chromatographic immunoassay
Princeton, NJ (test clusters)

CARDIAC BIOMARKER POCT


www.pbmc.com MLC-1b
NT-proBNPb
NycoCard Reader II CRP Sample: 5 mL heparinized/EDTA/ NycoCard Reader II interprets CRP and D-dimer test strips.
(Portable) D-dimer citrated WB or serum (dilution
Axis-Shield to 50 mL required)
Oslo, Norway Analysis time: 3 min
www.axis-shield.co.uk Other analytes: HbA1c, microalbumin
Method: Reectometric immunoassay
Nexus Dx cTnI myoa Sample: 120 mL WB, plasma or serum NT-proBNP test was approved in Europe mid-2004 and is
(Handheld, disposable) cTnI myo CK-MBa for cTnI pending FDA approval in the United States. Other tests are
Syn X Pharma Inc. (test clusters) Analysis time: 15 min, cTnI available only in Canada and Europe [30].
Toronto, CAN cTnIa Method: Chromatographic immunoassay
www.synxpharma.com NT-proBNPa,b
One-Step Marker Test cTnI myo CK-MBa Sample: 120 mL WB, plasma, or serum Device is only available outside the United States. Quality
(Handheld, disposable) (test cluster) Analysis time: 15 min controls are built into the test card.
ACON Labs Method: Chromatographic immunoassay
San Diego, CA
www.aconlabs.com
(continued on next page)

471
472
Table 1 (continued )
Test/Device name Biomarker(s) Device characteristics Reference notes
RAMP Reader Myoglobin Sample: 70 mL WB At 95% CI, Sn (cTnI) 90%, Sp (cTnI) 86%,
(Portable) CK-MB Analysis time: w15 min Sn (CK-MB) 59%, and Sp (CK-MB) 90% for AMI [31].
Response Biomedical cTnI Other analytes: WNV antigen
British Columbia, CAN Method: Fluorescence-based
www.responsebio.com immunoassay
RAPICHEKb H-FABP Sample: 150 mL WB H-FABP panel test achieved 98.4% agreement with H-FABP
(Handheld, disposable) Analysis time: 15 min mass concentration ELISA [22,32]. At 3 hours following
Dainippon Pharmaceuticals Method: Chromatographic onset of symptoms of AMI, RAPICHEK had Sn 100%,
Suita City, Osaka, Japan immunoassay Sp 63%, PPV 44%, and NPV 100%. RAPICHEK
www.dainippon-pharm.co.jp versus TropT sensitivity were 100% versus 50%, respectively.
Stratus CS STAT CK-MB Sample: 2.7 mL heparainzed WB or plasma Device correlated well with cTnI tests performed in a central
(Transportable NPT) Myoglobin Analysis time: 14 min for rst result, 4 min laboratory; imprecision of troponin method has a CV of less
Dade Behring cTnI for each additional result than 10% at the 99th percentile of the reference population
Deereld, IL D-dimera Other analytes: b-hCG [33,34].

KOST & TRAN


www.dadebehring.com Method: Fluorescence-based immunoassay
Triage Cardiac Panel CK-MB myo cTnI Sample: 225 mL EDTA WB or plasma Triage MeterPlus interprets Triage Cardiac Panels, includes
(Portable) cTnI CK-MB myo Analysis time: w15 min built-in quality controls, and provides simultaneous analyses
Biosite Diagnostics BNP (test clusters) Other analytes: acetaminophen, of multiple analytes. Study showed Triage Cardiac Panel
San Diego, CA D-dimer amphetamines, methamphetamines, comparable to established methods for detection of AMI
www.biosite.com cocaine, opiates, phencyclidine, [35]. Sn for cTnI, CK-MB, and Myo were 98%, 95%, and
tetrahydrocannabinol, barbiturates, 81% respectively, and the Sp were 100%, 91%, and 92%
benzodiazepines, propoxyphen, respectively for AMI [35,36].
tricyclic antidepressants
Method: Fluorescence-based immunoassay
TBA Sample: ngerstick WB A faster quantitative ngerstick-based test strip using whole
Analysis time: 25 min blood is being developed (Julie Doyle, M.D., Biosite
Method: under development Diagnostics, e-mail communication.).
Abbreviations: ACT, activated clotting time; ALT, alanine aminotransferase; b-hCG, b-human chorionic gonadotropin; BUN, blood urea nitrogen; ELISA, enzyme-linked
immunoabsorbant assay; F, female; Hb, hemoglobin; Hct; hematocrit; HDL, high density lipoprotein; H-FABP, human-type fatty acid binding protein; M, male; MLC-1,
myosin light chain-1; myo, myoglobin; NA, not available; NPT, near-patient testing; NPV, negative predictive value; PPV, positive predictive value; PT, prothombin time;
QC, quality control; QUAL, qualitative; Sn, sensitivity; Sp, specicity; TBA, to be announced; TnI, troponin I; WB, whole blood; and WNV, West Nile Virus.
Notes: a) Under development, b) Not available in US, ) Indicates test cluster. Methods were provided by manufacturers or their web sites. Reference 20 cited for com-
pleteness. No current information available for the First Medical Alpha Dx system [37,38]. Chemometrics data given for emergency room or early diagnostic applications. See
Jae chapter for general information of sensitivity, specicity, and predictive values.
CARDIAC BIOMARKER POCT 473

Table 2
Clinical decision levels for cardiac biomarker POCT
Biomarker Device(s) Qualitative Quantitative Decision level
BNP Triage Cardiac Panel X 100 pg/mL
CK-MB Cardiac STATus X 5.0 ng/mL
LifeSign MI X 5.0 ng/mL
Nexus Dxa,b X Under development
One-Step Marker Test X 5.0 ng/mL
RAMP Reader X 5.0 ng/mL
Stratus CS STAT X 3.5 ng/mL
Triage Cardiac Panel X 4.3 ng/mL
CRP Cholestech LDX X Under development
NycoCard CRP X Not available
cTnI Cardiac STATus X 0.5 ng/mL
i-STAT 1 and PCA X 0.1 ng/mL
LifeSign MI X 1.5 ng/mL
One-Step Marker Test X 0.5 ng/mL
RAMP Reader X 0.12 ng/mL
Stratus CS STAT X 0.06 ng/mL
Triage Cardiac Panel X 0.4 ng/mL
Nexus Dxa,b X Under development
cTnT Cardiac Reader X 0.1 ng/mL
TropT Sensitive X 0.1 ng/mL
D-dimer Cardiac Reader X 0.5 mg/mL
NycoCard D-dimer X Not available
Triage Cardiac Panel X 100 ng/mL
Stratus CS STATa X Under development
H-FABP RAPICHECK X 6.2 ng/mL
MLC-1 LifeSign MIb X Not available
Myoglobin Cardiac Reader X 76 ng/mL (M), 64 ng/mL (F)
Cardiac STATus X 80 ng/mL
LifeSign MI X 50 ng/mL
Nexus Dxa,b X Under development
One-Step Marker Test X 50 ng/mL
RAMP Reader X 99.3 ng/mL
Stratus CS STAT X 98 ng/mL (M), 56 ng/mL (F)
Triage Cardiac Panel X 107 ng/mL
NT-proBNP Cardiac Readera X Under development
LifeSign MIb X Not available
Nexus Dxa,b X Under development
Abbreviations: CRP, C-reactive protein; cTnT, cardiac troponin T; F, female; H-FABP, human-type fatty acid bind-
ing protein; M, male; MI, myocardial infarction; MLC-1, myosin light chain-1; NT-proBNP, N-terminal proBNP; TnI,
troponin I; TnT, troponin T.
Notes: a) Under development; b) Not available in United States.
Disclaimer: Tables 1 through 3 were compiled from reliable sources including company web sites and cited literature.
However, during evaluation before implementation users should update and verify all data directly from manufacturers
and product inserts. Users also should verify decision levels and reference methods, as well as FDA-approved clinical
applications of each cardiac biomarker.

Qualitative versus quantitative testing: the College of Cardiology/American Heart Associa-


American College of Cardiology/American Heart tion (ACC/AHA) 2002 guideline update [65] for
Association guidelines patients who present with unstable angina (UA)
Cardiac biomarkers linked with ecient triage and nonST-segment elevation myocardial in-
and treatment strategies can facilitate resource farction (NSTEMI) states that point-of-care as-
use, risk stratication, therapeutic management, says at present are qualitative or, at best,
and clinical outcomes [5964]. The American semiquantitative. The evolution of technology
474 KOST & TRAN

Table 3
Special applications of cardiac biomarker POCT
Special applications Biomarker/concept Device/index
ACS. Rapid test for ACS risk stratication utilizing cMyo Spectral Dx cMyo Test
cardiac myoglobin and CAIII for early specicity and (under development)
sensitivity (Chris Wayne, Spectral Diagnostics, e-mail
communication).
CHF. Patients who died or were readmitted tend to have BNP Triage BNP
an increase in BNP concentration during their
hospitalization ( 239 G 233 pg/mL). Patients who
had successful treatment tend to have decreases in
their BNP concentration during their hospitalization
(216 G 69 pg/mL). The dierence between these two
groups were signicant (P ! .05). Therefore, POCT
of BNP may be an eective way to improve in-
hospital management of patients with decompensated
CHF [45].
Correctional facility. When the Cardiac STATus is Multivariate analysis Cardiac STATus
combined with EKG results and medical history,
a positive result leads to early diagnosis and
subsequently faster, more eective treatment. A
negative result can provide the condence needed to
safely treat the inmate on premises. In the end the
correctional facility saved substantial funds (Chris
Wayne, Spectral Diagnostics, e-mail communication).
Diabetes risk. Bhalla and colleagues [46] state BNP to be BNP Triage BNP
the most signicant (P ! .001) predictor of all-cause
mortality in diabetic patients suspected of cardiac
dysfunction.
Diagnostic synthesis. MMX combines multiple Multimarker index MMX (Anderberg J.
biomarker test results in a continuous parameter that and colleagues.
shows potential to discriminate patients with ACS AACC Oak Ridge
from those without ACS (Julie Doyle, M.D., Boisite Conference, 2005;63)
Diagnostics, e-mail communication). From ROCs,
MMX AUC appears to outperform individual analyte
(cTnI, CK-MB, myoglobin, or BNP) AUCs in ACS
versus non-ACS and in other comparisons, such as
MI versus UA or non-ACS, including for subsets of
patients within 6 hours after the onset of chest pain.
MMX algorithms have the potential to aid in the
dierential diagnosis of other conditions, such as
stroke. Independent multimarker analysis also can be
used for risk stratication [47,48].
Early dysfunction. Mueller and colleagues [49] showed NT-proBNP (Roche Elecsys) POCT
NT-proBNP comparable to BNP biomarkers for under development
dierential diagnosis in symptomatic and
asymptomatic structural heart disease. However,
results also suggested NT-proBNP was able to discern
early cardiac dysfunction compared to BNP.
General practice environment. Dahler-Eriksen and CRP NycoCard Reader
colleagues [50] showed CRP testing in general practice
identied signicantly (P .002) more new diseases
(eg, infectious disease, inammatory disease)
compared to the control group. Additionally,
cost-eectiveness analysis showed a reduction of
$110,000 per year.
CARDIAC BIOMARKER POCT 475

Table 3 (continued )
Special applications Biomarker/concept Device/index
Home testing. Each year, more than a million persons in Various cardiac biomarkers Waived tests (to be
the United States have a heart attack, and about 50% developed)
of them die [51]. About 50% who die do so within 1
hour of the start of symptoms, such as pre-infarction
angina. Home testing could improve outcomes.
Military. Cardiac troponin T can help exclude cTnT TropT Sensitive and
myocardial damage in patients who experience an Cardiac T
elevated CK caused by skeletal muscle trauma [52].
Military. The Cardiac STATus is available to every Multimarker Analysis Cardiac STATus
surface ship in the entire US Navy eet (Chris Wayne,
Spectral Diagnostics, e-mail communication).
Mobile intensive care unit. When used within 2 to Multimarker Analysis Cardiac STATus
12 hours from the onset of symptoms, the CK-MB/
myoglobin multimarker test can prevent misdiagnosis
of AMI or unnecessary hospitalization [53].
Paramedic use. Use of the cTnI test helps Onslow cTnI Cardiac STATus
County paramedics to quickly diagnose chest pain,
and condently rule-in or rule-out heart attacks while
patients are in transit to the hospital. In certain cases,
if the paramedic or ED physician conrms the patient
is having a heart attack they can, following a protocol,
implement therapy before they reach the hospital
doors (Chris Wayne, Spectral Diagnostics, e-mail
communication).
Potential new biomarker. For the diagnosis of AMI, H-FABP RAPICHEK
quantitative ELISA-based measurement of H-FABP
has been shown to be more sensitive than myoglobin
or CK-MB, and more specic than myoglobin [22].
However, the ELISA process takes R 90 minutes
whereas the RAPICHEK takes 15 minutes.
Additionally, the RAPICHEK test exhibited superior
negative predictive values, which reached nearly 100%
for patients in all time frames. Seino and colleagues
[22] concluded that the H-FABP could be useful for
emergencies.
Prehospital admission. Prehospital cTnT testing appears cTnT TropT Sensitive and
to be an objective marker for patients with poor Cardiac T
outcomes [54].
Renal dysfunction. McCullough and colleagues [55] cTnI Triage Cardiac Panel
showed myoglobin and CK-MB correlated with
corrected creatinine clearance (r 0.36, P ! 0.01,
and r 0.10, P .01, respectively), while cTnI did
not (r 0.10, P .12). Multiple receiver operating
characteristic curve testing showed cTnI to be the
most consistent marker of myocardial injury over all
strata of renal dysfunction including end-stage renal
disease on dialysis [56]. Therefore, cTnI is applicable
and superior to myoglobin or CK-MB in the
evaluation of chest pain in patients with renal
dysfunction [55].
Renal failure. In patients with severe renal failure, cTnT cTnT TropT Sensitive and
could not be broken down and cleared by the kidneys, Cardiac T
resulting in high serum cTnT levels [56].
(continued on next page)
476 KOST & TRAN

Table 3 (continued )
Special applications Biomarker/concept Device/index
Secondary referral. European Society of Cardiology BNP BNP
recommendations include BNP testing to refer
patients to secondary care.
Thrombosis. As a cardiac biomarker, Dempe and D-dimer Cardiac D
colleagues [57] found the Roche Cardiac D Rapid
Assay was also useful as a rule-out diagnostic tool for
patients with suspected acute thrombosis [58]. The
Roche Cardiac D Rapid Assay also can be used for
diagnostic work-up for deep vein thrombosis [57].
Abbreviations: AUC, area under curve; CAIII, cardiac anhydrase III; cMyo, cardiac myoglobin; CorrCrCl, corrected
creatinine clearance; CRP, C-reactive protein; cTnT, cardiac troponin T; ELISA, enzyme-linked immunoabsorbant as-
say; F, female; H-FABP, human-type fatty acid binding protein; M, male; MI, myocardial infarction; MMX, multi-
marker index; NA, not available; NT-proBNP, N-terminal pro B-type natriuretic peptide; ROC, receiver operator
curve; and UA, unstable angina.

that will provide quantitative assays of multiple and colleagues [25] studied the potential usefulness
markers that are simple to use will improve the di- of rapid bedside qualitative myoglobin/CK-MB
agnosis and management of patients with sus- and cTnT tests in initiating revascularization ther-
pected acute coronary syndrome (ACS) in the apy. James and colleagues [48] found rapid qualita-
ED. The 2004 ACC/AHA guidelines [66] for the tive cTnI suboptimal for prediction of subsequent
management of patients with STEMI stated, Al- cardiac events at suspicion of unstable coronary
though handheld bedside (point-of-care) assays syndromes. Hirschl and colleagues [71] found reli-
may be used for a qualitative assessment of the pres- able clinical performance of qualitative cTnT test-
ence of an elevated level of a serum cardiac bio- ing, whether performed in laboratories or by nurses
marker, subsequent measurements of cardiac and physicians in critical care units. For patients
biomarker levels should be performed with a quan- with suspicious AMI, Seino and colleagues [22]
titative test.. in general, bedside assays are less showed improved early sensitivity with a rapid
sensitive and less precise than quantitative assays. whole-blood quantitative assay for heart-type fatty
Further, A positive bedside test should be con- acid binding protein versus qualitative cTnT test-
rmed by a conventional quantitative test [66]. ing. Wu and colleagues [31] reported equivalent re-
These guidelines statements regarding qualitative sults for quantitative cardiac biomarker POCT and
assays allude to the various types of handheld for- laboratory testing. Other studies [21,72] address
mats and disposable test kits noted for convenience quantitative bedside testing. Additionally, POCT
[67] and listed in Table 1. Qualitative versus quan- encompasses near-patient testing (NPT), such as
titative assays are identied in Table 2. Much in ED satellite laboratories, where automated ana-
room for improvement exists in the detection of mi- lyzers that provide quantitative testing (eg, Stratus
nor myocardial damage. The sensitivity and preci- CS STAT) often are placed.
sion at the cuto concentrations of qualitative In the ED, analytically sensitive quantitative
and quantitative assays should be checked and cTn assays have blurred the distinction between
deemed clinically appropriate before implementa- patients who present with and without classically
tion [68] in chest pain centers. dened acute myocardial infarction (AMI) and
Assay types, clinical applications, and testing have focused attention on the continuum of ACS
sites create many alternatives, and the relevance of from angina to transmural Q-wave myocardial
published studies will depend in part on the infarcation (MI) [73,74]. With rst-draw speci-
conguration and stang of individual chest mens in the ED and qualitative POCT, Kratz
pain centers. Hamm and colleagues [69] found and colleagues [75] showed that triple cardiac bio-
qualitative POC cardiac troponin T (cTnT) and markers (ie, a test cluster) may be needed to avoid
cTnI tests to be sensitive and useful for ED triage. weak positive predictive values for individual
Schwartz and colleagues [70] documented diag- tests, and quantitative conrmation in the clinical
nostically similar results from CK-MB and myo- laboratory yields additional improvement. In
globin qualitative testing in the ED and a ve-hospital study of acute chest pain patients,
quantitative testing in the laboratory. Panteghini Goldman and colleagues [18] reported equivalent
CARDIAC BIOMARKER POCT 477

results for quantitative POC cTnT and myoglobin identifying the bottleneck route, that is, the crit-
testing in the ED versus testing in the clinical lab- ical path of acute cardiac care in chest pain centers
oratory and stated that POC cTnT proved advan- and EDs. Expedited treatment inuences patient
tageous for rapid decision making. In the ED, survival, and TTAT can assess eciency. If blood
optimal clinical sensitivity for ACS on rst-draw specimens are obtained rst or before verbal or-
specimens derives from myoglobin, a better ad- ders are documented, as may occur in critical sit-
junct test than CK-MB [76]. Although qualitative uations, TTAT starts with collection. In the
POC methods appear in routine and also several algorithmic evaluation and treatment of chest
special settings (see Table 3), to expedite clinical pain using critical pathways and in concepts rec-
assessment without missing cases of AMI, quanti- ommended by the ACC/AHA [65,66,88] (also
tative and highly sensitive POC assays are needed. ESC [89,90]), the AMI and the UA/NSTEMI
POC assays also allow monitoring of release and pathways call for rapid cTn testing but not neces-
clearance dynamics in the form of bedside changes sarily at the point of care. POCT, however,
(D, delta values) [65]. A conservative approach, eliminates preanalytic delays. Caragher and col-
the evidence to date, and the guidelines above in- leagues [91] showed that POCT for myoglobin,
dicate a need for highly sensitive and quantitative CK-MB, and cTnI expedited diagnosis in the ED
cardiac biomarker POCT, and given a choice, by decreasing the time to test results by 55% (39
chest pain centers should implement quantitative versus 87 minutes) compared with the clinical lab-
assays. oratory. Hsu and colleagues [33] found a reduction
from 65 to 26 minutes with POCT. McCord and
colleagues [6] reported a median time from sam-
Ischemia markers: critical need
pling to reporting of 24 minutes, and others
Any ischemia biomarker that works well clini- [92,93] reported results availability in about 20 mi-
cally, that is, can detect ischemia in the absence of nutes. Stubbs and Collinson [94,95] showed that
necrosis, would be an immediate success if used for POCT of cardiac biomarkers decreased turn-
POCT. Aggressive early triaging demands im- around time by 52 minutes (central laboratory,
proved cardiac biomarkers for that purpose. A 72; POCT, 20) and shortened hospital stay. Altin-
new biomarker, ischemia-modied albumin (IMA) ier and colleagues [96] reported a turnaround time
(ACB, albumin Cobalt binding test; Ischemia of 82.5 minutes (50th percentile) versus 17 minutes
Technologies, Denver, Colorado), has been ap- with POCT, which facilitated faster discharge for
proved and licensed by the Food and Drug some patients in the ED. In a study of ve hospi-
Administration (FDA). Unfortunately, clinical tals, Gaze and colleagues [97] documented that
data show that the current generation of the IMA bedside cTnT testing in ED produced results in
assay lacks requisite specicity [7787] like that 12 to 22 minutes, a gain in time of 65 minutes
possibly aorded, for example, by ultrahigh-sensi- (range 34 to 135 minutes) compared with central
tivity assays for troponins. In addition, IMA has laboratory measurements. Thus, these studies
not been implemented on any of the POC plat- have established POCT as fast and capable of pro-
forms listed in Table 1. Biomarkers of ischemia un- ducing clinical results in time spans comparable to
der evaluation include free fatty acid, glycogen the analysis times shown in Table 1.
phosphorylase isoenzyme BB, myeloperoxidase, Lee-Lewandrowski and colleagues [98] showed
nourin-1, pregnancy-associated plasma protein A, that when POC cardiac biomarkers were com-
sphingosine-1-phosphate, soluble CD40L, and bined with other POC tests (ie, pregnancy testing
whole-blood choline. Future clinical studies of is- and urine dipstick), the integration decreased test
chemia biomarkers should target low-risk patients turnaround time by 87% and ED length of stay
in whom quick evaluation in the chest pain center by 41.3 minutes. Performed at a time when Bos-
can exclude AMI, and then attention can turn to ton ambulance diversion was an issue, this study
the underlying cause of the chest pain. found that implementation of a POCT satellite
laboratory in the ED decreased divert hours by
27%. POCT increased physician satisfaction
[68]. Integration of testing plus synthesis of test
Therapeutic turnaround time
results and other clinical evidence represents an
Because it measures the total time from test or- essential pivot for knowledge optimization and
dering to treatment, TTAT (Fig. 1) represents an decision making [3]. Collinson [94] suggested
overall performance monitor (metric) useful for that POCT must be used if cardiac biomarker
478 KOST & TRAN

results are delayed more than 25% of the clinical injury markers (cTn or CK-MB), while most
decision time. In reality, POCT integrates deci- (75%) ED physicians wanted results within 45 mi-
sion making at the bedside. Widespread availabil- nutes, in regards to patients presenting to the hos-
ity of POCT now makes fast TTAT an pital ED with symptoms of AMI. Neither goal
expectation in critical care management, the early was met. Among the fastest performing 25% of
recognition of life-threatening conditions, and the laboratories, median order-to-report cTn and
titration of commonly applied therapies [2,99 CK-MB response times equaled 50 and 48.3 mi-
101]. Experience in United States legal delibera- nutes, respectively. On the average, 90% of results
tions has demonstrated that POCT may be were reported in slightly more than 90 minutes
deemed clinically necessary to achieve fast measured from the time the tests were ordered.
TTAT. Hence, well-conceived and mutually In about 50% of the hospitals, specimens were or-
agreed upon goals for TTAT that are based on dered by protocol before clinicians saw their
evidentiary needs [102] will improve cycles of in- patients. In 35.8% of participating hospitals, lab-
dividual patient care and overall performance in oratory policies dictated that no results be re-
chest pain centers and emergency medicine. ported until all results became available. In
10%, clinicians waited more than 2 hours for re-
sults. Shorter response times were associated
Timeliness with having specimens obtained by the laboratory
rather than by nonlaboratory personnel and with
Perception versus performance
performing marker assays in EDs or other periph-
Discordant viewpoints about how to measure eral laboratories compared with performing tests
the timeliness of diagnostic testing frustrate team- in central laboratories. For the CAP respondents,
work. Most clinicians measure response time as however, only 7.2% used an ED or other periph-
the interval starting with test ordering and ending eral laboratory, and only 4.1% used POC
with result reporting [103]. In contrast, laborator- instruments.
ians typically start the clock when they receive the
sample in the laboratory and thereby discount
Evolution of guidelines and the standard
preanalytic delays [103] that may degrade sam-
of care for timeliness
ples. Even for ED tests, such as potassium and he-
moglobin, specimen transit times account for The discipline inherent in applying practice
approximately one third of the total response guidelines can improve teamwork and outcomes
time [104]. Active monitoring improves perfor- [107109]. In 1993 over one decade ago, the Na-
mance when the laboratory does not control the tional Institutes of Health (NIH) published,
specimen handling process, but invariably, re- Emergency Department: Rapid Identication
sponse time goals for EDs still are not met most and Treatment of Patients with Acute Myocardial
of the time [105]. Signicantly delayed outliers in- Infarction [110], a product of the 60 Minutes to
terrupt diagnostic-therapeutic processes. When - Treatment Working Group of the National
nally reported, results from such delayed samples Heart Attack Alert Program Coordinating Com-
may no longer represent physiologically or medi- mittee, which had nearly forty member organiza-
cal relevant data, or they may sit in the laboratory tions at the time. The working group proposed
information system unnoticed if the laboratory a goal of 60 minutes for laboratory test results
has not listed them as critical values. For en- .in patients with questionable symptoms and
hanced clarity, communication, understanding, ECG ndings., and 30 minutes for algorithmic
and teamwork, chest pain center physicians and door-to-drug treatment of patients with a clear
laboratorians should agree on denitions of re- diagnosis of AMI. The document stated, howev-
sponse time, jointly develop performance goals, er, that the value of rapidly available assays.
and perpetually monitor key indicators. [CK isoforms, troponin, myoglobin] is unknown
In 2004, a College of American Pathologists at this [1993] time [110].
(CAP) survey [106] of 159 hospital subscribers Five years ago the ACC/AHA 2000 guidelines
(mostly in the United States) to the Q-Probes pro- [88] for the management of patients with UA and
gram revealed critical deciencies in the timeliness NSTEMI recommended a target of 60 minutes
of cardiac biomarker services. Most (82%) labo- with 30 minutes preferred for completion of cen-
ratory participants deemed 1 hour or less a reason- tral laboratory test results, and stated further
able order-to-report response time for myocardial that, Point-of-care systems, if implemented at
CARDIAC BIOMARKER POCT 479

the bedside, have the advantage of reducing delays sent home prematurely. In chest pain centers,
due to transportation and processing in a central therefore, POCT should be approached not only
laboratory and can eliminate delays due to the from the perspective of cardiac biomarkers but
lack of availability of central laboratory assays also from the broader standpoint of garnering
at all hours [88]. The guidelines also stated of benets that arise when integrating several test
POCT that The evolution of technology that clusters. Delays could impede critical pathways,
will provide quantitative assays of multimarkers increase legal liability, aggravate risk exposure,
that are simple to use will improve the diagnosis and possibly also degrade medical and economic
and management of patients with suspected ACS outcomes [3]. For some chest pain centers, fast
in the ED [88]. The 2002 update reiterated [65] TTAT and decision making will be achieved
that When a central laboratory is used, results only by using POCT, in the form of bedside or
should be available within 60 minutes, preferably near-patient testing.
within 30 minutes. The timeliness envelope (curve) in Fig. 2 traces
The web-based draft of the National Academy progressive contraction of the time interval, test
of Clinical Biochemistry (NACB) guidelines [111] ordering / results receipt, the rst of two inter-
proposed that The laboratory should perform vals that make up TTAT (see Fig. 1), versus the
cardiac marker testing with a turnaround time year and conceptual evolution in POCT. To the
(TAT) of 1 hour, optimally 30 minutes, or less. right, standard-of-care milestones for cardiac bio-
The TAT is dened as the time from blood col- markers demarcate points that NIH and ACC/
lection to the reporting of results. Institutions AHA published guidelines [88,110], practice
that cannot consistently deliver cardiac marker standards, and chest pain centers have called for
TATs of approximately 1 hour should implement or actualized by virtue of published literature cit-
POC testing devices. The phrase, .approxi- ed herein, faster response time. The performance
mately 1 hour. requires clarication to avoid (map) vector [2] consisting of POCT, focused mul-
misunderstandings. Nevertheless, over the past timarker innovations, faster immunotesting, and
12 years these NIH, ACC/AHA, and NACB test cluster integration points to diagnostic-thera-
guidelines have pointed consistently toward peutic (Dx-Rx) process optimization by which
faster cardiac biomarker testing within 30 mi- physicians in chest pain centers intuitively synthe-
nutes. One must assess, therefore, whether POC size risk indices for individual patients. POCT
technologies can meet or exceed expectations in provides not only immediate results but also acute
these guidelines, positions tantamount to the awareness. The whole picture reects a paradigm
standard of care. Additionally, the statistical dis- shift enabled by POCT and increasingly ecient
tribution of performance times and outlier de- solutions for the management of ACS patients.
lays, and standards in critical care testing using In the future, real-time in vivo and noninvasive
WBA, must be considered to formulate cohesive function monitors, now embryonic or yet to be
rapid response criteria for chest pain centers. discovered, will supplant discrete in vitro diagnos-
Lastly, the Society of Chest Pain Centers should tic tests and lead to continuous knowledge optimi-
consider adopting timeliness as a criterion for ac- zation at the bedside, thus further improving the
creditation, as suggested in later discussion. standard of care.
POCT has evolved substantially during the
past decade, and so has the way it ts into clinical
practice [112]. Now ubiquitous in the United
Cost-eectiveness, algorithmic strategies,
States, POCT has established a new precedent
and multimarkers
for speed in critical care testing. Timeliness ex-
pectations for test results, such as electrolytes, Cardiac biomarkers, especially cTn and myo-
blood gases, pH, and hematocrit or hemoglobin, globin [114,115], play a pivotal role in cost-eec-
in critical care medicine now have contracted to tively triaging patients, assessing damage,
only 5 to 15 minutes, basically a little longer gauging risk, determining prognosis, and econo-
than WBA time [44,113]. This expectation for mizing length of stay [116121]. Decision analysis
timeliness, a de facto legal requirement in the modeling showed that in patients who presented
care of critically ill patients, prohibits diagnostic with acute undierentiated chest pain, biochemi-
testing from prolonging TTAT and certainly cal testing, observation for 2 to 6 hours, and
from ever degrading eciency in cases where then repeat testing, had a low incremental cost
AMI might be missed in an ED and patients per quality adjusted life year, and that strategies
480 KOST & TRAN

Fig. 2. The standard of care for timeliness of cardiac biomarker testing. Forceful clinical drivers mold the shape of the
rapid response curve while the underlying performance vector enables it, aiming at optimization, rst discrete, then in the
future, continuous. Contemporary timeliness in critical care testing has set a continuum precedent for chest pain centers,
which now can consider 15 minutes as reasonable criterion for accreditation. However, all cardiac biomarker results
should be available within 30 minutes, the preference stated in ACC/AHA guidelines and the optimal stated in draft
NACB guidelines.

requiring hospital admission had poor value [122]. Algorithmic strategies, however, using cardiac bi-
Brogan and colleagues [26] recommended serial omarkers need to be adapted to local settings [40].
testing of CCU admissions to eliminate non- The relative ecacy of diverse testing strategies
AMI causes and conserve resources. While serial reported in the literature remains uncertain, in
testing of cardiac biomarkers, a hallmark for diag- part because sensitivity and specicity for AMI
nosis of AMI, comes in at the low end of the cost depend on: biomarkers selected, decision levels,
spectrum of invasive procedures and hospitaliza- timing of testing after symptom onset, natural dis-
tion expenses [40], new POC multimarker ap- cordance, unique settings, risk and age strata of
proaches will redene serial testing needs and patients [128], infarct size [129], and random sto-
temporal patterns. In the ED, cost and other is- chastic events encountered commonly during
sues, such as lack of connectivity or apparently emergencies. Nonetheless, Mant and colleagues
modest value in STEMI, have slowed acceptance [130] used Monte Carlo statistical simulation of
of POCT [123,124]. However, at a meeting in management strategies for suspected ACS and
2004, Kugelmass and colleagues [125] reported found that POCT with cardiac troponins proved
that in a broad group of four community-based cost-eective.
hospitals, for 2248 consecutive admissions of pa- In the CHECKMATE study, Newby and
tients who received a primary diagnosis of AMI, colleagues [131] showed that rapid quantitative
revascularization (PCI, CABG) rates increased multimarker (myoglobin, CK-MB, and cTnI)
signicantly, and in-hospital mortality declined NPT identied positive patients earlier and pro-
24.7% (P .049) following 1 year of implementa- vided better risk stratication for mortality than
tion of POCT. a local laboratory-based, single-marker approach.
In the context of ACS and chest pain evalua- NPT was used to facilitate early detection. The
tion, chest pain centers can provide the dual study related marker status with 30-day death or
eciencies of expediting denitive treatment while infarction. All three markers discriminated 30-
managing patients with accelerated diagnostic- day death better than just CK-MB and cTnI,
therapeutic protocols that proved cost-eective and the report identied the eect of NPT with
and benecial to patient care [69,126,127]. multimarkers, despite lack of standardization
CARDIAC BIOMARKER POCT 481

and completion before guidelines suggested use of doubling of the mortality risk for each additional
the 99th percentile of the normal interval and the biomarker that was elevated (P .01). The in-
10% coecient of variation precision level as the vestigators suggested using the number of elevated
cuto [132]. In 2001, McCord and colleagues [6] biomarkers at presentation to risk stratify patients
showed POC myoglobin and cTnI can help ex- for short- and long-term major cardiac events.
clude AMI rapidly during the rst 90 minutes af- Multimarker testing as historically congured
ter presentation to the ED. Using NPT in the [49,50] now embraces BNP, CRP, and other cardi-
CCU, Eggers and colleagues [133] showed that ac biomarkers to support patient evaluation and
when a 10% coecient of variation cuto was treatment. Speed, per se, can be achieved through
used, POC multimarkers (cTnI CK-MB myo- ecient integration of all clinical pathology serv-
globin) did not necessarily oer additional diag- ices, including the main laboratory (eg, using an
nostic value versus cTnI alone in the exclusion automated transport system), NPT (eg, by placing
of MI. Thus, decision levels may aect the clinical a satellite laboratory in the emergency area), and
eciency of multivariate (multimarker) ap- bedside testing. The extent to which each of these
proaches, which need to be optimized mathemat- modalities can be used conveniently for multi-
ically and statistically. marker testing depends on individual health
Biomarkers of neurohormonal activation in system facilities, organization, and priorities.
heart failure include BNP and the N-terminal However, to take advantage of multivariate anal-
fragment of its prohormone (NT-proBNP), col- ysis (eg, using cTnI, CK-MB, myoglobin, BNP,
lectively called BNP here. In a recent University and CRP) and multimarker benetsdincluding
HealthSystem Consortium comprehensive review, improved triage, facilitated decision making, con-
Cummings and colleagues [134] cataloged the sistent ACS work-up of NSTEMI, enhanced com-
merits of BNP-facilitated decision making for munication, ecient cardiology consultation,
early intervention, fewer echocardiograms, severity shortened intervention wait times, scaling of treat-
assessment, risk stratication, prognostic value, ment, decreased risk for inappropriate discharge,
therapeutic monitoring, discharge decisions, post- and reduced length of stay [47,133,136140]dtest-
surgical status, transplant rejection, sudden death, ing should be early, simultaneous, and quick,
device alternatives, and so on. For example, a which can be accomplished conveniently with
rapid bedside BNP assay recently was shown POCT.
[135] cost-eective and ecient in accelerating
hospital discharge for patients presenting to the ED
Chest pain centers, risk management,
with acute dyspnea. Outpatient reimbursement
and customer satisfaction
for BNP testing increased in 2003. BNP assays
now reside, however, on several automated chemis- The largest dollar amount for liability recovery
try analyzers appropriate for main laboratories in still involves chest pain with subsequent missed
hospitals where inpatient testing is covered under transmural MI [141]. Rather than treating pa-
the capitated DRG reimbursement system. As tients in their oces, more physicians refer patients
noted earlier, the 2004 POCT survey [13] showed to EDs, and high-risk specialists progressively be-
that institutions may move BNP to the main labo- come unwilling to provide ED on-call coverage be-
ratory. Thus, the cost-eectiveness of future BNP cause of malpractice concerns and skyrocketing
testing depends on multiple dynamic factors that premiums [142]. Eciency mandates that chest
may change often. Chest pain centers are well ad- pain centers should use prompt rule-out schemes
vised to establish guidelines for BNP testing and and that rapid cardiac biomarker testing should
its timeliness now and, with the aim of improving be performed (along with rapid EKG) to help
outcomes, reevaluate the guidelines periodically physicians avoid missing patients who present
during the next 2 years. with AMI, currently estimated to occur in 2% to
BNP, other biomarkers, and POC test clusters 5% [109,143,144]. Chest pain centers that adopt
(see Table 1) work synergistically. For example, a macroscopic viewpoint, aggressive leadership,
although not POCT, Sabatine and colleagues and legal precautions generate excellent communi-
[47] documented the value of including BNP and ty relations and customer satisfaction [145]. Expe-
C-reactive protein (CRP) along with cTnI for rience in legal case analysis has revealed patients
risk stratication, When patients were catego- were sent home before clinical laboratory results
rized on the basis of the number of elevated bio- became available, only to discover too late an
markers at presentation, there was a near elevated cTnI or CK-MB. Also, physicians may
482 KOST & TRAN

order cardiac biomarker tests for which blood POCT, whether performed at the bedside or in
collection is delayed, and attend to other patients near-patient satellite laboratories.
or leave the area, thereby inadvertently overlook- The 2004 United States survey of POCT [13]
ing important results. POCT and physician cap- discussed earlier showed that some main laborato-
ture [2] can help alleviate these types of errors ries may move tests, such as BNP, away from the
and simultaneously provide other critical test re- point of care; however, timeliness must be pre-
sults necessary for patient evaluation and triaging. served. The authors recommend, therefore, that
To reduce risk exposure each institution the timeliness of cardiac biomarker testing and
should identify bottlenecks in cardiac critical also the elimination of all delayed outlier results
pathways and then determine specically if car- be adopted as criteria for accreditation by the So-
diac biomarker POCT can accelerate evaluation ciety of Chest Pain Centers. These criteria should
and therapy. TTAT actually achieved not only be reassessed at least every 2 years as improved
should meet ACS/AHA guidelines but also should POC technologies for WBA grow progressively
satisfy the current standard of care for rapid re- faster, more accurate, and better integrated.
sponse testing of patients with life-threatening
conditions. Table 1 includes tests (other than car-
Therapeutic turnaround time and risk reduction
diac biomarkers) that are available on the same
platforms. These other tests can help physicians POCT has set a general precedent for timeli-
synthesize critical care solutions [2,3]. For exam- ness in the standard of care for critically ill
ple, clinicians evaluating acutely ill patients who patients. WBA facilitates that timeliness. TTAT
present with chest pain can order electrolyte mea- (see Fig. 1) represents a common denominator for
surements, such as potassium (K) and ionized performance. Diagnostic testing must not com-
calcium (Ca), and these analytes, already avail- promise TTAT. Chest pain centers not using
able on several POC platforms (possibly the same POCT or fast WBA of cardiac biomarkers eventu-
ones that perform cardiac biomarker testing), can ally may have to prove, perhaps in a court of law
be integrated increasingly as market demand and if delayed test results lead to a missed diagnosis of
manufacturing eciencies drive instrument con- AMI and adverse outcomes, that they consistently
solidation in the cardiac biomarker eld. achieve requisite speed and TTAT without POCT.
Even if courier and pneumatic tube systems
serve the ED, testing performed in a main labo-
ratory may generate signicantly and unpredict-
Summary: planning and implementing
ably delayed results that impair patient evaluation
point-of-care testing
or treatment. Decision-makers must be cognizant
Table 4 presents 12 fundamental objectives and of the medicolegal ramications and reduce risk
recommendations to consider when planning and by optimizing performance, that is, by minimiz-
implementing POCT in chest pain centers. If accu- ing TTAT. TTAT represents an integrative per-
rate, precise, safe, and cost-eective in the context formance indicator that can be applied to entire
of holistic cycles of patient care, no a priori reason cycles of patient care. The time required to
exists to not use POCT. The discussion below perform diagnostic testing should t comfortably
highlights points from Table 4 and key considera- within the TTAT goals for dierent diagnoses in
tions for chest pain centers. individual chest pain centers, so that testing does
not delay the critical paths of care and increase
risk exposure for patients.
Timeliness accreditation and the Society
of Chest Pain Centers
User-friendly new cardiac biomarkers
For cardiac biomarkers, the evolution in clin-
and instruments
ical guidelines has shaped the standard-of-care
envelope for timeliness (see Fig. 2). Response time In view of rapid growth, anticipate the appear-
is being decreased by minimizing the WBA time ance of several new cardiac biomarkers, some
for on-site quantitative testing in vitro. Future developed exclusively for POCT. For novel tests
ex vivo and in vivo techniques for assessing myo- and new indices, chest pain centers should docu-
cardial function can generate immediate real-time ment clinical justication and assure cost-eec-
data. Needs assessment by leadership in individual tiveness. The benets of on-site testing should
chest pain centers will reveal how best to optimize outweigh the costs of innovating, implementing,
CARDIAC BIOMARKER POCT 483

Table 4
Planning and implementing future POCT in chest pain centers
Objective Recommendation
 Algorithmic integrative strategies Embedding POCT within diagnostic-therapeutic algorithms, care
paths, and critical pathways improves the overall eciency and
eectiveness of cardiac biomarker testing.
 Connectivity CPCs should use primarily POC devices that provide connectivity
to the electronic medical record and other computerized information
systems.
 Critical limits and urgent notication Urgent notication of critical results assures physician awareness
and may help prevent premature discharge to home of patients with
evidence of AMI.
 Error prevention and biohazard control Security of instrument operation, validation of personnel skills,
and certication of qualied operators who perform POCT help
prevent errors, decrease risk, and contain biohazards.
 Home testing and early detection Since up to 50% of AMI patients die before reaching the hospital,
future home testing with automated communication of cardiac
biomarker results may facilitate early life-saving detection and
transport.
 Integration of critical care test clusters POC technologies that combine simultaneous measurements of
blood gasses, pH, electrolytes (eg, K and Ca), metabolites (eg,
glucose), and other essential critical care test clusters along with
cardiac multimarkers will streamline holistic evaluation, diagnosis,
and treatment of critically ill patients.
 Minimum analysis time and critical path CPCs can provide cardiac biomarker test results as expeditiously as
permitted by newer, smaller, and faster analytical methods, which
soon may provide quantitative results on ngerstick whole-blood
samples in 2 to 3 minutes, thereby facilitating the critical path of
patient care.
 Multimarker synthesis and indices Multimarker synthesis will facilitate dierential diagnosis and risk
proling through evidence-based joint probabilities and
mathematical heuristics using indices that help decipher myocardial
ischemia, NSTEMI, STEMI, heart failure, and inammation.
 POCT coordinator and hybrid sta POCT should be managed by POC Coordinators who work
collaboratively with nurses and physicians to enhance performance
through education, quality control, compliance monitoring, and
prociency testing.
 Quantitative assay performance Quantitative POC assays should achieve equivalent or better
performance (detection threshold, accuracy, and 10% coecient
of variation for precision at the 99th percentile of normals) and
clinical performance (sensitivity, specicity, and predictive values)
versus main laboratory counterparts.
 Standardization, continuity, and consistency Use of identical POC assays in the CPC and critical care areas
(eg, CCU, ICU, and OR) promotes local consistency, while for
clinical continuity dissimilar assays, if used, should be referenced to
the same comparison method or better, standardized, and all
methods should have compatible reference intervals, decision levels,
and critical limits.
 Timeliness and accreditation Expanding numbers of CPCs in the US will lead to higher expectations
for fast cardiac biomarker results and codication of timeliness,
potentially a Society of CPCs accreditation requirement, here
suggested as the analysis time of 15 minutes for POCT, with all
results received by the clinical team within the ACC/AHA guidelines
preferred time of 30 minutes (order-to-receipt, 24 h/d, 7 d/wk).
Abbreviations: ACC, American College of Cardiology; AHA, American Heart Association; CCU, cardiac/coronary
care unit; CPC, chest pain center; ICU, intensive care unit; NSTEMI, non-ST-elevation myocardial infarction; OR, op-
erating room; STEMI, ST-elevation myocardial infarction; TTAT, therapeutic turnaround time.
484 KOST & TRAN

and operating POCT or proprietary miltimarker is recommended by the ACC/AHA, ECS, and
indices. Economic analysis should assess the total NACB guidelines discussed earlier, and also here.
benets and costs of care episodes and determine Qualitative testing remains benecial in some
the marginal eects of POCT alternatives. Anal- cases to determine quickly whether a patient is
ysis also should take into account hospital admis- hot or not. Highly sensitive quantitative testing
sions that could have been prevented through with well-dened detection limits [146], however,
timely cardiac biomarker POCT. Already valu- has several advantages and better ts the multidi-
able risk stratication will build as new POCT mensional, multimarker, and multivariate algo-
inventions and innovations appear, but lack of rithmic practice of the future in chest pain
denitive biomarkers for myocardial ischemia centers.
represents an unmet need. Research should target
discovery of highly sensitive and specic bio-
Leadership vision and chest pain centers
chemical tests predictive for myocardial ischemia.
Because the 2004 POCT United States survey With more than 150 sites in the United States
[13] showed hospitals evaluated but then rejected in 2005, chest pain centers now can pool pro-
cardiac biomarker instrument platforms more fre- fessional resources to: (a) consolidate care algo-
quently than other forms of POCT, user-friendly rithms, (b) write guidelines for POCT, (c)
POC instrument designs also are needed badly. systematize bedside and NPT, (d) consider adopt-
Solutions currently available will not satisfy the ing timeliness accreditation criteria, (e) form
present or future needs of chest pain centers purchasing groups to obtain POC devices and
or busy EDs. Poorly integrated systems that reagents inexpensively, (f) facilitate rapid discov-
lack connectivity to the electronic medical re- ery of evidence to avoid missing patients with
cord threaten quality control monitoring and AMI, (g) monitor performance indicators, such as
make the management tasks of the POC coordi- TTAT, regularly, and (h) provide outreach pro-
nator extremely dicult. Additionally, existing grams to extend cardiac biomarker testing to eld
devices are poorly adapted for emergency eld and home sites to help reduce unacceptably high
response and must be simplied substantially mortality rates there.
to garner waived status for home testing under POCT should be approached from the stand-
federal regulations. point of the laboratory-clinical-community inter-
face and optimized by physicians, nurses, POC
Algorithmic practice and quantitative testing coordinators, and laboratorians working collabo-
ratively [2,147]. Professional leadership groups,
POCT can speed diagnostic synthesis in chest such as the Society of Chest Pain Centers, can
pain centers, and speed is essential in EDs over- help guide the current rapid growth of cardiac
whelmed with patients. For most cardiac bio- biomarker POCT to fulll important practice
markers, accelerated algorithmic strategies will goals, including accurate diagnosis of myocardial
help position cardiac biomarkers temporally and ischemia, improved triage, administration of
optimally within the diagnostic decision tree and proven treatment (eg, glycoprotein IIb/IIIa inhib-
its treatment branches to take advantage of time- itors), quantitative risk stratication, early inva-
dependent kinetics, sensitivities, specicities, pre- sive strategies, and improved economic and
dictive values, and multivariate functions. The medical outcomes.
choice of test clusters, types of instruments, and
sites, such as bedside or near-patient (eg, an ED
satellite laboratory) and home or eld (eg, ambu- Acknowledgments
lance, helicopter, ships, and airplanes), rests with
collaborative care teams who must assure the We thank Ms. Claudia Graham, University of
continuity, eciency, and ecacy of critical path- California Artist, for drawing Fig. 1. We are in-
ways in emergency situations. debted to Ms. Susan Farber of Enterprise Analy-
Whether using cardiac biomarker POCT on- sis Corporation for sharing the 2004 US POCT
site or testing performed elsewhere, United States survey data on cardiac biomarkers. We thank col-
physicians (and physicians in United States juris- leagues in industry who responded when asked to
dictions abroad) must be aware of FDA-approved review the data in Tables 1 through 3. Figures and
manufacturer claims to avoid clinical misuse of tables were used permission and courtesy of
tests and unnecessary liability. Quantitative testing Knowledge Optimization, Davis, CA 95616.
CARDIAC BIOMARKER POCT 485

References [15] Rosen S. Market outlook for IVDs. IVD


Tech 2004;10(7):3943. Available at: http://www.
[1] Kost GJ. Principles and practice of point-of-care devicelink.com/ivdt/archive/04/06/001.html.
testing. Philadelphia: Lippincott, Williams, and Accessed December 28, 2004.
Wilkins; 2002. [16] Kost GJ. Goals, guidelines, and principles for
[2] Kost GJ. The hybrid laboratory, therapeutic turn- point-of-care testing. In: Kost GJ, editor. Princi-
around time, critical limits, performance maps, and ples and practice of point-of-care testing. Phila-
Knowledge Optimization. In: Kost GJ, editor. delphia: Lippincott, Williams, and Wilkins; 2002.
Principles and practice of point-of-care testing. p. 312.
Philadelphia: Lippincott, Williams, and Wilkins; [17] Kost GJ. Preventing medical errors in point-of-
2002. p. 1325. care testing: security, validation, safeguards, and
[3] Kost GJ. Controlling economics, preventing connectivity. Arch Pathol Lab Med 2001;125:
errors, and optimizing outcomes in point-of-care 130715.
testing. In: Kost GJ, editor. Principles and practice [18] Goldmann BU, Langenbrink L, Matschuck G,
of point-of-care testing. Philadelphia: Lippincott, et al. Quantitative bedside testing of troponin T: is
Williams, and Wilkins; 2002. p. 577600. it equal to laboratory testing? The Cardiac Reader
[4] Bourke SE, Kirk JD, Kost GJ. Point-of-care test- Troponin T (CARE T) study. Clin Lab 2004;50:
ing in emergency medicine. In: Kost GJ, editor. 110.
Principles and practice of point-of-care testing. [19] Agewall S. Evaluation of point-of-care test systems
Philadelphia: Lippincott, Williams, and Wilkins; using the new denition of myocardial infarction.
2002. p. 99118. Clin Biochem 2003;36:2730.
[5] Kost GJ, Bullock J, Despotis GJ. On-site and near- [20] Saadeddin S, Habbab M, Siddieg H, et al. Reliabil-
patient testing in the operating room. In: Kost GJ, ity of the rapid bedside whole-blood quantitative
editor. Principles and practice of point-of-care cardiac troponin T assay in the diagnosis of myo-
testing. Philadelphia: Lippincott, Williams, and cardial injury in patients with acute coronary syn-
Wilkins; 2002. p. 11932. drome. Med Sci Monit 2004;10:436.
[6] McCord J, Nowak RM, McCullogh PA, et al. [21] Yamamoto M, Komiyama N, Koizumi T, et al.
Ninety-minute exclusion of acute myocardial in- Usefulness of rapid quantitative measurement of
farction by use of quantitative point-of-care testing myoglobin and troponin T in early diagnosis of
of myoglobin and troponin I. Circulation 2001;104: acute myocardial infarction. Circ J 2004;68:
14838. 63944.
[7] Ng SM, Krishnaswamy P, Morissey R, et al. Nine- [22] Seino Y, Tomita Y, Takano T, et al. Oce cardiol-
ty-minute accelerated critical pathway for chest ogists cooperative study on whole blood rapid
pain evaluation. Am J Cardiol 2001;88:6117. panel tests in patients with suspicious acute myocar-
[8] Maisel AS. Point-of-care diagnosis and manage- dial infarction. Comparison between heart-type
ment of myocardial infarction and congestive heart fatty acid-binding protein and troponin T tests.
failure. In: Kost GJ, editor. Principles and practice Circ J 2004;68:1448.
of point-of-care testing. Philadelphia: Lippincott, [23] Schwartz JG, Gage CL, Farley NJ, et al. Evalua-
Williams, and Wilkins; 2002. p. 55466. tion of the cardiac STATus CK-MB/myoglobin
[9] Hainut C, Gade W. The emerging roles of BNP card test to diagnose acute myocardial infarctions
and accelerated cardiac protocols in emergency in the ED. Am J Emerg Med 1997;15:3037.
laboratory medicine. Clin Lab Sci 2003;16:16679. [24] Heeschen C, Goldmann BU, Moeller RH, et al.
[10] Amsterdam EA, Lewis WR, Kirk JD, et al. Acute Analytical performance and clinical application of
ischemic syndromes. Chest pain center concept. a new rapid bedside assay for the detection of serum
Cardiol Clin 2002;20:11736. troponin I. Clin Chem 1998;44:192530.
[11] Society of Chest Pain Centers. eNews Update. Jan- [25] Panteghini M, Cuccia C, Pagani F, et al. Compar-
uary 24, 2005. Available at: http://www.scpcp.org/ ison of the diagnostic performance of two rapid
accreditation/index.html. bedside biochemical assays in the early detection
[12] Kugelmass AD, Anderson AL, Brown PB, et al. of acute myocardial infarction. Clin Cardiol 1998;
Does having a chest pain center impact the treat- 21:3948.
ment and survival of acute myocardial infarction [26] Brogan GX, Bock JL, McCuskey CF, et al. Evalu-
patients [abstract no. 1932]? J Amer Heart Assoc ation of cardiac STATus CK-MB/myoglobin de-
(Suppl to Circulation) 2004;110(17):111409. vice for rapidly ruling out acute myocardial
[13] Farber S. US hospitals point-of-care survey. Car- infarction. Clin Lab Med 1997;17:65568.
diac makers section excerpt. January 24, 2005. Stan- [27] Shindelman J, Bellet N, Haley K, et al. Develop-
ford (CT): Enterprise Analysis Corporation; 2004. ment of a high-sensitivity immunoassay for c-reac-
[14] Diller W. Hospitals hang-ups with point-of-care tive protein on Cholestech LDX, a point-of-care
testing. Around the Industry. Available at: http:// analyzer. Point of Care: the Journal of Near-Pa-
www.windover.com; 2004. December: 2,8,9. tient Test and Technology 2004;3:1914.
486 KOST & TRAN

[28] Apple FS, Murakami MM, Christenson RH, et al. [42] Poe SS, Case-Cromer DL. Nursing strategies for
Analytical performance of i-STAT cardiac tropo- point-of-care testing. In: Kost GJ, editor. Princi-
nin I assay. Clin Chim Acta 2004;345:1237. ples and practice of point-of-care testing. Philadel-
[29] Drug Testing. Available at: http://www.drugtesting. phia: Lippincott, Williams, and Wilkins; 2002. p.
com.au. Accessed January 7, 2005. 21435.
[30] Syn X Pharma. Available at: http://www.synxpharma. [43] Anderson M. Assessing point-of-care testing pro-
com/news_2004_1.html. Accessed January 7, 2005. grams. In: Kost GJ, editor. Principles and practice
[31] Wu AHB, Smith A, Christenson RH, et al. Evalu- of point-of-care testing. Philadelphia: Lippincot,
ation of a point-of-care assay for cardiac markers Williams, and Wilkins; 2002. p. 41721.
for patients suspected of acute myocardial infarc- [44] Kost GJ. New whole blood analyzers and their im-
tion. Clin Chim Acta 2004;346:2119. pact on cardiac and critical care. Crit Rev Clin Lab
[32] Wantanabe T, Ohkubo Y, Matsuoka H, et al. De- Sci 1993;30:153202.
velopment of a simple whole blood panel test for [45] Cheng V, Kazanagra R, Garcia A, et al. A rapid
detection of human heart-type acid binding pro- bedside test for B-type peptide predicts treatment
tein. Clin Biochem 2001;34:25763. outcomes in patients admitted for decompensated
[33] Hsu LF, Koh TH, Lim YL. Cardiac marker point- heart failure: a pilot study. J Am Coll Cardiol
of-care testing: evaluation of rapid on-site bio- 2001;37:38691.
chemical marker analysis for diagnosis of acute [46] Bhalla MA, Chiang A, Epstheyn VA, et al. Prog-
myocardial infarction. Ann Acad Med Singapore nostic role of b-type natriuretic peptide levels in
2000;29:4217. patients with type 2 diabetes mellitus. J Am Coll
[34] Christenson RH, Cervelli DR, Bauer RS, et al. Car- Cardiol 2004;44:104752.
diac troponin I method: performance characteris- [47] Sabatine MS, Morrow DA, de Lemos JA, et al.
tics including imprecision at low concentrations. Multimarker approach to risk stratication in
Clin Biochem 2004;37:67983. non-ST elevation acute coronary syndromes. Si-
[35] Apple FS, Christenson RH, Valdes R, et al. Si- multaneous assessment of troponin I, C-reactive
multaneous rapid measurement of whole blood protein, and B-type natriuretic peptide. Circulation
myoglobin, creatine kinase MB, and cardiac tro- 2002;105:17603.
ponin I by the Triage Cardiac Panel for detection [48] James SK, Lindahl B, Armstrong P, et al. A rapid
of myocardial infarction. Clin Chem 1999;45: troponin I assay is not optimal for determination
199205. of troponin status and prediction of subsequent
[36] Clark TJ, McPherson PH, Buechler KF. Triage cardiac events at suspicion of unstable coronary
cardiac panel: cardiac markers for the triage sys- syndromes. Int J Cardiol 2004;93:11320.
tem. Point of Care: the Journal of Near-Patient [49] Mueller T, Gegenhuber A, Poelz W, et al. Head-to-
Test and Technology 2002;1:426. head comparison of diagnostic utility of BNP and
[37] Apple FS, Anderson FP, Collinson P, et al. Clinical NT-proBNP in symptomatic and asymptomatic
evaluation of the rst medical whole blood, point- structural heart disease. Clin Chim Acta 2004;
of-care testing device for detection of myocardial 341:418.
infarction. Clin Chem 2000;46:16049. [50] Dahler-Eriksen BS, Lauritzen T, Lassen JF, et al.
[38] Mulcahy K, Montalvo J, Tran L, Oltjenbruns X. Near-patient test for c-reactive protein in general
The alpha dx point-of-need system: an evaluation practice: assessment of clinical, organizational,
with reference to the national association of clinical and economic outcomes. Clin Chem 1999;45:
biochemistry recommendations for the use of cardi- 47885.
ac markers in coronary artery disease. Point of [51] http://www.nhlbi.nih.gov/health/dci/Diseases/Heart
Care: the Journal of Near-Patient Test and Tech- Attack/HeartAttack_WhatIs.html. Accessed Febru-
nology 2002;1:5861. ary 13, 2005.
[39] Tang Z, Louie RF, Kost GJ. Principles and perfor- [52] Collinson PO, Chandler HA, Stubbs PJ, et al. Mea-
mance of point-of-care testing instruments. In: surement of serum troponin T, creatine kinase MB
Kost GJ, editor. Principles and practice of point- isoenzyme, and total creatine kinase following ar-
of-care testing. Philadelphia: Lippincott, Williams, duous physical training. Ann Clin Biochem 1995;
and Wilkins; 2002. p. 6792. 32:4503.
[40] Cummings JP, Kost GJ, Matuszewski KA, et al. [53] Roth A, Malov N, Bloch Y, et al. Assessment of
POC tests for cardiac injury markers: clinical im- a creatine kinase-MB/myoglobin in the prehospital
provement and eectiveness. Oak Brook (IL): Uni- setting in patients presenting with acute nontrau-
versity HealthSystem Consortium; 2000. matic chest pain: the shahal experience. Crit
[41] Ammirati EB. Regulatory qualication of new Care Med 1999;27:10859.
point-of-care diagnostics. In: Kost GJ, editor. Prin- [54] Schuchert A, Hamm C, Scholz J, et al. Prehospital
ciples and practice of point-of-care testing. Phila- testing for troponin T in patients with suspected
delphia: Lippincott, Williams, and Wilkins; 2002. acute myocardial infarction. Am Heart J 1999;
p. 46980. 138:458.
CARDIAC BIOMARKER POCT 487

[55] McCullough PA, Nowak RM, Foreback C. Perfor- [68] Lee-Lewandrowski E, Benzer T, Corboy D, et al.
mance of multiple cardiac biomarkers measured in Cardiac marker testing as part of an emergency de-
the emergency department in patients with chronic partment point-of-care satellite laboratory in
kidney disease and chest pain. Acad Emerg Med a large academic medical center: practical issues
2002;9:138996. concerning implementation. Point of Care: the
[56] Diris JH, Hackeng CM, Kooman JP, et al. Im- Journal of Near-Patient Testing and Technology
paired renal clearance explains elevated troponin 2002;1:14554.
T fragments in hemodialysis patients. Circulation [69] Hamm CW, Goldmann BU, Heeschen C, et al.
2004;109:235. Emergency room triage of patients with acute chest
[57] Dempe C, Schraml M, Besenthal I, et al. Multi- pain by means of rapid testing for cardiac troponin
centre evaluation of a new point-of-care test for T or troponin I. N Engl J Med 1997;337:164853.
the quantitative determination of D-dimer. Clin [70] Schwartz JC, Gage CL, Farley NJ, et al. Evaluation
Chim Acta 2001;307:2118. of the Cardiac STATus CK-MB/myoglobin card
[58] Legnani C, Fariselli S, Cini M, et al. A new rapid test to diagnose acute myocardial infarctions in
bedside assay for quantitative testing of D-dimer (car- the ED. Am J Emer Med 1997;15:3037.
diac D-dimer) in the diagnostic work-up for deep [71] Hirschl MM, Herkner H, Laggner AN, et al. Ana-
vein thrombosis. Thromb Res 2003;111:14953. lytical and clinical performance of an improved
[59] Kost GJ, Omand K, Kirk JD. A strategy for the use qualitative troponin T rapid test in laboratories
of cardiac injury markers (troponin I and T, crea- and critical care units. Arch Pathol Lab Med
tine kinase-MB mass and isoforms, and myo- 2000;124:5837.
globin) in the diagnosis of acute myocardial [72] Muller-Bardo M, Rauscher T, Kampmann M,
infarction. Arch Pathol Lab Med 1998;122:24551. et al. Quantitative bedside assay for cardiac tropo-
[60] Hudson MP, Christenson RH, Newby LK, et al. nin T: a complementary method to centralized lab-
Cardiac markers: point of care testing. Clin Chim oratory testing. Clin Chem 1999;45:10028.
Acta 1999;284:22337. [73] Plebani M, Zaninotto M. Cardiac markers: central-
[61] Ohman EM, Armstrong PW, White HD, et al. Risk ized or decentralized testing? Clin Chem Lab Med
stratication with a point-of-care cardiac troponin 1999;37:11137.
test in acute myocardial infarction. GUSTOIII. [74] Storrow AB, Gibler WB. The role of cardiac
Am J Cardiol 1999;84:12816. markers in the emergency department. Clin Chim
[62] Giannitsis E, Lehrke S, Wiegand UK, et al. Risk Acta 1999;284:18796.
stratication in patients with inferior acute myo- [75] Kratz A, Januzzi JL, Lewandrowski KB, et al. Pos-
cardial infarction treated by percutaneous coro- itive predictive value of a point-of-care testing
nary interventions: the role of admission troponin strategy on rst-draw specimens for the emergency
T. Circulation 2000;102:203844. department-based detection of acute coronary syn-
[63] Azzazy HM, Christenson RH. Cardiac markers of dromes. Arch Pathol Lab Med 2002;126:148793.
acute coronary syndromes: is there a case for point- [76] Melanson SF, Lee-Lewandrowski E, Januzzi JL,
of-care testing? Clin Biochem 2002;35:1327. et al. Reevaluation of myoglobin for acute chest
[64] Lewandrowski K, Chen A, Januzzi J. Cardiac pain evaluation. Would false-positive results on
markers for myocardial infarction. A brief review. rst-draw specimens lead to increased hospital
Am J Clin Pathol 2002;118:S939. admissions? Am J Clin Pathol 2004;121:8048.
[65] Braunwald E, Antman EM, Beasley JW, et al. [77] Apple FS, Quist HE, Otto AP, et al. Release char-
ACC/AHA guideline update for the management acteristics of cardiac biomarkers and ischemia-
of patients with unstable angina and non-ST-seg- modied albumin as measured by the albumin
ment elevation myocardial infarctiond2002: sum- cobalt-binding test after a marathon race. Clin
mary article: a report of the American College of Chem 2002;48:1097100.
Cardiology/American Heart Association Task [78] Quiles J, Roy D, Gaze D, et al. Relation of ische-
Force on Practice Guidelines (Committee on the mia-modied albumin (IMA) levels following elec-
Management of Patients with Unstable Angina). tive angioplasty for stable angina pectoris to
Circulation 2002;106:1893900. Full text available duration of balloon-induced myocardial ischemia.
at: www.acc.org. Accessed February 16, 2005. Am J Cardiol 2003;92:3224.
[66] Antman EM, Anabe DT, Armstrong PW, et al. [79] Sinha MK, Gaze DC, Tippins JR, et al. Ischemia
ACC/AHA guidelines for the management of modied albumin is a sensitive marker of myocar-
patients with ST-elevation myocardial infarctiond dial ischemia after percutaneous coronary interven-
executive summary. J Amer Coll Cardiol 2004;44: tion. Circulation 2003;107:24035.
671719. Full text available at: www.aacc.org. [80] Wu AH. The ischemia-modied albumin biomark-
Accessed January 22, 2005. er for myocardial ischemia. Med Lab Obser 2003;
[67] Antman EM, Grudzien C, Sacks DB. Evaluation of 35:3640.
a rapid bedside assay for detection of serum cardiac [81] Borderie D, YAllanore, Meune C, et al. High
troponin T. JAMA 1995;273:127982. ischemia-modied albumin concentration reects
488 KOST & TRAN

oxidative stress but not myocardial ischemia in sys- kinase MB, myoglobin, and troponin T. Am Heart
tem sclerosis. Clin Chem 2004;50:21903. J 1998;135:67783.
[82] Garrido IP, Roy D, Calvino R, et al. Comparison [94] Collinson PO. The need for point of care testing: an
of ischemia-modied albumin levels in patient un- evidence-based appraisal. Scand J Clin Lab Invest
dergoing percutaneous coronary intervention for 1999;230(Suppl):6773.
unstable angina pectoris with versus without coro- [95] Stubbs P, Collinson PO. Point-of-care testing: a car-
nary collaterals. Am J Cardiol 2004;93:8890. diologists view. Clin Chim Acta 2001;15:5761.
[83] Roy D, Quiles J, Aldama G, et al. Ischemia modi- [96] Altinier S, Zaninotto M, Mion M, et al. Point-of-
ed albumin for the assessment of patients present- care testing of cardiac markers: results from an ex-
ing to the emergency department with acute chest perience in an emergency department. Clin Chim
pain but normal or non-diagnostic 12-lead electro- Acta 2001;311:6772.
cardiograms and negative cardiac troponin T. Int J [97] Gaze D, Collinson PO, Haass M, et al. The use of
Cardiol 2004;97:297301. a quantitative point-of-care system greatly reduces
[84] Roy D, Quiles J, Sharma R, et al. Ischemia-modi- the turnaround time of cardiac marker determina-
ed albumin concentration in patients with periph- tion. Point of Care: the Journal of Near-Patient
eral vascular disease and exercise-induced skeletal Testing and Technology 2004;3:1568.
muscle ischemia. Clin Chem 2004;50:165660. [98] Lee-Lewandrowski E, Corboy D, Lewandrowski
[85] Roy D, Quiles J, Sinha M, et al. Eect of direct-cur- K, et al. Implementation of a point-of-care satel-
rent cardioversion on ischemia-modied albumin lite laboratory in the emergency department of
levels in patients with atrial debrillation. Am J an academic medical center. Impact on test turn-
Cardiol 2004;93:3668. around time and patient emergency department
[86] Roy D, Quiles J, Sinha M, et al. Eect of radiofre- length of stay. Arch Pathol Lab Med 2003;127:
quency catheter ablation on the biochemical marker 45660.
ischemia modied albumin. Am J Cardiol 2004;94: [99] Kilgore ML, Steindel SJ, Smith JA. Evaluating stat
2346. testing options in an academic health center: thera-
[87] Zapico-Muniz E, Santalo-Bel M, Merce-Muntanola peutic turnaround time and sta satisfaction. Clin
J, et al. Ischemia-modied albumin during skeletal Chem 1998;44:1597603.
muscle ischemia. Clin Chem 2004;50:10635. [100] St-Louis P. Status of point-of-care testing: promise,
[88] Braunwald E, Antman EM, Beasley JW, et al. realities, and possibilities. Clin Biochem 2000;33:
ACC/AHA guidelines for the management of 42740.
patients with unstable angina and non-ST-segment [101] Drenck N. Point of care testing in critical care med-
elevation myocardial infarction: a report of the icine: the clinicians view. Clin Chim Acta 2001;307:
American College of Cardiology/American Heart 37.
Association Task Force on Practice Guidelines. J [102] Christenson RH, Collinson PO. Point-of-care test-
Am Coll Cardiol 2000;36:9701062 and Circula- ing for cardiac markers: an outcome-based ap-
tion 2000;102:1193209. praisal. In: Price CP, St John A, Hicks JM,
[89] Joint European Society of Cardiology/American editors. Point-of-care testing. Washington DC:
College of Cardiology Committee. Myocardial in- AACC Press; 2004. p. 37594.
farction redenedda consensus document of the [103] Howanitz PJ, Cembrowski GS, Steindel SJ, et al.
joint European Society of Cardiology/American Physician goals and laboratory test turnaround
College of Cardiology Committee for the redeni- times. A College of American Pathologists Q-
tion of myocardial infarction. J Am Coll Cardiol Probes study of 2763 clinicians and 722 institutions.
2000;36:95969. Arch Pathol Lab Med 1993;117:228.
[90] European Society of Cardiology Task Force for the [104] Howanitz PJ, Steindel SJ, Cembrowski GS, et al.
Diagnosis and Treatment of Chronic Heart Fail- Emergency department stat test turnaround times.
ure. Guidelines for the diagnosis and treatment of A College of American Pathologists Q-Probes
chronic heart failure. Eur Heart J 2001;22:152760. study for potassium and hemoglobin. Arch Pathol
[91] Caragher TE, Fernandez BB, Jacobs FL, et al. Lab Med 1992;116:1228.
Evaluation of quantitative cardiac biomarker [105] Steindel SJ, Howanitz PJ. Changes in emergency
point-of-care testing in the emergency department. department turnaround time performance from
J Emer Med 2002;22:17. 1990 to 1993. A comparison of two College of
[92] Collinson PO, Gerhardt W, Katus HA, et al. Mul- American Pathologists Q-Probes studies. Arch
ticentre evaluation of an immunological rapid test Pathol Lab Med 1997;121:103141.
for the detection of troponin T in whole blood [106] Novis DA, Jones BA, Dale JC, et al. Biochemical
samples. Eur J Clin Chem Clin Biochem 1996;34: markers of myocardial injury test turnaround
5918. time: a College of American Pathologists Q-Probes
[93] Sylven C, Lindahl S, Hellkvist K, et al. Excellent re- study of 7020 troponin and 4368 creatine kinase-
liability of nurse-based bedside diagnosis of acute MB determinations in 159 institutions. Arch Pathol
myocardial infarction by rapid dry-strip creatine Lab Med 2004;128:15864.
CARDIAC BIOMARKER POCT 489

[107] Panteghini M, Apple FS, Christenson RH, et al. [119] Beyerle K. POC testing of cardiac markers enhan-
Use of biochemical markers in acute coronary syn- ces ED care. Nurs Manage 2002;33:379.
dromes. Clin Chem Lab Med 1999;37:68793. [120] Panteghini M. The measurement of cardiac
[108] Wu AHB, Apple FS, Gibler WB, et al. National markers: where should we focus? Am J Clin Pathol
Academy of Clinical Biochemistry standards for 2002;118:35461.
laboratory practice: recommendations for the use [121] Panteghini M. Acute coronary syndrome: bio-
of cardiac markers in coronary diseases. Clin chemical strategies in the troponin era. Chest
Chem 1999;45:110421. 2002;122:142835.
[109] Kamineni R, Alpert JS. Acute coronary syn- [122] Goodacre S, Calvert N. Cost eectiveness of diag-
dromes: initial evaluation and risk stratication. nostic strategies for patients with acute, undieren-
Prog Cardiovas Dis 2004;46:37992. tiated chest pain. Emer Med J 2003;20:42933.
[110] Lambrew CT, Smith MS, Annas GG, et al. Nation- [123] Gibler WB, Hoekstra JW, Weaver WD, et al. A
al Heart Attack Alert Program Coordinating Com- randomized trial of the eects of early cardiac se-
mittee 60 Minutes to Treatment Working Group. rum marker availability on reperfusion therapy in
Emergency department: rapid identication and patients with acute myocardial infarction: the serial
treatment of patients with acute myocardial infarc- markers, acute myocardial infarction and rapid
tion Bethesda (MD): National Institutes of Health; treatment trial (SMARTT). J Am Coll Cardiol
1993. Ann Emerg Med 1994;23:31129. 2000;36:15006.
[111] Storrow AB. NACB recommendationsdlogistics [124] Fermann GJ, Suyama J. Point of care testing in the
(administrative, cost-eectiveness, point-of-care, emergency department. J Emer Med 2002;22:
what are the rules?) Laboratory Medicine 393404.
Practice Guidelines (LMPG)dbiomarkers of acute [125] Kugelmass A, Anderson A, Katz M, et al. Point-of-
coronary syndrome and heart failure. Available care biomarkers: does it make a dierence in acute
at: http://www.nacb.org/lmpg/card_biomarkers_ myocardial infarction outcomes? Circulation 2004;
LMPG_ draft_PDF.stm. Accessed September 23, 109:50 [abstract no. P248].
2005. [126] Dadkhah S, Fisch C, Zonia C, et al. Accelerated
[112] Males RG, Stephenson J, Harris P. Cardiac coronary reperfusion through the use of rapid bed-
markers and point-of-care testing: a perfect t. side cardiac markers. Case Reports. Angiology.
Crit Care Nurs Q 2001;24:5461. J Vascular Dis 1999;50:5562.
[113] Henderson AR, Gerhardt W, Apple FS. The use of [127] Edmond JJ, French JK, Henny H, et al. Prospective
biochemical markers in ischemic heart disease: evaluation of a chest pain pathway at Green Lane
summary of the roundtable and extrapolations. Hospital. N Z Med J 2002;115:U103.
Clin Chim Acta 1998;272:93100. [128] Polancyzk CA, Kuntz KM, Sacks DB, et al. Emer-
[114] Kellett J, Hirschl MM, Derhaschnig U, et al. Bed- gency department triage strategies for acute chest
side testing of cardiac troponin T and myoglobin pain using creatine kinase-MB and troponin I
for the detection of acute myocardial infarction in assays: a cost-eectiveness analysis. Ann Intern
patients with a nondiagnostic electrocardiogram Med 1999;131:90918.
in the emergency department. Point of Care: the [129] de Winter RJ, Koster RW, Sturk A, et al. Value of
Journal of Near-Patient Testing and Technology myoglobin, troponin T, and CK-MB mass in ruling
2004;3:15961. out an acute myocardial infarction in the emergen-
[115] Derhaschnig U, Hirschl MM, Collinson PO, et al. cy room. Circulation 1995;92:34017.
Diagnostic eciency of a point-of-care system for [130] Mant J, McManus RJ, Oakes RA, et al. Systematic
quantitative determination of troponin T and myo- review and modeling of the investigation of acute
globin in the coronary care unit. Point of Care: the and chronic chest pain presenting in primary care.
Journal of Near-Patient Testing and Technology Health Tech Assess 2004;8:1158.
2004;3:1624. [131] Newby LK, Storrow AB, Gibler WB, et al. Bedside
[116] Gomez MA, Anderson JL, Karagounis LA, et al. multimarker testing for risk stratication in chest
An emergency department-based protocol for rap- pain units. The chest pain evaluation by creatine
idly ruling out myocardial ischemia reduces hospi- kinase-MB, myoglobin, and troponin I (CHECK-
tal time and expense: results of a randomized study MATE) study. Circulation 2001;103:18327.
(ROMIO). J Am Coll Cardiol 1996;28:2533. [132] Apple FS, Jae AS. Bedside multimarker testing
[117] Anderson FP, Fritz ML, Kontos MC, et al. Cost- for risk stratication in chest pain units: the chest
eectiveness of cardiac troponin I in a systematic pain evaluation by creatine kinase-MB, myoglobin,
chest pain evaluation protocol: use of cardiac tro- and troponin I (CHECKMATE) study [Letter and
ponin I lowers length of stay for low-risk patients. Newby response]. Circulation 2001;104:E125.
Clin Lab Manage Rev 1998;12:639. [133] Eggers KM, Oldgren J, Nordenskjold A, et al. Di-
[118] Blomkalns AL, Gibler WB. Markers and the initial agnostic value of serial measurement of cardiac
triage and treatment of patients with chest pain. markers in patients with chest pain: limited value
Cardiovasc Toxicol 2001;1:1115. of adding myoglobin to troponin I for exclusion
490 KOST & TRAN

of myocardial infarction. Am Heart J 2004;148: [140] Blick KE. Economics of point-of-care (POC) test-
57481. ing for cardiac markers. In: the value of critical
[134] Cummings JP, Kost GJ, Matuszewski KA, Ratko care and point of care testing: who benets? Pro-
TA (contributing reviewers). UHC high-impact ceedings of the 20th International Symposium, Crit-
technology brief: B-type natriuretic peptide blood ical and Point of Care Testing Division, American
test. Oak Brook (IL): University HealthSystem Association for Clinical Chemistry. Wurzburg,
Consortium; 2003. Germany. June 912, 2004.
[135] Mueller C, Scholer A, Laule-Kilian K, et al. Use of [141] Berenson RA, Kuo S, May JH. Medical malprac-
B-type natriuretic peptide in the evaluation and tice liability crisis meets markets: stress in unexpect-
management of acute dyspnea. N Engl J Med ed places. Issue Brief Cent Stud Health Syst Change
2004;350:64754. 2003;68:17.
[136] Fonarow GC, Horwich TB. Combining natriuretic [142] Amon E, Winn HN. Review of professional medi-
peptides and necrosis markers in determining prog- cal liability insurance crisis: lessons learned from
nosis in heart failure. Rev Cardiovasc Med 2003; Missouri. Am J Obstet Gynecol 2004;190:153440.
(Suppl 4):S208. [143] Pope HJ, Aufderheide TP, Ruthazer R, et al. Missed
[137] Gibler WB, Blomkalns AL, Collins SP. Evaluation diagnosis of acute cardiac ischemia in the emergency
of chest pain and heart failure in the emergency de- department. N Engl J Med 2000;342:116370.
partment: impact of multimarker strategies and B- [144] Vukmir RB. Medical malpractice: managing the
type natriuretic peptide. Rev Cardiovasc Med risk. Med Law 2004;23:495513.
2003;4(Suppl 4):S4755. [145] Worthington K. Customer satisfaction in the emer-
[138] de Lemos JA, Morrow DA. Combining natriuretic gency department. Emerg Med Clin North Am
peptides and necrosis markers in the assessment of 2004;22:87102.
acute coronary syndromes. Rev Cardovasc Med [146] Clinical Laboratory and Standards Institute
2003;4(Suppl 4):S3746. (CLSI). Protocols for Determination of Limits of
[139] James SK, Armstrong P, Barnathan E, et al. Tro- Detection and Limits of Quantitation. NCCLS-
ponin and C-reactive protein have dierent rela- IFCC (ISO) Joint Project. Wayne (PA): CLSI
tions to subsequent mortality and myocardial EP17-A; 2005.
infarction after acute coronary syndrome: a GUS- [147] Kost GJ, Ehrmeyer SS, Chernow B, et al. The lab-
TO-IV substudy. J Am Coll Cardiol 2003;41: oratory-clinical interface: point-of-care testing.
91624. Chest 1999;115:114054.
Cardiol Clin 23 (2005) 491501

Markers of Cardiac Ischemia and Inammation


Tracy Y. Wang, MD, MSa, Wael A. AlJaroudi, MD, MSb,
L. Kristin Newby, MD, MHSc,*
a
Division of Cardiology, Box 31246, Duke University Medical Center, Durham, NC 27710, USA
b
Department of Medicine, Box 31184, Duke University Medical Center, Durham, NC 27710, USA
c
Duke Clinical Research Institute, P.O. Box 17969, Durham, NC 27715-7969, USA

Coronary artery disease (CAD) is one of the occurs. Insights into the pathophysiology of
leading causes of death in the United States. The atherothrombosis have allowed development of
challenge is not only to diagnose acute coronary novel markers to detect not only early ischemia
syndromes (ACSs) in patients who present with without myocyte death, but also early indicators
the traditional combination of anginal pain and of coronary inammation in patients with pre-
electrocardiography (ECG) changes but also to clinical atherosclerosis.
recognize disease in patients who have atypical
symptoms and initially nondiagnostic ECGs. Ad- Inammation and atherothrombosis
vances in the understanding of the pathophysiol-
There is extensive literature supporting the role
ogy of cellular injury and in analytical technology
of inammation in CAD. Hemodynamic forces
have made possible the development of molecular
from hypertension and oxidative stressors, such as
markers that can be measured in serum or plasma.
tobacco and hyperglycemia, result in vascular
This development has already made a tremendous
endothelial injury. The attachment of leukocytes,
impact on our clinical assessment of myocardial
transformation of monocytes into macrophages,
injury. In September 2000, a committee compris-
and subsequent uptake of cholesterol lipoproteins
ing members representing the American College
initiate the fatty streak. Cytokine release from the
of Cardiology and the European Society of Cardi-
fatty streak recruits further inammatory cells
ology redened acute myocardial infarction (MI),
(macrophages, mast cells, activated T cells), with
including the manner in which biomarkers of myo-
resulting uptake and oxidation of low-density
cardial necrosis, preferentially cardiac troponins,
lipoprotein (LDL). These cytokines also stimulate
are used in addition to ischemic symptoms and
smooth muscle cell proliferation and development
ECG changes [1].
of a collagenous brous cap that covers this
Because current biomarkers of myocardial
inammatory mixture to form the mature athero-
necrosis only become positive in the setting of
sclerotic plaque [2].
MI and disruption of cellular integrity, the di-
Recent data show that the risk for an acute
agnosis of MI can only be made in retrospect.
coronary event has less to do with the degree of
Ideally, one would like to identify patients at risk
angiographic luminal stenosis than with the un-
for complications before myocardial necrosis
derlying pathology of the atherosclerotic plaque
that makes it susceptible to rupture [3]. Activated
T cells within the atheromatous core secrete inter-
Dr. Newby has received consulting honoraria from
feron gamma that decreases smooth muscle cell
Ischemia Diagnostics, Inc. and Ortho-Clinical
Diagnostics and research grant support from Roche
production of collagen [4]. Inammatory cyto-
Diagnostics Corporation. kine-activated macrophages secrete matrix metal-
* Corresponding author. loproteinases that degrade the extracellular
E-mail address: newby001@mc.duke.edu matrix, further weakening the brous cap and
(L.K. Newby). making it prone to rupture [5,6]. Disruption of
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.007 cardiology.theclinics.com
492 WANG et al

the cap exposes the atheronecrotic core to blood, independent of troponin levels; there was no asso-
after which further inammatory reactions lead ciation with recurrent MI [14]. However, in stable
to platelet activation, coagulation cascade, further post-MI patients, elevated hsCRP predicted a sig-
vasomotor dysfunction [7,8], and ultimately lumi- nicantly higher risk for recurrent nonfatal MI or
nal occlusion. fatal coronary events (75% higher in the highest
Thus, all stages of atherothrombosis implicate versus lowest quintile of hsCRP), suggesting that
inammation as a key pathogenic mechanism. it is not merely a marker for the extent of myocar-
Several clinical studies have therefore targeted dial damage [15].
inammatory factors as potential markers for In the European Concerted Action on Throm-
cardiovascular risk assessment. bosis and Disabilities Angina Pectoris Study, 2121
patients admitted with angina were followed over
a 2-year period. A plasma concentration of CRP
Markers of inammation greater than 3.6 mg/L at study entry was associ-
ated with a 45% increase in the relative risk for
Several inammatory factors involved in the
nonfatal MI or sudden cardiac death (95% con-
cascade described earlier have been studied in
dence intervale [CI], 1.151.83) [16]. However, the
clinical trials in the last decade. These include the
usefulness of measuring hsCRP in patients who
proinammatory cytokines interleukin (IL)-1, IL-
have ACS is limited because there is already su-
6, tumor necrosis factor a), catalysts of the inam-
cient evidence warranting maximal lipid-lowering,
matory reaction (CD40 ligand), cellular adhesion
antiplatelet, and other cardioprotective drug ther-
molecules (intercellular cell adhesion molecule
apies in these patients, so that measurement of
[ICAM]-1, vascular cell adhesion molecule
hsCRP does not provide incremental information.
[VCAM]-1, selectins), and acute phase reactants
The setting in which hsCRP may be most
(brinogen, white blood count, C-reactive protein
useful is primary prevention. A population-based
[CRP], serum amyloid A).
cross-sectional study in Great Britain showed that
the prevalence of CAD increased 1.5 fold (95%
C-reactive protein
CI, 1.251.92) for each doubling of hsCRP among
CRP is the best studied of the inammatory men aged 50 to 69 years [17]. In the Multiple Risk
markers in cardiovascular disease. It is an acute- Factor Intervention Trial cohort of middle-aged
phase protein that has been shown to be a marker men who had traditional high cardiovascular
of systemic inammation, elevated in response to risk factors, CRP was elevated more in smokers
acute injury, infection, and other inammatory versus nonsmokers. Over 17 years of follow-up,
stimuli [9]. Hepatic production is directly related an elevated baseline CRP was associated with
to IL-6 stimulation and, unlike other acute phase a 2.8-fold increased risk for coronary heart disease
reactants, its levels remain stable over long peri- mortality (95% CI, 1.45.4) [18].
ods of time in the absence of new stimuli [10]. Tra- Several prospective studies of hsCRP in appar-
ditional CRP assays with limits of quantication ently healthy individuals have also shown that
of 3 to 8 mg/L lack adequate sensitivity to detect elevated baseline levels of hsCRP are correlated
levels required for atherosclerotic risk prediction. with higher risk for future cardiovascular morbid-
The development of a standardized high-sensitivity ity and mortality after adjustment for potential
CRP (hsCRP) assay has improved precision at confounders. The Physicians Health Study, a pro-
low concentrations of CRP that permits its use spective, nested case-control study of men who did
in cardiovascular risk assessment. not have prior history of CAD and had low rates of
In several studies of patients who have ACS cigarette use, showed that men in the highest
(Chimeric c7E3 AntiPlatelet Therapy in Unstable quartile of hsCRP (R2.1 mg/L) had a signicant
angina REfactory to standard treatment, Throm- 2.9-fold increase in risk for MI that was indepen-
bolysis In Myocardial Infarction [TIMI]-11a, and dent of smoking status, lipid levels, and other tra-
FRagmin during InStability in Coronary artery ditional risk factors for CAD [19]. This nding was
disease [FRISC] trials), elevated hsCRP at hospi- conrmed in the Monitoring Trends and Deter-
tal admission independently predicted increased minants of Cardiovascular Disease (MONICA)
mortality [1113]. In a Global Use of Strategies to Augsburg prospective study in Europe that
open Occluded arteries IV substudy, hsCRP ele- followed 936 healthy, middle-aged men over 8 years
vation during the acute stage of unstable CAD and noted a 19% increase in risk for future cor-
was associated with an increased 30-day mortality onary event for each standard deviation increase
CARDIAC ISCHEMIA & INFLAMMATION MARKERS 493

in baseline hsCRP after adjustment for multiple assessed by electron-beam CT was similar across
risk factors [20]. all hsCRP quartiles [27]. Therefore, although
This relationship also bears out in women. In hsCRP has good risk assessment value, there is
a prospective nested case-control study involving no denitive evidence supporting its role in select-
postmenopausal women enrolled in the Womens ing patients for coronary angiography.
Health Study, Ridker and colleagues [21] showed The role of hsCRP in predicting CAD has been
hsCRP to be the most powerful predictor of cardio- established. What remains uncertain at this time is
vascular risk compared with other inammatory whether CRP should be a target of therapy or
markers, baseline lipid levels, and homocysteine. used to guide therapy. In vitro and in vivo studies
Women in the highest quartile had a relative risk show that hsCRP contributes to plaque develop-
of 4.4 (95% CI, 2.28.9, P ! .001) compared ment by increasing monocyte adherence, inducing
with those in the lowest quartile. Addition of expression of cell surface adhesion molecules [28],
hsCRP to cholesterol measurement increased the and increasing LDL scavenger cell uptake of cho-
area under the receiver operating characteristic lesterol [29]. Furthermore, high CRP is associated
(ROC) curve from 0.59 to 0.66 (P ! .001). Further- with decreased nitric oxide availability impacting
more, in women who had LDL levels less than 130 vasomotor activity [30] and increased monocyte
mg/dL (the target level recommended for primary production of tissue factor [31]. CRP is also
prevention by the National Cholesterol Education shown to activate complement and neutrophils
Program), those who had elevated baseline CRP [32] and decrease brinolytic capacity [33], there-
were still at increased risk for future events with by contributing to plaque instability and throm-
a 3.1 relative risk in the highest quartile compared bus formation.
with the lowest (95% CI, 1.711.3, P .002) after In the Physicians Health Study, use of aspirin
adjustment for traditional risk factors and high- was associated with a statistically signicant,
density lipoprotein (HDL) cholesterol levels. 55.7% risk reduction for rst MI among men
Prospective data show that hsCRP levels who had hsCRP levels in the highest quartile
minimally correlate with lipid levels and are compared with a nonsignicant 13.9% reduction
a stronger predictor of risk than LDL cholesterol in those who had hsCRP levels in the lowest
(area under the ROC curve 0.64 versus 0.60) [22]. quartile [19]. No follow-up levels were drawn to
Therefore, the role of CRP is adjunctive to lipid deduce whether aspirin directly lowered hsCRP
screening and the additive predictive value of levels.
hsCRP to the Framingham 10-year risk score, The anti-inammatory eects of statins are still
LDL, total, and HDL cholesterol measurements unclear; experimental evidence posits statin-
has been demonstrated in several studies [23,24]. mediated reduction in macrophage activation,
However, there is not yet a consensus that the de- antiproliferative eects on smooth muscle cells, im-
gree of hsCRP elevation correlates with athero- provements in endothelial function and vasomo-
sclerotic burden. tion, and antithrombotic eects [34,35]. At 5-year
Tataru and colleagues [25] showed that in survi- follow-up of stable post-MI patients randomized
vors of MI, there was a signicant association of in the Cholesterol and Recurrent Events trial,
hsCRP level with angiographically detected degree mean hsCRP levels decreased by 37.8% in those
of coronary artery stenosis. In addition, patients randomized to pravastatin therapy, whereas levels
who had coronary disease who had sonographi- increased in the placebo patients [36]. This reduc-
cally detectable peripheral arterial disease had tion in hsCRP was independent of the magnitude
even higher levels of hsCRP compared with those of changes in lipid levels. Furthermore, patients
who had CAD alone. In a small Denmark study randomized to pravastatin therapy had a twofold
of 269 patients referred for elective coronary angi- reduction in risk for recurrent MI or fatal coro-
ography, hsCRP levels were signicantly higher in nary event in the elevated hsCRP group compared
patients who had coronary stenoses than those with the group that did not have elevated hsCRP,
who did not, but no dierence in hsCRP levels even though baseline lipid levels were identical in
were found comparing groups with single, 2-, or both groups [15].
3-vessel disease [26]. In a nested case-control sub- More recently, PRavastatin Or atorVastatin
study of the Prospective Army Coronary Calcium Evaluation and Infection Therapy (PROVE-IT)
trial that looked at healthy men between the age of trial demonstrated that intensive statin therapy
40 and 45 years who did not have coronary disease, that lowered hsCRP levels to a mean of less than 2
the prevalence of coronary artery calcium as mg/L resulted in a decreased risk for recurrent MI
494 WANG et al

or death from coronary causes, irrespective of the Lipoprotein-associated phospholipase A2


degree of LDL lowering [37]. In the PRavastatin
Lipoprotein-associated phospholipase A2 (Lp-
Inammation CRP Evaluation (PRINCE) study,
PLA2), previously described as platelet-activating
pravastatin was shown to also reduce hsCRP lev-
factor acetylhydrolase, is a novel marker whose
els in patients who did not have prior history of
role in atherosclerosis has been heavily debated.
cardiovascular disease [38]. There is not yet evi-
Initially considered atheroprotective because of
dence linking hsCRP reduction to a reduction in
its ability to degrade platelet-activating factor,
cardiovascular events in the primary prevention
this enzyme has since been discovered to cleave
setting.
oxidized phosphatidylcholine into lysophosphati-
The Centers for Disease Control and Preven-
dylcholine and oxidized free fatty acids which pro-
tion/American Heart Association consensus state-
mote the inammatory process of atherosclerosis.
ment advocates use of hsCRP for risk assessment
Lp-PLA2 is produced by macrophages and T lym-
in patients who are at intermediate risk for
phocytes and can be detected in human athero-
cardiovascular events (10% to 20% 10-year risk
sclerotic lesions [43]. In humans, Lp-PLA2 is
of coronary event) [39]. However, there is no pro-
predominantly bound to LDL cholesterol par-
spective randomized clinical trial yet examining
ticles and is activated once LDL particles undergo
the benets or harm of screening with hsCRP.
oxidative damage [44].
Presently the guidelines rate a hsCRP level less
The West of Scotland Coronary Prevention
than 1 mg/L as normal, 1 to 3 mg/L intermediate,
Study rst demonstrated that baseline Lp-PLA2
and more than 3 mg/L as high. A level greater than
elevation in hyperlipidemic men predicted risk
10 mg/L indicates a noncardiovascular source of
for coronary events independently of other in-
inammation and should prompt a search for
ammatory markers, such as CRP and brinogen.
a source of infection or other inammation. For
There was a 60% statistically signicant increase
patients who have ACS, hsCRP may provide
in risk between the highest and lowest quintile of
some prognostic information, and a cuto level
Lp-PLA2 [45]. However, in a lower risk popula-
of more than 10 mg/L is most predictive of risk
tion of women (in a nested case-control analysis
for adverse outcomes.
of the Womens Health Study), the predictive abil-
ity of Lp-PLA2 was not statistically signicant af-
ter adjustment for traditional risk factors [46].
Interleukin-6 The Atherosclerosis Risk in Communities study
enrolled men and women who had a wide range of
IL-6 is an inammatory cytokine that induces
LDL levels and found that in patients who had
hepatic synthesis of all the acute phase proteins
LDL levels below 130 mg/dL, Lp-PLA2 was signif-
and is the primary determinant of CRP levels. The
icantly and independently associated with CAD
presence of IL-6expressing macrophages in cor-
[47]. This nding suggests a role for Lp-PLA2, sim-
onary atherosclerotic plaques [40] suggests that in-
ilar to CRP, in identifying high-risk patients who
creased IL-6 levels are not just a marker of
may benet from drug therapy and are not tar-
inammation but also play a direct role in increas-
geted for statin use under the current Adult Treat-
ing plaque vulnerability to ssuring and thrombo-
ment Panel (ATP) III guidelines. In the most recent
sis. In addition to its proinammatory eects, it
study by Brilakis and colleagues [48], Lp-PLA2
also has procoagulant eects by modulating -
levels correlated with extent of angiographic
brinogen synthesis, enhancing platelet adhesion,
CAD; however, this was not independently predic-
and impairing endothelial vasodilation. High IL-
tive after adjusting for CRP, lipid status, and other
6 levels in healthy men correlated with increased
traditional risk factors. Inhibition of Lp-PLA2 in
risk for future MI independently of hsCRP
animal models has been shown to be eective in re-
[41]. In patients who had ACS, IL-6 elevation
versing atherosclerosis in animal models [49], and
(O5 ng/mL) was associated with 3.5-fold higher
phase II trials of SB-480848, a specic Lp-PLA2 in-
1-year mortality than levels less than 5 ng/mL,
hibitor, as a potential treatment of atherosclerosis
independent of troponin and hsCRP [42]. Fur-
are currently underway.
thermore, in this study it appeared that patients
who had inammation as determined by IL-6 lev-
CD40 ligand
els greater than 5 ng/mL had a greater response to
an invasive versus conservative strategy than pa- CD40 ligand (CD40L) is a transmembrane
tients who had IL-6 levels less than 5 ng/mL. protein expressed on CD4 T cells, macrophages,
CARDIAC ISCHEMIA & INFLAMMATION MARKERS 495

and activated platelets. Its interaction with the P-selectin levels in healthy women in the Womens
CD40 receptor on endothelial cells, smooth mus- Health Study were also associated with increased
cle cells, and phagocytes has been shown to cardiovascular risk [57]. The levels of ICAM-1
stimulate expression of proinammatory cyto- and E-selectin, but not VCAM-1, have been shown
kines, cellular adhesion molecules, matrix metal- to correlate with atherosclerotic burden as mea-
loproteinases, and procoagulant tissue factor [50]. sured by carotid ultrasound, suggesting the useful-
Furthermore, CD40 ligand has a KGD sequence ness of circulating adhesion molecules as indicators
that is a known binding motif for platelet integrin of subclinical disease [58]. Despite these intriguing
aIIb3. T cells expressing CD40L are found in ath- data, the challenges of sample handling (samples
erosclerotic plaques [51], and disruption of the are unstable unless frozen) currently limit the clin-
CD40-CD40L interaction in vitro diminishes ath- ical usefulness of these markers.
eroma formation and promotes stabilization of
the established plaque [52].
Biologically active soluble CD40L (sCD40L) Serum amyloid A
released from stimulated lymphocytes can be Like CRP, serum amyloid A (SAA) is an acute-
measured in plasma. Healthy women who have phase protein that is synthesized in the liver in
high levels of sCD40L have been shown to be at response to stress, injury, or inammation. During
increased risk for cardiovascular events [53]. the acute phase, SAA becomes the predominant
In patients who have ACS, sCD40L elevation apolipoprotein on HDL cholesterol, thus altering
(O5 mg/L) correlated with increased 6-month HDL-mediated cholesterol delivery to cells [59]. In
mortality (18.6% versus 7.1%, P ! .001) [54]. the Womens Ischemia Syndrome Evaluation
Treatment of these patients with abciximab before study of women referred for coronary angiogra-
coronary angioplasty reduced risk for death or phy for suspected myocardial ischemia, SAA level
nonfatal MI (hazard ratio 0.12; 95% CI, 0.08 correlated with 3-year risk for cardiovascular
0.49; P!.001). However, patients who had lower events and with angiographic severity of CAD
levels of sCD40L did not experience the same [60].
treatment benet. Among troponin-negative pa-
tients, those who had elevated sCD40L had an in-
creased risk for cardiac events (13.6%) similar to Interleukin-10
that of patients who were troponin-positive
(14%). Treatment of these patients with abcixi- IL-10 is a cytokine whose function is to limit
mab also reduced risk for cardiac events (5.5% the inammatory reaction once the pathogen is
versus 13.6%, P .03). eliminated. Its eects on macrophages include
down-regulation of proinammatory cytokine
production and adhesion molecule expression. It
Selectins also inhibits synthesis of matrix metalloproteinases
Selectins (P-selectin, E-selectin) and cellular implicated in plaque destabilization [61]. Overex-
adhesion molecules (VCAM-1, ICAM-1) mediate pression of human IL-10 compensates for the
the initial leukocyte rolling along the endothelium rapid atherosclerosis of LDL-receptor knockout
and attachment/subintimal transmigration, respec- mice [62]. A small study in the United Kingdom
tively, that are the initial steps of atherosclerosis. showed higher IL-10 levels in patients who had sta-
Immunohistochemical studies have demonstrated ble versus unstable coronary syndrome [63].
expression of these molecules, stimulated by oxi-
dized LDL, in the endothelium overlying the
Markers of ischemia
atherosclerotic plaque. Circulating levels of
VCAM-1, ICAM-1, and P- and E-selectin have Commonly used biomarkers such as creatine
been detected in plasma and are thought to arise kinase (CK), CK-MB, and troponin are not
from proteolytic cleavage from endothelial cells, detected in blood until several hours after symp-
particularly during inammatory conditions [55]. tom onset, and their presence is almost always
In a nested case-control analysis of the Physicians correlated with myocardial necrosis. The chal-
Health Study, risk for future MI was 60% higher lenge is to identify patients who develop symp-
in patients who had soluble ICAM-1 elevation, toms of ischemia, such as plaque rupture with
and this increased risk was independent of smoking partial occlusion of the vessel lumen, before myo-
status or lipid levels [56]. Elevated baseline soluble cyte death. In this early stage of ACS, traditional
496 WANG et al

biomarkers are not yet detectable in blood. Re- Unbound free fatty acid
cent eorts have focused on identifying new
Unbound free fatty acid (u-FFA) is a new
markers of ischemia. The main diculty is the
marker recently found to be elevated in patients
lack of a gold standard to diagnose myocardial is-
who have ACS before more traditional biomarkers
chemia, and many studies rely on clinical diagno-
become positive. It is thought to be released as
sis with inconsistent objective data end points.
a marker of ischemia and plaque instability, with
greater than 90% sensitivity [68]. In patients un-
Ischemia modied albumin dergoing PCI, the elevation of u-FFA correlated
with ST-segment changes and PCI-induced tran-
Human serum albumin is a complex protein
sient ischemic changes [69]. The clinical applicabil-
with an N-terminus capable of binding metals such
ity of this test still needs to be determined and
as cobalt. During ischemic events, free radicals are
validated in a larger population.
released, resulting in acetylation of the N-terminus
and alteration of the cobalt binding site. There-
B-type (brain) natriuretic peptide
fore, ischemia-modied albumin (IMA) is unable
to bind cobalt, and the free metal can be detected B-type natriuretic peptide (BNP) is a peptide
using a specic assay. The higher the amount of hormone released from cardiac ventricles in re-
free cobalt detected, the greater the degree of sponse to myocardial stretch or increased wall
ischemia [64]. The advantage of this test is that tension. Its actions include vasodilation, natriure-
IMA becomes positive within a few minutes of the sis, and inhibition of the renin-angiotensin aldo-
ischemic event. A positive test despite negative tro- sterone and sympathetic nervous systems [70].
ponin places the patient at high risk for a cardiac BNP is produced as a prohormone, pro-BNP,
event [65]. It has a sensitivity of 83%, specicity which is enzymatically cleaved into BNP and N-
of 69%, negative predictive value (NPV) of 96%, terminal pro-BNP (NT-pro-BNP). Initially identi-
and a positive predictive value (PPV) of 33%. Sen- ed as highly sensitive diagnostic markers for con-
sitivity increases to 92% when combined with pos- gestive heart failure, BNP and NT-pro-BNP levels
itive ECG ndings, and 95% when combined with are highly correlated, although NT-pro-BNP may
positive troponin and ECG. Conversely, a combi- be slightly more sensitive for ventricular dysfunc-
nation of negative IMA, negative troponin, and tion. BNP and NT-pro-BNP have been found to
negative ECG has 99% NPV of a cardiac event predict increased mortality independently of left
[66]. Rapid hepatic clearance (6 hours) allows ventricular function when elevated in patients
specicity for an acute event, but the main disad- who have acute MI and unstable angina [7173].
vantage is that IMA is not specic for cardiac is- In patients who have unstable angina or non
chemia; any ischemia can result in a positive test ST-elevation MI, a BNP greater than 80 pg/mL
[67]. It can also be elevated in patients who have was associated with angiographically tighter cul-
cancer, acute infections, cirrhosis, and end-stage prit lesions or left anterior descending coronary
renal disease. artery involvement, implicating more severe ische-
The Food and Drug Administration recently mia as a cause of increased mortality [74]. How-
approved the Albumin-Cobalt Binding Test, which ever, the role of BNP levels in terms of therapy
measures ischemia-modied albumin, for use in decisions in ACS is still unclear. In the Treat An-
evaluating patients who have suspected myocar- gina with Aggrastat and Determine Cost of Ther-
dial ischemia. In its labeling, this test is intended apy with Invasive or Conservative Strategy
to be used in conjunction with troponin testing and (TACTICS-TIMI 18) population, an elevated
ECG evaluation to facilitate risk stratication of baseline BNP did not predict signicant benet
patients who have chest pain or other symptoms of from an early invasive management strategy [75].
suspected ischemic origin. Based on its sensitivity However, in the FRISC II study, baseline NT-
and specicity prole and excellent NPV, a nega- pro-BNP in concert with IL-6 elevation were use-
tive ischemia-modied albumin level in conjunc- ful for identifying patients who derive survival
tion with a negative troponin and nondiagnostic benet from an early invasive strategy [76].
ECG can rule out ACS in low-risk patients. Thus, Recent data have shown that BNP is elevated
it is ideal for use in chest pain units. The results of in the setting of myocardial ischemia even in the
an ongoing prospective study are expected to absence of necrosis. Plasma NT-pro-BNP levels
provide more information on the prognostic value are higher in patients who have unstable angina
of a positive IMA result. compared with age-matched controls who have
CARDIAC ISCHEMIA & INFLAMMATION MARKERS 497

stable angina or no coronary disease [77]. Patients specicity of 86%. Unlike troponin, choline is
who have stable CAD and BNP elevation are not a marker of myocardial cell necrosis. Its use
more likely to have inducible ischemia on exercise as an early biomarker of MI is limited, with a sen-
treadmill testing [78]. BNP increases transiently sitivity of 41%, specicity of 79%, PPV of 51%,
after exercise in patients who have stable CAD and NPV of 71%. The weak association of choline
and is correlated with size of ischemic territory with ST-elevation MI has been attributed to the
as assessed by single-photon emission computed possibility a dierent mechanism of plaque rup-
tomography [79]. In patients undergoing PCI, ture, whereby rapid red thrombus formation
plasma BNP levels increase immediately after bal- leaves limited time for the injured endothelium
loon ination [80] and remain elevated 24 hours and collagen tissue to activate phospholipase D
postintervention independent of intracardiac ll- [85]. Intermittent formation of nonoccluding
ing pressures [81]. white thrombi in the setting of unstable angina
Experimental studies conrm that acute car- leads to phospholipase D activation in platelets
diac hypoxia induces BNP expression in the and choline release. Therefore, choline may pro-
ventricular myocardium [82] and BNP is released vide early detection of patients who have high-
in the setting of coronary artery occlusion [83]. risk unstable angina, indicating subendocardial
In the rat heart model, the degree of BNP eleva- rather than transmural injury.
tion correlates with duration of ischemia (induced
up to 20 minutes by reversible ligation of the left Pregnancy-associated plasma protein-A
main coronary artery). Exogenous infusion of
Another emerging marker of plaque instability
BNP limits infarct size in a dose-dependent man-
is PAPP-A, a zinc-binding matrix metalloprotei-
ner, suggesting that BNP is released in response
nase that activates insulin-like growth factor I
to ischemia such that its vasodilatory eect at-
(IGF-I), a mediator of atherosclerosis. IGF-I
tenuates ischemic injury. These basic and clinical
induces migration of vascular smooth muscle
data suggest that plasma measurement of BNP
cells, monocyte chemotaxis, and release of cyto-
and NT-pro-BNP can be useful in detecting myo-
kines. Histologically, PAPP-A is abundantly ex-
cardial ischemia. However, this nding still needs
pressed in ruptured or eroded plaques but not in
to be validated in larger studies.
stable plaques, suggesting a potential enzymatic
role in extracellular matrix degradation and -
brous cap weakening that contributes to plaque
Markers of plaque instability or rupture
rupture [86]. Circulating levels of PAPP-A were
Recognition of the role of plaque rupture in signicantly higher in patients who had unstable
the pathogenesis of cardiac ischemia has led to angina or MI compared with patients who did
investigation of serum markers for plaque in- not have an ACS; the areas under the receiver op-
stability. These markers include whole blood erating curve were 0.94 and 0.88, respectively, at
choline levels, pregnancy-associated plasma pro- the threshold level of 10 mIU/L. The sensitivity
tein-A (PAPP-A), and malondialdehyde-modied and specicity were 89.2% and 81.3%, respec-
LDL. tively. PAPP-A levels did not correlate with mark-
ers of cardiac necrosis such as troponin and
Whole blood choline CK-MB, suggesting its role as a marker of early
ischemia but not infarction [86]. In troponin-neg-
Choline is thought to be released from leuko-
ative patients presenting with ACS, those who had
cytes and platelets in response to activation of
elevated PAPP-A levels had a 4.6-fold higher risk
phospholipase D during plaque rupture. In pa-
for cardiovascular morbidity at 6 months com-
tients presenting with symptoms of ACS and
pared with those who had low PAPP-A levels [87].
negative troponin levels on admission, an elevated
choline level predicts higher risk for cardiac death,
Malondialdehyde-modied low-density lipoprotein
nonfatal cardiac arrest, life-threatening arrhyth-
mias, heart failure, and future angioplasty within During ischemic injury of the endothelium,
30 days [84]. Furthermore, in the absence of MI, a cascade of platelet adhesion and activation
choline detects patients who have high-risk unsta- involving the prostaglandin pathway results in
ble angina (prolonged chest pain, ischemia-related the release of aldehydes that bind to the apo-B100
pulmonary edema, rest pain with dynamic ECG moiety of LDL, forming malondialdehyde
changes, or hypotension) with a sensitivity and (MDA)-modied LDL [88]. Plasma levels of
498 WANG et al

MDA-modied LDL become elevated in patients myocardial damage and other complications. The
who have ACS within the rst 6 hours of symp- challenge for clinicians will be to integrate marker
tom onset. In patients presenting with ACS, the testing into practice without overwhelming, or
sensitivity of admission MDA-modied LDL substituting for, clinical judgment and rational
was 95% compared with 38% for troponin I in decision making.
the diagnosis of unstable angina [89], and the
specicity was 95%. The use of troponin I and
MDA-modied LDL in combination achieved References
a sensitivity of 98%. MDA-modied LDL levels
[1] Alpert JS, Thygesen K, Antman E, et al. Myocardial
were threefold higher in patients who had unsta- infarction redenedda consensus document of the
ble angina compared with those who had stable Joint European Society of Cardiology/American
CAD. In acute MI, MDA-modied LDL is 95% College of Cardiology Committee for the redeni-
sensitive when used alone, troponin I is 90% tion of myocardial infarction. J Am Coll Cardiol
when used alone, and the combination had a sen- 2000;36:95969.
sitivity approaching 100% [89]. This evidence [2] Mantovani A, Bussolino F, Dejana E. Cytokine reg-
suggests that MDA-modied LDL, reecting en- ulation of endothelial cell function. FASEB J 1992;6:
dothelial injury and plaque rupture, is useful in 25919.
discriminating unstable from stable angina, in [3] Libby P. Molecular bases of the acute coronary syn-
dromes. Circulation 1995;91:284450.
contradistinction to troponin, which helps to dis-
[4] Amento EP, Ehsani N, Palmer H, et al. Cytokines
criminate between unstable angina and myocar- and growth factors positive and negatively regular
dial infarction. interstitial collagen gene expression in human vascu-
lar smooth muscle cells. Arterioscler Thromb Vasc
Biol 1991;11:122330.
[5] Galis ZS, Muszynski M, Sukhova GK, et al. Cyto-
Summary
kine stimulated human vascular smooth muscle cells
Recent advances have identied many promis- synthesize a complement of enzymes required for ex-
ing new molecular markers of ischemia and in- tracellular matrix digestion. Circ Res 1994;75:1819.
ammation. Their clinical usefulness requires [6] Galis ZS, Sukhova GK, Lark MW, et al. Increase ex-
pression of matrix metalloproteinases and matrix
satisfaction of the following criteria: they must
degrading activity in vulnerable regions of human
(1) be easily available, simple, and standardized atherosclerotic plaques. J Clin Invest 1994;94:
assay, (2) have trial-based evidence of association 2493503.
with clinical end points, (3) be generalizable to [7] Moreno PR, Bernardi VH, Lopez-Cuellar J, et al.
various population groups, and (4) demonstrate Macrophages, smooth muscle cells, and tissue factor
adequate sensitivity and specicity. in unstable angina: Implications for cell-mediated
There is a growing recognition of the role of thrombogenicity in acute coronary syndromes. Cir-
hsCRP in further risk assessment, particularly in culation 1996;94:30907.
patients at intermediate cardiovascular risk and [8] Neri Serneri G, Abbate R, Gori A, et al. Transient
for risk stratication in patients who have ACS. intermittent lymphocyte activation is responsible
for the instability of angina. Circulation 1992;86:
However, it has limited diagnostic usefulness in
7907.
chest pain patients. Ischemia-modied albumin is [9] Deodhar SD. C-reactive protein: the best laboratory
now available clinically to rule out acute coronary indicator available for monitoring disease activity.
ischemia in conjunction with standard troponin Cleve Clin J Med 1989;56:12630.
and ECG evaluation, and whole blood choline [10] Macy E, Hayes T, Tracy R. Variability in the mea-
and MDA-modied LDL are two markers of surement of C-reactive protein in healthy subjects:
plaque instability that have been shown in small Implications for reference interval and epidemio-
trials to diagnose unstable angina with good logic applications. Clin Chem 1997;43:528.
sensitivity and specicity. However, larger clinical [11] Heeschen C, Hamm CW, Bruemmer J, et al. Predic-
trials will be needed to validate their clinical value. tive value of C-reactive protein and troponin T in
patients with unstable angina: a comparative analy-
As understanding of the pathophysiology of
sis. CAPTURE Investigators. Chimeric c7E3 Anti-
atherothrombosis evolves, the number of candi- Platelet Therapy in Unstable angina REfractory to
date markers meriting clinical consideration will standard treatment trial. J Am Coll Cardiol 2000;
increase exponentially. The hope is that these 35:153542.
markers will permit earlier detection of disease, [12] Morrow DA, Rifai N, Antman EM, et al. C-reactive
allowing time for interventions to circumvent protein is a potent predictor of mortality
CARDIAC ISCHEMIA & INFLAMMATION MARKERS 499

independently of and in combination with troponin [25] Tataru MC, Heinrich J, Junker R, et al. C-reactive
T in acute coronary syndromes: a TIMI 11a sub- protein and the severity of atherosclerosis in myo-
study. J Am Coll Cardiol 1998;31:1460. cardial infarction patients with stable angina pecto-
[13] Lindahl B, Toss H, Siegbahn A, et al. Markers of ris. Eur Heart J 2000;21:10008.
myocardial damage and inammation in relation [26] Masden T, Skou HA, Hansen VE, et al. C-reactive
to long-term mortality in unstable coronary artery protein, dietary n-3 fatty acids, and the extent of cor-
disease. FRISC Study Group. Fragmin during In- onary artery disease. Am J Cardiol 2001;88:113942.
stability in Coronary Artery Disease. N Engl J [27] Hunt ME, OMalley PG, Vernalis MN, et al. C-reac-
Med 2000;343:113947. tive protein is not associated with the presence or
[14] James S, Armstrong P, Barnathan E, et al. Troponin extent of calcied subclinical atherosclerosis. Am
and C-reactive protein have dierent relations to Heart J 2001;141:20610.
subsequent mortality and myocardial infarction af- [28] Pasceri V, Willerson J, Yeh E. Direct pro-inamma-
ter acute coronary syndrome: a GUSTO-IV study. tory eect of C-reactive protein on human endothe-
J Am Coll Cardiol 2003;41:91624. lial cells. Circulation 2000;102:21658.
[15] Ridker P, Rifai N, Pfeer M, et al. Inammation, [29] Zwaka T, Homback V, Torzewski J. C-reactive pro-
pravastatin, and the risk of coronary events after tein-mediated low density lipoprotein uptake by
myocardial infarction in patients with average cho- macrophages: implications for atherosclerosis. Cir-
lesterol levels. Circulation 1998;98:83944. culation 2001;103:11947.
[16] Haverkate F, Thompson SG, Pyke SD, et al. Pro- [30] Venugopal S, Devaraj S, Yuhanna I, et al. Demon-
duction of C-reactive protein and risk of coronary stration that C-reactive protein decreases ENOS ex-
events in stable and unstable angina. European pression and bioactivity in human aortic endothelial
Concerted Action on Thrombosis and Disabilities cells. Circulation 2002;106:143941.
Angina Pectoris Study Group. Lancet 1997;349: [31] Cermak J, Key NS, Bach RR, et al. C-reactive pro-
4626. tein induces human peripheral blood monocytes to
[17] Mendall MA, Patel P, Ballam L, et al. C-reactive synthesize tissue factor. Blood 1993;82:51320.
protein and its relation to cardiovascular risk fac- [32] Wolbink GJ, Brouwer MC, Buysmann S, et al.
tors: a population base cross sectional study. BMJ CRP-mediated activation of complement in-vivo:
1996;312:10615. assessment by measuring circulating complement-
[18] Kuller LH, Tracy RP, Shaten J, et al. Relation of C- C-reactive protein complexes. J Immunol 1996;157:
reactive protein and coronary heart disease in the 4739.
MRFIT nested case control study. Multiple Risk [33] Deveraj S, Xu D, Jialal I. C-reactive protein
Factor Intervention Trial. Am J Epidemiol 1996; increases plasminogen activator inhibitor-1 expres-
144:53747. sion and activity in human aortic endothelial cells:
[19] Ridker PM, Cushman M, Stampfer MJ, et al. In- implications for the metabolic syndrome and athero-
ammation, aspirin and the risk of cardiovascular thrombosis. Circulation 2003;107:398404.
disease in apparently healthy men. N Engl J Med [34] Vaughan CJ, Murphy MB, Buckley BM. Statins do
1997;336:9739. more than just lower cholesterol. Lancet 1996;348:
[20] Koenig W, Sund M, Frohlich M, et al. C-reactive 107982.
protein, a sensitive marker of inammation, predicts [35] Rosenson RS, Tangney CC. Antiatherothrombotic
future risk of coronary heart disease in initially properties of statins: Implications for cardiovascular
healthy middle aged men. Circulation 1999;99: event reduction. JAMA 1998;279:164350.
23742. [36] Ridker P, Rifai N, Pfeer M, et al. Long term eects
[21] Ridker P, Hennekens CH, Buring JE, et al. C-reac- of pravastatin on plasma concentration of C-reac-
tive protein and other markers of inammation in tive protein. Circulation 1999;100:2305.
the prediction of cardiovascular disease in women. [37] Ridker PM, Cannon CP, Morrow D, et al. C-reac-
N Engl J Med 2000;342:83643. tive protein levels and outcomes after statin therapy.
[22] Ridker P, Rifai N, Rose L, et al. Comparison N Engl J Med 2005;352:208.
of C-reactive protein and low-density lipoprotein [38] Albert MA, Danielson E, Rifai N, et al. Eect of sta-
cholesterol levels in the prediction of rst car- tin therapy on C-reactive protein levels: the pravas-
diovascular events. N Engl J Med 2002;347: tatin inammation/CRP evaluation (PRINCE):
155765. a randomized trial and cohort study. JAMA 2001;
[23] Ridker P, Glynn R, Hennekens C. C-reactive protein 286:6470.
adds to the predictive value of total and HDL choles- [39] Pearson TA, Mensah GA, Alexander RW, et al.
terol in determining risk of rst myocardial infarc- Markers of inammation and cardiovascular dis-
tion. Circulation 1998;97:200711. ease application to clinical and public health prac-
[24] Ridker PM. High sensitivity C-reactive protein: tice: a statement for healthcare professionals for the
potential adjunct for global risk assessment in the Center for Disease Control and Prevention and the
primary prevention of cardiovascular disease. Circu- American Heart Association. Circulation 2003;107:
lation 2001;103:18138. 499511.
500 WANG et al

[40] Scheier B, Scheier E, Hilfker-Kleiner D, et al. Ex- established atherosclerosis in mice. Proc Natl Acad
pression of angiotensin II and interleukin-6 in hu- Sci U S A 2000;97:745663.
man coronary atherosclerotic plaques: potential [53] Schonbeck U, Varo N, Libby P, et al. Soluble
implications for inammation and plaque instabil- CD40L and cardiovascular risk in women. Circula-
ity. Circulation 2000;101:13728. tion 2001;103:22668.
[41] Ridker P, Rifai N, Stampfer MJ, et al. Plasma con- [54] Heeschen C, Dimmeler S, Hamm C, et al. Soluble
centration of interleukin-6 and the risk of future CD-40 ligand in acute coronary syndromes.
myocardial infarction among apparently healthy N Engl J Med 2003;348:110411.
men. Circulation 2000;101:176772. [55] Gearing AJ, Newman W. Circulating adhesion mol-
[42] Lindmark E, Diderholm E, Wallentin L, et al. Rela- ecules in disease. Immunol Today 1993;14:50612.
tionship between interleukin-6 and mortality in [56] Ridker P, Hennekens CH, Roitman-Johnson B,
patients with unstable coronary artery disease. et al. Plasma concentration of soluble intracellular
JAMA 2001;286:210713. adhesion molecule 1 and risks of future myocardial
[43] Hakkinen T, Luoma JS, Hiltunen MO, et al. Lipo- infarction in apparently healthy men. Lancet 1998;
protein-associated phospholipase A2, platelet acti- 351:8892.
vating factor acetylhydrolase, is expressed by [57] Ridker P, Buring JE, Rifai N. Soluble P-selectin and
macrophages in human and rabbit atherosclerotic the risk of future cardiovascular events. Circulation
lesions. Arterioscler Thromb Vasc Biol 1999;19: 2001;103:4915.
290917. [58] Hwang SJ, Ballantyne CM, Sharrett R, et al. Circu-
[44] Caslake MJ, Packard CJ, Suckling KE, et al. Lipo- lating adhesion molecules VCAM-1, ICAM-1, and
protein-associated phospholipase A2, platelet acti- E-selectin in carotid atherosclerosis and incident
vating factor acetylhydrolase: a potential new risk coronary heart disease cases: the Atherosclerosis
factor for coronary artery disease. Atherosclerosis Risk in Communities Study. Circulation 1997;96:
2000;50:4139. 421925.
[45] Packard CJ, OReilly DS, Caslake MJ, et al. Lipo- [59] Artl A, Marsche G, Lestavel S, et al. Role of serum
protein-associated phospholipase A2, as an indepen- amyloid A during metabolism of acute phase HDL
dent predictor of coronary heart disease. N Engl J by macrophages. Arterioscler Thromb Vasc Biol
Med 2000;343:114855. 2000;20:76372.
[46] Blake GJ, Dada N, Fox JC, et al. A prospective eval- [60] Johnson BD, Kip KE, Marroquin OC, et al. Serum
uation of lipoprotein-associated phospholipase A2 amyloid A as a predictor of coronary artery disease
levels and the risk of future cardiovascular events and cardiovascular outcome in women. Circulation
in women. J Am Coll Cardiol 2001;38:13026. 2004;109:72632.
[47] Ballantyne CM, Hoogeveen RC, Bang H, et al. Li- [61] Girndt M, Kohler H. Interleukin-10 (IL-10): an up-
poprotein-associated phospholipase A2, high sensi- date on its relevance for cardiovascular risk. Neph-
tivity C-reactive protein, and risk for incident rol Dial Transplant 2003;18:19769.
coronary heart disease in middle aged men and [62] Von der Thusen JH, Kuiper J, Fekkes ML, et al. At-
women in the Atherosclerosis Risk in Communities tenuation of atherogenesis by systemic and local ad-
(ARIC) study. Circulation 2004;109:83742. enovirus-mediated gene transfer of interleukin-10 in
[48] Brilakis ES, McConnell JP, Lennon RJ, et al. Asso- LDLr/ mice. FASEB J 2001;15:27302.
ciation of lipoprotein-associated phospholipase A2 [63] Smith DA, Irving SD, Sheldon J, et al. Serum levels
levels with coronary artery disease risk factors, an- of the anti-inammatory cytokine interleukin-10 are
giographic coronary artery disease, and major ad- decreased in patients with unstable angina. Circula-
verse events at follow-up. Eur Heart J 2005;26: tion 2001;104:7469.
13744. [64] Bar-Or D, Lau E, Winkler JV. A novel assay for co-
[49] Leach CA, Hickey DM, Ife RJ, et al. Lipoprotein- balt-albumin binding and its potential as a marker
associated PLA2 inhibitiona novel, non-lipid low- for myocardial ischemia: A preliminary report.
ering strategy for atherosclerosis therapy. Farmaco J Emerg Med 2000;19:3115.
2001;56:4550. [65] Pollack CV, Sieck S. Use of ischemia-modied albu-
[50] Henn V, Slupsky JR, Grafe M, et al. CD40 ligand on min in emergency department risk stratication of
activated platelets triggers an inammatory reaction chest pain is both clinically eective and cost eec-
of endothelial cells. Nature 1998;351:5914. tive. Ann Emerg Med 2003;(Suppl 42):S37.
[51] Mach F, Schonbeck U, Sukhova GK, et al. Func- [66] Christenson RH, Duh SH, Sanhai WR, et al. Char-
tional CD40 ligand is expressed on human vascular acteristics of an albumin cobalt binding test for as-
endothelial cells, smooth muscle cells, and macro- sessment of acute coronary syndrome patients: A
phages: implications for CD40CD40 ligand signal- multicenter study. Clin Chem 2001;47:46470.
ing in atherosclerosis. Proc Natl Acad Sci U S A [67] Kontos MC, Schorer S, Kirk JD. Ischemia modied
1997;94:19316. albumin, a new biomarker of myocardial ischemia,
[52] Schonbeck U, Sukhova GK, Shimizu K, et al. Inhi- for the emergency diagnosis of acute coronary syn-
bition of CD40 signaling limits evolution of drome. J Am Coll Cardiol 2003;41:340A.
CARDIAC ISCHEMIA & INFLAMMATION MARKERS 501

[68] Panteghini M. Role and importance of biochemical with stable coronary disease. Circulation 2003;108:
markers in clinical cardiology. Eur Heart J 2004; 298792.
25:118796. [79] Marumoto K, Hamada M, Hiwada K. Increased se-
[69] Kleinfeld AM, Prothro D, Brown DL, et al. cretion of atrial and brain natriuretic peptides dur-
Increases in serum unbound free fatty acid levels fol- ing acute myocardial ischemia induced by dynamic
lowing coronary angioplasty. Am J Cardiol 1996;78: exercise in patients with angina pectoris. Clin Sci
13504. 1995;88:5516.
[70] Stein BC, Levin RI. Natriuretic peptides: physiol- [80] Kyriakides ZA, Markianos M, Michalis L, et al.
ogy, therapeutic potential and risk stratication in Brain natriuretic peptide increases acutely and
ischemic heart disease. Am Heart J 1998;135:91423. much more prominently than atrial natriuretic pep-
[71] Omland T, Aakvaag A, Bonarjee VV, et al. Plasma tide during coronary angioplasty. Clin Cardiol
brain natriuretic peptide as an indicator of left 2000;23:285.
ventricular systolic function and long term survival [81] Tateishi J, Masutani M, Ohyanagi M, et al. Tran-
after acute myocardial infarction. Comparison with sient increase in plasma brain (B-type) natriuretic
plasma atrial natriuretic peptide and N-terminal peptide after percutaneous transluminal coronary
proatrial natriuretic peptide. Circulation 1996;93: angioplasty. Clin Cardiol 2000;23:77680.
19639. [82] Goetze JP, Gore A, Moller CH, et al. Acute myocar-
[72] de Lamos JA, Morrow DA, Bentley JH, et al. The dial hypoxia increases BNP gene expression. FASEB
prognostic value of B-type natriuretic peptide in J 2004;18:192830.
acute coronary syndromes. N Engl J Med 2001; [83] DSouza SP, Yellon DM, Martin C, et al. B-type na-
345:101421. triuretic peptide limits infarct size in rat isolated
[73] James SK, Lindahl B, Seigbahn A, et al. N-terminal hearts via KATP channel opening. Am J Physiol
pro-brain natriuretic peptide and other risk markers Heart Circ Physiol 2003;284:H1592600.
for the separate prediction of mortality and subse- [84] Danne O, Mockel M, Lueders C, et al. Prognostic
quent myocardial infarction in patients with unsta- implications of elevated whole blood choline levels
ble coronary artery disease. Circulation 2003;108: in acute coronary syndromes. Am J Cardiol 2003;
27581. 91:10607.
[74] Sadanandan S, Cannon CP, Chekuri K, et al. Asso- [85] Mizuno K, Satomura K, Miyamoto A, et al. Angio-
ciation of elevated B-type natriuretic peptide levels scopic evaluation of coronary artery thrombi in
with angiographic ndings among patients with un- acute coronary syndromes. N Engl J Med 1992;
stable angina and non-ST-segment elevation myo- 326:28791.
cardial infarction. J Am Coll Cardiol 2004;44:5648. [86] Bayes-Genis A, Conover CA, Overgaard MT, et al.
[75] Morrow DA, de Lemos JA, Sabatine MS, et al. Eval- Pregnancy-associated plasma protein A as a marker
uation of B-type natriuretic peptide for risk as- of acute coronary syndromes. N Engl J Med 2001;
sessment in unstable angina/non ST-elevation 345:10229.
myocardial infarction: B-type natriuretic peptide [87] Lund J, Qin QP, Ilva T, et al. Circulating preg-
and prognosis in TACTICS-TIMI 18. J Am Coll nancy-associated plasma protein A predicts out-
Cardiol 2003;41:126472. come in patients with acute coronary syndrome
[76] Jernberg T, Lindahl B, Siegbahn A, et al. N terminal but no troponin I elevation. Circulation 2003;
pro-brain natriuretic peptide in relation to inam- 108:19246.
mation, myocardial necrosis and the eect of inva- [88] Lynch SM, Morrow JD, Roberts LJ, et al. Forma-
sive strategy in unstable coronary syndrome. J Am tion of non-cyclooxygenase-derived prostanoids
Coll Cardiol 2003;42:1909. in plasma of low density lipoprotein exposed to
[77] Talwar S, Squire IB, Downie PF, et al. Plasma N term- oxidative stress in vitro. J Clin Invest 1994;93:
inal pro-brain natriuretic peptide and cardiotrophin 9981004.
1 are raised in unstable angina. Heart 2000;84:4214. [89] Holvoet P, Collen D, van de Werf F. Malondialde-
[78] Bibbins-Domingo K, Ansari M, Schiller NB, et al. hyde-modied LDL as a marker of acute coronary
B-type natriuretic peptide and ischemia in patients syndromes. JAMA 1999;281:171921.
Cardiol Clin 23 (2005) 503516

Exercise Testing in Chest Pain Units: Rationale,


Implementation, and Results
Ezra A. Amsterdam, MD, FACCa,*, J. Douglas Kirk, MD, FACEPb,
Deborah B. Diercks, MDb, William R. Lewis, MD, FACCa,
Samuel D. Turnipseed, MDb
a
Department of Internal Medicine, University of California School of Medicine (Davis) and Medical Center,
4860 Y Street, Suite 2820, Sacramento, CA 95817, USA
b
Department of Emergency Medicine, University of California School of Medicine (Davis) and Medical Center,
4150 V Street, PSSB Suite 2100, Sacramento, CA 95817, USA

Although new diagnostic approaches have patient management, and inecient resource use.
enhanced the evaluation of patients presenting This problem is reected by the recent demonstra-
to the emergency department (ED) with chest tion that an appreciable number of patients with
pain, this syndrome remains a major clinical coronary artery disease (CAD) who present with
challenge [1]. This symptom accounts for more acute chest pain respond to proton pump inhibi-
than 8 million ED visits per year in this country, tors, reecting the gastrointestinal cause of their
accounting for more than 2 million hospital ad- symptoms [7].
missions at a cost of $8 billion for presumed acute The development of chest pain units (CPUs) is
coronary syndrome (ACS) [2]. However, a coro- a response to the need for a strategy to eect accu-
nary event is actually conrmed in only a minority rate, safe, and cost-eective management of pa-
of these patients [3]. This population poses a di- tients presenting with possible ACS. Although
lemma to the clinician of inadvertent discharge their initial purpose was to facilitate rapid coronary
of those with a life-threatening condition versus reperfusion therapy, these units have evolved into
unnecessary admission for a benign process with centers for management of the lower risk popula-
its associated expense and the potential risks of tion that composes the majority of patients present-
further tests. A low threshold for admission of ing with chest pain. The latter include those without
these patients was advocated early in the coronary initial, objective evidence of myocardial ischemia/
care unit (CCU) era by the admonition that pa- infarction in whom accelerated risk stratication
tients should be admitted to the CCU solely on can identify those requiring admission and those
suspicion of having a myocardial infarction [4]. who can be safely discharged with outpatient
This approach has persisted because of the focus follow-up [815]. A basic element of this accelerated
on patient welfare as well as the litigation poten- diagnostic protocol (ADP) is stress testing after
tial of missed ACS [5]. Inadvertent discharge of a negative initial assessment for myocardial infarc-
patients with ACS persists at a rate of 4% to tion (MI) or unstable angina. The primary method
5% and the mortality and morbidity of this group of testing has been treadmill exercise electrocardiog-
are substantial [6]. However, a consequence of the raphy (ECG) in the context of the ADP. Fundamen-
low threshold for admission has been large num- tal to this approach is the identication of patients
bers of unnecessary hospitalizations, suboptimal with low clinical risk on presentation to the ED.

Indicators of low clinical risk


* Corresponding author.
E-mail address: eaamsterdam@ucdavis.edu There is abundant evidence that low risk in
(E.A. Amsterdam). patients presenting with chest pain can be
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.016 cardiology.theclinics.com
504 AMSTERDAM et al

recognized on presentation and that this group aords: (1) early identication of clinical risk
neither requires nor benets from traditional and (2) further risk stratication of low-risk
intensive care or extended observation in a moni- patients to identify those who require admission
toring unit. Lee and colleagues [16] reported that and those who can be discharged [815,21]. CPUs
in patients admitted to rule out a coronary event, vary in form and may either occupy a designated
those with ! 5% probability of acute MI could structural area or function as virtual units com-
be identied by type of chest pain, past history, prising primarily personnel and process. Close
and initial ECG. Extension of this approach to coordination between ED physicians and cardiol-
over 4600 such patients demonstrated that the ini- ogists is an essential element for successful func-
tial clinical assessment could distinguish those tioning of the unit. The strategy is based on
with ! 1% probability of major complications a protocol-driven process that uses current stand-
[17]. The prognostic utility of the initial ECG ards of care for ecient and timely treatment in
alone in patients admitted to rule out MI was conformity with the guidelines of the American
demonstrated by Brush and colleagues [18] in College of Cardiology (ACC) and American
their report that a negative ECG on admission Heart Association (AHA) [25] as depicted in
was associated with a 0.6% rate of serious compli- Fig. 1.
cations during hospitalization compared with
a 14% incidence in those with an abnormal
ECG [18]. An earlier study indicated that a normal Accelerated diagnostic protocols
ECG in patients admitted for preinfarction angina ADPs have been increasingly used in low-risk
predicted benign early and late outcomes in con- patients, usually culminating in one of the meth-
trast to ECG evidence of ischemia, which corre- ods of cardiac stress evaluation, depending on the
lated with markedly increased cardiac morbidity results of the clinical assessment. This process
and mortality [19]. These ndings were con- usually entails 6 to12 hours of clinical observa-
rmed by Schroeder and colleagues [20] in their tion, serial 12-lead ECGs, continuous ECG mon-
report that in patients in whom MI was ruled itoring, and measurement of serial serum markers
out, ECG evidence of ischemia was associated of cardiac injury [11,13,15,21,25]. Positive ndings
with a 1-year mortality similar to that of post- indicate ACS (usually non-ST elevation ACS,
MI patients. An important concept to emerge rarely ST elevation ACS) and mandate admission
from these studies is that although the cause for further management. Negative ndings are
of chest pain is frequently elusive, basic clinical consistent with absence of MI and no evidence
tools provide powerful estimates of cardiac risk. of ischemia at rest. In these cases, the evaluation
Recognition of low clinical risk stimulated proceeds to a stress test to determine if the patient
alternative approaches to conventional coronary has inducible ischemia. Those with a positive test
care, such as reduced time in the CCU [21,22], direct are admitted and those with a negative result are
admission to a step-down unit [23], and observation discharged to outpatient follow-up. Multiple
in a short stay unit [24]. Recent innovations in the methods are currently available to detect stress-
management of low-risk patients include guide- induced ischemia, the most widely available and
lines, critical pathways, new serum markers of car- readily applicable of which is treadmill exercise
diac injury, novel ECG monitoring systems, early testing. The utility of ADPs has been well demon-
noninvasive cardiac imaging, early treadmill exer- strated with this method, as indicated by its safety
cise testing, coronary calcium screening in the and the very low clinical risk in patients desig-
ED, noninvasive coronary angiography by CT and nated as appropriate for early discharge.
conventional coronary angiography [815,21].

Development and evolution of early exercise


testng
Chest pain units
Initial recommendations
Although risk is low in patients selected for
admission to a CPU, it is not negligible. Contem- Incorporation of treadmill exercise testing into
porary CPUs (also known as chest pain observa- current ADPs is a recent development based on
tion units, chest pain emergency units, and chest ample data that has overcome initial concern
pain evaluation units) provide an integrated regarding the possible hazards of this technique
approach to the patient with chest pain that in potentially unstable patients. This early,
EXERCISE TESTING IN CHEST PAIN UNITS
Fig. 1. Management of patients presenting with symptoms suggestive of an acute coronary syndrome (ACS). ACC, American College of Cardiology; AHA, American Heart
Association. (Adapted from Braunwald E, Antman E, Beasley JW, et al. ACC/AHA guidelines for the management of patients with unstable angina and non-ST elevation

505
myocardial infarction. http://www.acc.org/clincal/guidelines/unstable/unstable.pdf.)
506 AMSTERDAM et al

cautionary approach and its evolution into cur- contemporary studies conrmed the safety of
rent concepts is reected in the progressive alter- symptom-limited treadmill exercise ECG testing
ation of recommendations by expert panels that after 8 to 12 hours of evaluation in patients who
have considered exercise testing in patients pre- have been identied as being at low to intermedi-
senting with possible ACS. Such caution is re- ate risk by a clinical algorithm that uses serum
ected in the Clinical Practice Guideline on markers of myocardial necrosis and resting
Unstable Angina published in 1994 [26]. Noting ECGs [28]. This strategy is incorporated in the
that In many cases, noninvasive stress testing 2002 guidelines of the ACC/AHA for manage-
provides a useful supplement to clinically based ment of patients with non-ST elevation ACS in
risk assessment, this guideline recommended which it is recommended that exercise testing
that unless cardiac catheterization is indicated, can be performed in stable, low risk patients if
stress testing should be should be performed in a follow-up 12-lead ECG and cardiac marker
patients hospitalized with unstable angina who measurements after 6 to 8 hours of observation
have been free of angina and congestive heart fail- are normal [25]. Recent exercise testing guide-
ure for a minimum of 48 hours. The strength of lines are in accord with these recommendations
the evidence for this recommendation was rated [29,30]. These updated guidelines reect evolution
B (B evidence from well conducted clinical from earlier versions that advised exercise testing
studies but no randomized clinical trials). The re- only after patients had been symptom-free for
port of the American College of Emergency Physi- a minimum of 48 hours [31]. Thus, during the
cians in 1995 emphasized the ecacy of CPUs in course of less than a decade, recommendations
reducing both missed ACS and unnecessary hos- for exercise testing in low-risk patients progressed
pitalizations [27]. However, its only reference to from admonitions for a pretest 48-hour period of
stress testing as part of the patient evaluation clinical stability to expert consensus supporting its
was that Many patients, upon discharge from application after a much briefer interval.
the CPU, are scheduled for a form of testing
that is reliable for identifying ischemic heart dis-
Initial studies of early exercise testing
ease. A year later, the Working Group Report
of the National Heart Attack Alert Program pub- All studies of early exercise testing in patients
lished an extensive evaluation of technologies to presenting with chest pain have required that
detect ischemia in the ED [10]. This body con- patients are clinically stable with no ECG evi-
cluded that the expedited ECG exercise test may dence of ischemia/injury. The criteria for a positive
oer the benet of an expedited workup and may test for myocardial ischemia are the standard
reduce hospital admissions for chest pain. How- indicators: R 1.0 mm horizontal or downsloping
ever, ECG exercise stress testing in the ED cannot ST segment shift. Other exercise-induced altera-
be recommended in the absence of additional tions that indicate an abnormal test and the need
data demonstrating safety and eectiveness. The for further evaluation include angina, arrhyth-
quality of the evidence at the time of publication mias, and fall in blood pressure. Although the rst
was rated C, indicating its basis in limited stud- two studies of this method included only small
ies, and the test was viewed as modestly accurate numbers of patients, they demonstrated the safety
with little or unknown clinical impact. of exercise testing in low-risk patients in the ED
setting, and the utility of this method was con-
rmed in multiple subsequent investigations
Current guidelines
(Table 1). There have been no reports of adverse
Subsequent clinical studies demonstrating the events in any study of early exercise testing of
safety and ecacy of exercise testing in ADPs low-risk patients.
have provided rm support for this approach. The Tsakonis and colleagues [32] evaluated 28 pa-
31st Bethesda Conference on Emergency Cardiac tients several hours after hospital arrival with
Care held in 1999 noted that ADPs, including ex- treadmill ECG using a symptom-limited modied
ercise testing as a key element, have been associ- Bruce protocol (see Table 1). These patients had
ated with reduced hospital stay and lower costs unexplained chest pain consistent withdbut not
[11]. The absence of adverse eects and the accu- diagnostic ofdangina and normal baseline
rate identication of low post-discharge prognos- ECGs. The exercise test was negative in 23 pa-
tic risk were also recognized. A subsequent tients and positive in 5. The latter group was ad-
Science Advisory of the AHA concluded that mitted and the former was discharged. The
EXERCISE TESTING IN CHEST PAIN UNITS 507

Table 1
Studies of exercise electrocardiography (ECG) testing in chest pain centersa
Percent Percent Adverse
Number Percent negative positive exercise
of positive predictive predictive test
Author patients testsb valuec valuec events
Tsakonis et al [32] 28 17.8 100 0
Kerns et al [33] 32 0 100 0
Lewis and Amsterdam [34] 93 13.0 100 46 0
Gibler et al [35] 782 1.2 99 44 0
Gomez et al [37] 100 7 100 0 0
Zalenski et al [39] 224 8 98 16 0
Polanczyk et al [40] 276 24 98 15 0
Kirk et al [44] 212 12.5 100 57 0
Amsterdam et al [45] 1000d 13 88.7 33 0
a
Includes studies in which results of exercise ECG tests could be distinguished from those of other forms of stress
testing.
b
Positive exercise ECG.
c
Based on clinical follow-up or further cardiac evaluation.
d
Includes a small number of patients in Kirk and colleagues [44].

investigators reported no cardiac morbidity or than 1 hour from the decision to admit and in
mortality at 6 months in the patients with negative less than 24 hours in all patients. Positive tests oc-
exercise tests. These ndings were conrmed in the curred in 13% of patients, negative in 64%, and
subsequent report of Kerns and colleagues [33] nondiagnostic (no ischemia but peak heart rate
who performed exercise testing in 32 ED patients ! 85% of age-predicted maximum) in 13%. Is-
with atypical chest pain, normal ECGs, and car- chemic ECG changes occurred at a signicantly
diac risk factor stratication (see Table 1). Com- lower percentage of age-predicted maximal heart
pared with similar patients admitted for rate (70%) in patients with true-positive tests com-
evaluation of atypical chest pain, those with neg- pared with those with false-positives (O90%)
ative results during ED exercise testing had (Fig. 2). No complications were associated with
shorter length of stay (5.5 hours versus 2 days) exercise testing. Coronary angiography revealed
and lower costs ($467 versus $2340). All patients signicant CAD in 6 of the 13 patients with posi-
in the accelerated protocol had negative exercise tive tests, 5 of whom had multivessel involvement.
tests and no evidence of CAD at 6 months fol- A majority (54%) of the 81 patients with negative
low-up. In addition to their small numbers, limita- or nondiagnostic results was discharged immedi-
tions of these two studies include the very low risk ately after the exercise test. At 6 months follow-
of the patients (none had been designated for ad- up, there were no coronary events in patients
mission), the lack of any positive tests in Kerns with negative or nondiagnostic exercise tests. Sev-
patients, and no further evaluation for CAD in eral unique aspects of this study demonstrated the
Tskanosiss patients with positive tests. utility of early exercise testing in low-risk patients
The rst study of exercise testing in patients and provided the basis for the authors current ap-
with the clinical prole of those currently included proach of immediate exercise testing in low-risk
in CPU ADPs was published from the authors patients without excluding MI by a traditional se-
institution in 1994. In this investigational pro- ries of negative cardiac serum markers. In con-
tocol, selected patients presenting with chest pain trast to the preceding studies, it included only
who were designated for admission by ED physi- patients who were assigned to admission for a tra-
cians to rule out ACS underwent immediate ditional rule out MI protocol. In these patients,
treadmill testing (see Table 1) [34] . The study exercise testing was performed before the latter
group comprised 93 patients in whom ED symp- process, there were no adverse eects of testing,
tom-limited exercise testing was performed by true and false positive tests were related to the
a cardiologist using a modied Bruce protocol heart rate at ST-segment depression, the majority
without prior measurement of any markers. The of patients were discharged immediately after
test was performed within a median time of less a negative or nondiagnostic test, and there were
508 AMSTERDAM et al

100 *3 3
59 33 26 6

12 6 6

% Maximal Predicted Heart Rate


80
22 9 13 6 3
3

60

40

20

0
Negative Non-Diagnostic Positive True Positive False Positive

Fig. 2. Treadmill exercise results. Percent maximal predicted heart rate attained versus exercise electrocardiographic re-
sult for all patients (solid bars), men (open bars), women (shaded bars). Numbers above bars represent number of pa-
tients. *Percent maximal predicted heart rate for patients with true-positive tests versus false-positive tests, P ! .01.

no coronary events during the posthospital evaluation in only small numbers of patients (see
course. One patient with a positive test was found Table 1). The utility of this strategy was conrmed
to have non-ST elevation MI by subsequent serial in its ability to safely and eciently reduce unnec-
serum enzymes. Coronary angiography revealed essary admissions in low-risk patients. Indeed, the
right coronary artery disease, coronary angio- relative frequency of patients with true- and false-
plasty was performed, and the clinical course positive exercise tests in this setting is actually
was uncomplicated. similar to that seen in asymptomatic individuals
[36], conrming the value of clinical indicators
in identifying the probability of disease in pa-
tients presenting to the ED with chest pain
Exercise testing in chest pain units
[12,16,17,21,34,35].
Gibler and associates [34] demonstrated the The Rapid Rule-Out of Myocardial Ischemia
utility of early exercise testing in 782 patients in Observation (ROMIO) study of Gomez and co-
whom low risk was by an ADP (see Table 1) workers [37] was the rst prospective, controlled
[35]. Their protocol included serial ECGs and se- investigation of ADP-exercise testing (see Table
rum CK-MB, 9 hours of continuous ST-segment 1). It evaluated 100 patients with chest pain, half
monitoring, and a resting echocardiogram fol- of whom were randomly assigned to admission
lowed by symptom-limited exercise testing in for regular care and half to a chest pain center
those with negative ndings. There were no ad- protocol consisting of a 12-hour observation
verse eects and no mortality at 30-day follow-up period with standard ECGs, continuous ST-seg-
in the patients with negative tests. The negative ment monitoring, and serial CK-MBs. Observa-
predictive value of the exercise test was 99%. Al- tion and monitoring were negative in 44 patients
though the positive predictive value was less than in the accelerated protocol in whom symptom-
50%, only 9 of 782 patients (!2%) had positive limited exercise testing was performed without ad-
exercise tests, which was lower than in the authors verse eects, demonstrating normal tests in 93%
initial investigation [34]. and positive results in 7%. The latter were all
Emerging from these studies [3235] is recogni- false-positives based on coronary angiography.
tion of the safety and excellent predictive value of All patients with negative tests were discharged
negative exercise tests in patients identied as low after the exercise evaluation, and there were no
risk (see Table 1). Further, a large majority of pa- coronary events at 30-day follow-up. However,
tients selected for exercise testing by an ADP had compared with regular care, the observation proto-
negative ndings and, although the positive pre- col was associated with substantial reductions in
dictive value was modest, positive tests were in- length of stay (11.0 versus 22.8 hours) and total
frequent and resulted in the need for further cost ($624 per patient at 30 days). This was the rst
EXERCISE TESTING IN CHEST PAIN UNITS 509

study to document the anticipated cost savings by CK-MB data and serial ECGs, the exercise test
an observation unit protocol. had a high sensitivity (90%) and excellent negative
By the latter part of the last decade, the ecacy predictive value (98%); as in prior studies, speci-
of ADP-exercise test protocols was evident for city and positive predictive value were low
low-risk patients. However, although the potential (51% and 16%, respectively). Cost analysis dem-
for cost savings was suggested, this advantage had onstrated a savings of $567 per patient managed
not been rmly established. Additionally, there by the ADP-exercise test strategy.
was also the possibility, as suggested by Roberts The report of Polanczyk and colleagues [40]
and co-workers [38], that this strategy could actu- evaluated the prognostic signicance of the exer-
ally lead to higher costs by capturing very low-risk cise test performed within 48 hours of presenta-
patients who would previously have been dis- tion in 276 low-risk patients admitted for chest
charged directly from the ED. These investigators pain (see Table 1). Twenty-six percent of these pa-
addressed this issue in a prospective, randomized tients had a prior history of CAD. The test was
trial in which the primary outcomes were length performed within 12 hours in 7% of patients, at
of stay and total cost [38]. The study group com- 12 to 24 hours in 45%, and after 24 hours in
prised 165 low-risk chest pain patients enrolled in 48%. The Bruce treadmill protocol was used in
a 12-hour ADP followed by an exercise test in ap- 84% of the patients, and the modied version
propriate subjects. There were no deaths or seri- was used in 12%. A negative test was dened by
ous complications associated with the ADP, achievement of at least 3 METs (one stage of
which also resulted in fewer indeterminate evalua- the Bruce protocol) without evidence of ischemia.
tions than standard care (16% versus 54%, P ! Positive tests were those in which the results were
.001). Length of stay and cost were signicantly interpreted as highly predictive of signicant cor-
less for ADP patients than those receiving stan- onary disease or strongly positive and inconclu-
dard care (33 versus 45 hours, P ! .01; and sive tests were those consistent with but not
$1528 versus $2095, P ! .001) (Table 2). Extrap- diagnostic of ischemia or without evidence of is-
olation of this cost benet by the authors indi- chemia at a peak work of ! 3 METs. There were
cated a potential annual national saving of over no adverse eects of exercise testing. The test was
$238 million. negative in 71% of patients, positive in 24%, and
In a prospective, cross-sectional study, Zalen- inconclusive in 5%. Outcome data were available
ski and colleagues [39] evaluated 317 patients by at 6 months for 92% of the study group. Events
a 12-hour ADP followed by exercise testing (see during the follow-up period were dened as car-
Table 1). A unique aspect of this trial was admis- diac death, MI, or myocardial revascularization.
sion of all patients to obtain a diagnosis to which During this interval, there was no mortality and
the ADP evaluation could be compared. Exercise the event rate in the negative exercise test cohort
testing was performed in 224 patients with nega- was 2% compared with 15% in those with posi-
tive observation data and was negative in 66%, tive or equivocal tests. In the negative test group,
positive in 8%, and inconclusive in 26% because compared with those with positive/equivocal tests,
of failure of patients to reach 85% of age-pre- there were also fewer repeat ED visits (17% versus
dicted maximal heart rate. There were no adverse 21%, P ! .05) and fewer readmissions (12% ver-
eects of exercise testing. There were no follow-up sus 17%, P ! .01). The negative predictive value
data in this study, in which the accuracy of the of the exercise test was 98%; sensitivity and spec-
ADP-exercise strategy was based on diagnoses ob- icity were 73% and 74%, respectively. In addi-
tained during admission. In combination with the tion to the documented prognostic utility of the
exercise test, this study aorded other noteworthy
features. The investigators demonstrated the
Table 2
safety of the early exercise test in patients with a
Primary outcomes
history of CAD and their denition of test results
ADP Control yielded a low rate of inconclusive diagnoses, in-
Outcome (n 82) (n 83) P creasing the clinical utility of the test in clinical
Hospitalized, % 45.1 100 !.001 decision-making.
Total cost, mean 1528 2095 !.001 Mandatory stress testing in a chest pain center
LOS, mean, h 33.1 44.8 !.01 was evaluated by Mikhail and co-workers [41] for
Abbreviations: ADP, accelerated diagnostic protocol; its safety, utility, and cost-eectiveness. After neg-
CI, condence interval; LOS, length of stay. ative ndings for myocardial ischemia/infarction,
510 AMSTERDAM et al

a total of 424 patients underwent cardiac stress care during the follow-up period was signicantly
evaluation, which included 247 exercise treadmill higher in the hospital admission patients (P
tests. The remainder of the tests included stress .003), amounting to an estimated 61% increase
imaging studies; no data are given on the results in costs.
of the separate test modalities. The tests were neg- The impact of CPUs compared with regular
ative in 392 (92.6%) patients, and average stay in hospital care was analyzed by Gra and col-
the CPU was 12.8 hours. At 5 months follow-up, leagues [27] in the report of the Chest Pain Evalu-
there was no mortality or MI in any of the pa- ation Registry (CHEPER) Study. Results were
tients discharged after a negative evaluation. A - assessed in terms of the proportion of patients un-
nal diagnosis of ischemic heart disease was made dergoing a complete rule out MI investigation,
in 44 patients admitted from the chest pain center, the number of missed MIs, and costs. Outcomes
24 (55%) of whom were identied only on stress in over 23,000 patients managed in eight chest
testing with which there were no adverse eects. pain observation units were compared with results
Evaluation in the CPU with mandatory stress test- in over 12,000 patients in ve studies from hospi-
ing was associated with a cost-per-case saving of tals without these units. Although data on exercise
62% for each patient who would otherwise have testing were not provided in the report, the inves-
been admitted to the inpatient service. tigators note that all CPUs complied with stand-
The largest and most recent prospective, ran- ards of the American College of Emergency
domized trial comparing management by an Physicians [8,43] and the Clinical Practice Guide-
ADP-stress test with regular inpatient care is lines for Management of Unstable Angina [26],
that of Farkouh and co-investigators [42]. They which include recommendations on exercise test-
studied 424 patients with a diagnosis of unstable ing in their management algorithms. CPUs were
angina based on symptoms and considered to be associated with an increase in evaluations to rule
at intermediate risk. Patients with ECG evidence out MI (67% versus 57%, P ! .001), a lower
of ischemia were excluded. The observation pro- rate of missed MIs (0.4% versus 4.5%, P !
tocol, which included ECGs, ST-segment moni- .001) and a lower admission rate (47% versus
toring, and serial serum creatine kinaseMB 57%, P ! .001). The latter eect equated to a po-
measurements, diered from prior studies in its tential for 2314 admissions avoided and more
duration of only 6 hours. Patients with negative than $4 million saved.
ADP ndings who were suciently ambulatory The foregoing studies of the last decade reect
underwent exercise testing while pharmacologic the parallel and interdependent development of
stress imaging was performed in the others. No chest pain centers, ADPs, and early exercise
adverse events occurred with the stress tests but testing in low-risk patients presenting to the ED
the proportion of patients receiving the various with chest pain. Integration of these methods into
tests is not specied. Patients with negative tests a protocol-driven process has rmly established
were discharged and those with positive or equiv- the contemporary chest pain center and its capac-
ocal results were admitted. Forty-six percent of ity for safe, accurate, and rapid patient manage-
patients in the ADP-stress test group had a nega- ment. After a negative observation period of 12
tive overall evaluation and were directly dis- hours or less to exclude MI and ischemia at rest,
charged after a median stay of 9.2 hours. At 6 exercise testing has been feasible to detect or
months, there was no signicant dierence in car- exclude inducible ischemia. The negative predic-
diac events in the ADP-stress test group versus the tive accuracy of this strategy is very high (w98%)
regular care patients (6.6% versus 8.5%, respec- and the low positive predictive accuracy is not
tively). Events were broadly dened (primary: problematic because the frequency of positive
death, MI, heart failure, stroke, cardiac arrest; exercise tests is low, resulting in a small number
secondary: any revisit to the ED or hospitalization of patients requiring admission for further in-
for cardiac diagnosis or care). There were no pri- patient evaluation. Additionally, the exercise test
mary events in the ADP-stress patients with neg- avoids inappropriate discharge of patients not
ative evaluations and early discharge. The identied by the observation process. Estimates
investigators emphasized that this group repre- of cost-eectiveness indicate the potential for
sented 46% of patients who ordinarily would substantial savings by chest pain centers through
have been admitted but in whom the accelerated reduction of unnecessary admissions and decrease
protocol avoided admission. Finally, the use of of inadvertent discharges of ACS patients.
cardiac procedures and hospitalization for cardiac The chest pain center is a relatively recent
EXERCISE TESTING IN CHEST PAIN UNITS 511

development, and investigation is ongoing to


determine optimal implementation of this con- Box 1. Selection criteria for immediate
cept, which will vary with the resources of in- exercise test
dividual institutions and the expertise and
Chest pain suspicious for myocardial
experience of the responsible physicians. In this
ischemia
regard, as previously noted [1215,34], the authors
Able to exercise
have used a unique approach to the management
ECG normal, minor ST-T changes, or no
of low-risk patients at their institution.
change from previous abnormal ECG
Hemodynamically stable, no arrhythmia
A single negative serum marker is
Immediate exercise testing measured in selected patients
The authors evaluation of low-risk patients in
the University of California (Davis) CPU diers
from that of the previous studies in that they demonstrated CAD in 57% of those with a pos-
apply exercise testing immediately after identi- itive test and 30-day follow-up revealed no mor-
cation of low-risk patients on presentation. tality or morbidity in the negative or
Serial cardiac injury markers are not obtained in nondiagnostic groups. This study demonstrated
this selected group, as described in their rst study the safety of proceeding to exercise testing in
10 years ago [34]. Since then, more than 3000 pa- carefully selected patients on the basis of the ini-
tients have been assessed by this strategy with no tial presentation without serial ECGs or cardiac
adverse eects. The current approach embodies injury markers. However, its limited numbers re-
several modications based on their continuing quired a larger patient population for conrma-
experience, as documented in their second study tion of the feasibility of this strategy.
(see Table 1) [44]. In this investigation of 212 pa- In the largest single center study of exercise
tients, those with a history of CAD were not testing in low-risk patients, the authors reported
excluded, exercise testing was performed by in- results in 1000 patients (see Table 1) [45]. This
ternists trained in this technique, who serve as at- study incorporated their current approach, which
tending physicians on the CPU, and cardiac injury includes conrmation of a single negative cardiac
markers were not obtained before testing. Strict
selection criteria for exercise testing included, as
before, absence of hemodynamic dysfunction or Box 2. Immediate exercise test
cardiac arrhythmias, normal or near-normal procedure and end-points
ECGs, and no evidence of a noncardiac cause
of chest pain on screening examination (Box 1). Modified Bruce treadmill protocola
Symptom-limited treadmill exercise was per- Symptom-limited
formed by a modied Bruce protocol with the Other end-points
following endpoints: ischemic ST-segment shift Ischemia (R1.0 mm ST segment shift
(1.00 mm horizontal or downsloping depression for 80 msec after the J point)
60 to 80 msec after the J point), symptoms, or YBlood pressure (R10 mm Hg systolic)
arrhythmias, any of which is an indication for Significant arrhythmia (sustained
immediate termination of the test (Box 2). Neg- supraventricular tachyarrhythmia;
ative exercise tests were obtained in 59% of pa- high-grade ventricular ectopy
tients, positive in 13%, and nondiagnostic in [R2 consecutive beats, sustained
28% (negative exercise ECG but failure to reach bigeminy])
R 85% of age-predicted maximum heart rate). Positive result: R 1.0 mm horizontal ST
Patients with positive tests were admitted and segment shift
all patients with negative exercise tests and Nondiagnostic result: < 85% maximum
93% with nondiagnostic results were discharged predicted heart rate with no ST shift
directly from the ED. In the latter group, the
decision to discharge was based on achievement a
Includes two initial 3-minute states (1.7
of adequate functional capacity (eg, R 2 stages mph, 0% grade and 1.7 mph, 5% grade) before
of the Bruce protocol or R 75% of maximum the standard Bruce protocol.
predicted heart rate). Further evaluation
512 AMSTERDAM et al

injury marker before exercise testing. Almost two- method are prohibitive for many institutions. In
thirds of the patients had negative immediate ex- this regard, the authors have shown that O 70%
ercise tests, approximately 13% were positive of low-risk patients presenting with chest pain
and less than 25% were nondiagnostic (Fig. 3). qualify for immediate exercise testing and that
There was no mortality during the 30-day fol- more complex and expensive stress imaging tech-
low-up interval. The negative predictive value niques can be appropriately reserved for the
for a cardiac event was 99.7% at 30 days. In the remainder of patients [49]. Moreover, a compara-
nondiagnostic group, all events were accounted tive study of 239 low risk patients by Senaratne
for by revascularization in 32% of the 79 patients and colleagues [50] revealed that early treadmill
who had further evaluation. The positive predic- testing was as informative and more cost-eective
tive value of the exercise test was 33% for a car- than scintigraphy in identifying low-risk patients
diac event (four non-Q MIs detected after who did not require hospitalization. In this study,
admission, 12 myocardial revascularizations) or in which the follow-up period was 20 months, ex-
a conrmatory imaging test for CAD (two pa- ercise testing was applicable in all but 9.6% of pa-
tients). These results extended the authors pre- tients. These investigators report that, compared
vious ndings to a large, heterogeneous with scintigraphy as the initial cardiac study, exer-
population in that testing was uncomplicated, cise testing yielded a savings of more than $86,000
the bulk of patients had negative tests and could in this group of patients. In contrast to our strat-
be released directly from the ED, negative predic- egy in which noncardiologists assess patients and
tive value was excellent, and the low positive pre- perform immediate exercise testing, with cardiol-
dictive value involved a relatively small group ogy consultation available as required, Senaratne
with positive exercise tests. The safety and accu- and colleagues [50] emphasize the essential role
racy of this method has been conrmed in patients of cardiologists in these processes. In this regard,
with known CAD by a specic study of immediate the expertise of the authors noncardiology CPU
exercise testing of 100 consecutive patients with physicians in performing exercise testing has
this disease [46]. Although exercise testing is con- been conrmed by their accuracy in test interpre-
sidered to have limited value in women, the au- tation and the complete absence of complications
thors experience has conrmed the reliability of [51]. Analysis of 645 immediate exercise ECGs re-
a negative immediate exercise test in this group. vealed a concordance of greater than 98% be-
The negative predictive value of the test was tween the interpretations of these physicians and
99% in 661 women with a mean age of 54 years the authors sta cardiologists.
studied in the CPU [47].
Immediate exercise testing in special groups
Comparison of exercise testing and myocardial The authors have recently explored a number
scintigraphy of other issues presented by early exercise testing
The utility of myocardial scintigraphy has been in patients presenting with chest pain. The use of
well established in patients presenting with chest beta-blockers or rate-limiting calcium channel-
pain [48]. However, the cost and logistics of this blockers often precludes diagnostic exercise test-
ing because of attenuation of exertional heart rate
by these agents. These drugs are associated with
80 reduction of peak exercise heart rate and rate-
pressure product, resulting in an increased fre-
Percent of Patients

60 n=640 quency of nondiagnostic tests compared with


patients not receiving these agents [52]. However,
40 a majority (O60%) of patients taking these med-
ications did have a diagnostic test. Therefore, it is
20 n=235 our experience that use of these drugs should not
n=125 preclude early exercise testing.
0 Recent trials of non-STE ACS have demon-
Negative Non-Diagnostic Positive
strated the prognostic importance of multiple clin-
Fig. 3. Proportion of 1000 patients with negative, posi- ical factors in patients presenting with this
tive, and nondiagnostic immediate exercise tests. Num- diagnosis, including elevated cardiac injury
bers in bars indicate number of patients. markers, ST-segment deviation, age R 65 years,
EXERCISE TESTING IN CHEST PAIN UNITS 513

R three coronary risk factors, known CAD, warrant comment. As previously noted, patients
R two anginal episodes in the previous 24 hours in whom this method is used are carefully
and aspirin use in the previous 7 days. These fac- screened to conrm their low-risk status, as out-
tors comprise the TIMI risk score for prediction lined in Box 1. The test is terminated at the initial
of fatal and nonfatal coronary events, which in- appearance of any abnormality (see Box 2). Al-
crease directly with the number of these risk fac- though noncardiologists perform the clinical as-
tors [53]. However, in patients presenting to sessment and exercise tests in CPU patients,
their CPU, the authors have found that, with these specially trained physicians have ready
the exception of elevated cardiac injury markers availability of consultation by sta cardiologists.
and ST-segment deviation, both of which preclude In regard to the relatively large proportion
exercise testing, the test can be safely performed (O20%) of patients with nondiagnostic tests (neg-
for reliable risk stratication regardless of the ative but peak heart rate ! 85% of predicted
presence of the other TIMI risk factors [54]. Be- maximum), the authors have found that those
cause it has recently been reported that augment- with negative tests at R 80% of maximum pre-
ing the standard exercise ECG with additional dicted heart rate had uneventful outcomes on fol-
leads enhances its sensitivity for detecting ische- low-up [45]. Therefore, use of this lower heart rate
mia [55], the authors evaluated this innovation for a diagnostic test for the purpose of risk strat-
in their CPU patients. In this setting, the addition ication appears prudent and would reduce the
of four leads (two posterior and two right-sided) nondiagnostic group by 25% [45]. Although fol-
to the standard 12-lead ECG enhanced the sensi- low-up is only 30 days, the purpose of this ap-
tivity of the test minimally without altering speci- proach is to determine short-term risk. This
city [56]: sensitivity rose only minimally from strategy is predicated on timely follow-up and fur-
7.6% to 8.0% based on detection of two pa- ther outpatient evaluation. Finally, although clin-
tients who were positive only in the additional ical assessment is basically reliable in identifying
leads compared with 37 patients with positive low-risk patients, it is imperfect and can result in
ndings in the standard exercise ECG. These ad- inadvertent exercise testing in patients with ACS.
ditional leads were also not useful in detecting This possibility is minimized by physician exper-
ischemia or injury in the resting ECG in patients tise and experience together with continued cau-
admitted to the CPU [57]. In a study of 2021 tion in the selection of patients for testing and in
patients referred for elective outpatient treadmill the indications for test termination.
testing, the authors also found that the 16-lead A recurrent question concerns the necessity of
exercise ECG did not aord increased sensitivity performing exercise testing before discharge after
in this setting [58], thereby failing to conrm the a negative ADP rather than a short time after
prior study of Michaelides and colleagues [55]. discharge. The former approach provides the
A notable aspect of the authors CPU experi- most ecient completion of the evaluation and
ence has been the high recidivism rate of patients obviates concern regarding lack of return of
discharged from the unit with a negative evalua- patients for the outpatient test, which would
tion. During a 7.5-year period, 13% of 1960 contribute to the hazard of incomplete assessment
patients had R two or more negative immediate and missed ACS [6]. However, where testing is not
exercise tests and accounted for 26% of the CPU feasible and the system and patient characteristics
visits [59]. Further, of the latter group, almost are conducive to early return for testing (within 48
10% had R four negative exercise tests during hours), it is reasonable to consider this approach.
this period. The multi-exercise test patients were The authors have discharged selected very low-
relatively young (mean age 52 years) and most of risk patients from their CPU without a predis-
them were women. Beyond the traditional dier- charge exercise test. They have specically studied
ential diagnosis of multiple somatic causes of chest this approach in a group of very low-risk women
pain [60], common and underdiagnosed condi- presenting to the ED with chest pain who were les
tions responsible for this symptom include anxiety than 50 years old, nondiabetic, and nonsmokers
syndromes and somatoform disorders [11,6062]. [63]. Of the entire group of 346 women, 175
were discharged from the CPU without exercise
testing. At 30 days follow-up, none of these pa-
Further issues
tients had conrmatory evidence of CAD or
Several aspects of the authors application of ACS. The results suggest that the risk of ACS is
immediate exercise testing in CPU patients very minimal in women with low-risk proles
514 AMSTERDAM et al

who present with chest pain and that stress testing [7] Liuzzo JP, Ambrose JA, Diggs P. Proton pump
in the CPU may not be necessary to determine dis- inhibitors for patients with coronary artery disease
position in these patients. These ndings have im- with reduced chest pain, emergency department vis-
plications for optimal use of limited resources. its, and hospitalizations. Clin Cardiol 2005;28:
36974.
[8] Gra L, Joseph T, Andelman R, et al. American
Summary College of Emergency Physicians Information Pa-
per: chest pain units in emergency departmentsd
CPUs are now established centers for assess- a report from the short-term observation services
ment of low-risk patients presenting to the ED section. Am J Cardiol 1995;76(14):10369.
with symptoms suggestive of ACS. ADPs, of [9] Jesse RL, Kontos MC, Roberts CS. Evaluation of
which treadmill testing is a key component, have chest pain in the emergency department. Curr Prob
Cardiol 1997;22:151236.
been developed within these units for ecient
[10] Selker HP, Zalenski RJ, Antman EM, et al. An eval-
evaluation of these patients. Studies of the last
uation of technologies for identifying acute cardiac
decade have established the utility of early exer- ischemia in the emergency department. Executive
cise testing, which has been safe, accurate, and summary of a National Heart Attack Alert Program
cost-eective in this setting. Specic diagnostic Working Group report. Ann Emerg Med 1997;29:
protocols vary, but most require 6 to12 hours of 187.
observation by serial ECGs and cardiac injury [11] Hutter AM, Amsterdam EA, Jae AS. Task force 2:
markers to exclude infarction and high-risk un- acute coronary syndromes: section 2B-chest discom-
stable angina before proceeding to exercise test- fort evaluation in the hospital, 31st Bethesda Confer-
ing. However, in the CPU at UC Davis Medical ence. J Am Coll Cardiol 2000;35:85362.
[12] Lewis WR, Amsterdam EA. Chest pain emergency
Center, the approach includes immediate tread-
units. Curr Opin Cardiol 1999;14:3218.
mill testing without a traditional process to rule
[13] Kirk JD, Diercks DB, Turnipseed SD, Amsterdam
out MI. Extensive experience has validated this EA. Evaluation of chest pain suspicious for acute
approach in a large, heterogeneous population. coronary syndrome: use of an accelerated diagnostic
The optimal strategy for evaluating low-risk protocol in a chest pain evaluation unit. Am J Car-
patients presenting to the ED with chest pain diol 2000;85(5A):40B8B.
will continue to evolve based on current research [14] Lewis WL, Amsterdam EA. Observation units, clini-
and the development of new methods. cal decision units and chest pain centers. In:
Becker RC, Alpert JS, editors. Cardiovascular medi-
cine: practice and management. London: Arnold;
References 2001. p. 85103.
[1] Gibler WB, Cannon CP, Blomkains AL, et al. Prac- [15] Amsterdam EA, Lewis WR, Kirk JD, et al. Acute
tical implementation of the guidelines for unstable ischemic syndromes: chest pain center concept.
angina/non-st-segment elevation myocardial infarc- Cardiol Clin 2002;20:11736.
tion in the emergency department. Ann Emerg [16] Lee TH, Cook EF, Weisberg M, et al. Acute chest
Med 2005;46:18597. pain in the emergency room: identication and ex-
[2] Pozen MW, DAgostino RB, Selker HP, et al. Pre- amination of low-risk patients. Arch Intern Med
dictive instrument to improve coronary-care-unit 1985;145:659.
admission practices in acute ischemic heart disease: [17] Goldman L, Cook F, Johnson PA, et al. Prediction
a prospective multicenter clinical trial. N Engl J of the need for intensive care in patients who come
Med 1984;310:12738. to emergency departments with acute chest pain.
[3] Karlson BW, Herlitz J, Wiklund O, et al. Early pre- N Engl J Med 1996;334:1498504.
diction of acute myocardial infarction from clinical [18] Brush JE Jr, Brand DA, Acampora D, et al. Use of
history, examination and electrocardiogram in the the initial electrocardiogram to predict in-hospital
emergency room. Am J Cardiol 1991;68:1715. complications of acute myocardial infarction.
[4] Lown B, Vassaux C, Hood WB, et al. Unresolved N Engl J Med 1980;312:113741.
problems in coronary care. Am J Cardiol 1967;20: [19] Gazes PC, Mobley EM Jr, Faris HM, et al. Prein-
494508. farctional (unstable) angina, a prospective studyd
[5] Karcz A, Holbrook J, Burke MC, et al. Massachu- a ten year follow-up. Prognostic signicance of
setts emergency medicine closed malpractice claims: electrocardiographic changes. Circulation 1973;48:
19881990. Ann Emerg Med 1993;22:5539. 3317.
[6] Pope JH, Aufderheide TP, Ruthazer R, et al. [20] Schroeder J, Lamb IH, Hu M. Do patients in whom
Missed diagnoses of acute cardiac ischemia in the myocardial infarction has been ruled out have a bet-
emergency department. N Engl J Med 2000;342: ter prognosis after hospitalization than those surviv-
116370. ing infarction? N Engl J Med 1980;303:15.
EXERCISE TESTING IN CHEST PAIN UNITS 515

[21] Lee TH, Goldman L. Evaluation of the patient [33] Kerns JR, Shaub TF, Fontanarosa PB. Emergency
with acute chest pain. N Engl J Med 2000;342: cardiac stress testing in the evaluation of emergency
118795. department patients with atypical chest pain. Ann
[22] Mulley AG, Thibault GE, Hughes RA, et al. The Emerg Med 1993;22:7948.
course of patients with suspected myocardial infarc- [34] Lewis WL, Amsterdam EA. Utility and safety of im-
tion: the identication of low-risk patients for early mediate exercise testing of low-risk patients admit-
transfer from intensive care. N Engl J Med 1980; ted to the hospital for suspected acute myocardial
302:9438. infarction. Am J Cardiol 1994;74:98790.
[23] Fineberg HV, Scadden D, Goldman L. Care of [35] Gibler WB, Runyon JP, Levy RC, et al. A rapid di-
patients with a low probability of acute myocardial agnostic and treatment center for patients with chest
infarction: cost eectiveness of alternatives to coro- pain in the emergency department. Ann Emerg Med
nary-care-unit admission. N Engl JM Med 1984; 1995;25:18.
310:13017. [36] Laslett LJ, Amsterdam EA. Management of the
[24] Gaspoz JM, Lee TH, Weinstein MC, et al. Cost- asymptomatic patient with an abnormal exercise
eectiveness of a new short-stay unit to rule out ECG. JAMA 1984;252:17446.
acute myocardial infarction in low risk patients. [37] Gomez MA, Anderson JL, Labros AK, et al. An
J Am Coll Cardiol 1994;24:1249. emergency department-based protocol for rapidly
[25] Braunwald E, Antman EM, Beasley JW, et al. ACC/ ruling out myocardial ischemia reduces hospital
AHA 2002 guideline update for the management of time and expense: results of randomized study
patients with unstable angina and non-ST-segment (ROMIO). J Am Coll Cardiol 1996;28:2533.
elevation myocardial infarction. http://www.acc. [38] Roberts RR, Zalenski RJ, Mensah EK. Costs of an
org/clincal/guidelines/unstable/unstable.pdf. emergency department-based accelerated diagnostic
[26] Braunwald E, Mark DB, Jones RH, et al. Unstable protocol vs hospitalization in patients with chest
angina: diagnosis and management. Agency for pain. JAMA 1997;278:16706.
Health Care Policy and Research Publication No. [39] Zalenski RJ, McCarren M, Roberts R, et al. An
9406062, Rockville, MD. 1994. evaluation of a chest pain diagnostic protocol to ex-
[27] Gra LG, Dallara J, Ross MA, et al. Impact on the clude acute cardiac ischemia in the emergency de-
care of the emergency department chest pain patient partment. Arch Int Med 1997;157:108591.
from the chest pain evaluation registry (CHEPER) [40] Polanczyk CA, Johnson PA, Hartley LH, et al. Clin-
study. Am J Cardiol 1997;80:5638. ical correlates and prognostic signicance of early
[28] Stein RA, Chaitman BR, Balady GJ, et al. Safety negative exercise tolerance test in patients with acute
and utility of exercise testing in emergency room chest pain seen in the hospital emergency depart-
chest pain centers. An advisory from the Committee ment. Am J Cardiol 1998;81:28892.
on Exercise, Rehabilitation, and Prevention, Coun- [41] Mikhail MG, Smith FA, Gray M, et al. Cost-eec-
cil on Clinical Cardiology, American Heart Associa- tiveness of mandatory stress testing in chest pain
tion. Circulation 2000;102:14637. center patients. Ann Emerg Med 1997;29:8898.
[29] Fletcher GF, Balady GJ, Amsterdam EA, et al. [42] Farkouh MI, Smars PA, Reeder GS, et al. A clinical
Exercise standards for testing and training. A trial of a chest-pain observation unit for patients with
statement for healthcare professionals from the unstable angina. N Engl J Med 1998;339:18828.
American Heart Association. Circulation 2001; [43] Brillman J, Mathers-Dunbar L, Gra L, et al. Short
104:1694740. term observation services section of the American
[30] Gibbons RJ, Balady GJ, Bricker JT, et al. ACC/ College of Emergency Physicians. Management of
AHA 2002 guideline update for exercise testing: observation units. Ann Emerg Med 1995;25:92330.
summary article. A report of the American College [44] Kirk JD, Turnipseed S, Lewis WR, Amsterdam EA.
of Cardiology/American Heart Association task Evaluation of chest pain in low-risk patients present-
force on practice guidelines (committee to update ing to the emergency department: the role of immedi-
the 1997 exercise testing guidelines). J Am Coll Car- ate exercise testing. Ann Emerg Med 1998;32:17.
diol 2002;40:153140. [45] Amsterdam EA, Kirk JD, Diercks DB, et al. Imme-
[31] Gibbons RJ, Balady GJ, Beasley JW, et al. ACC/ diate exercise testing to evaluate low-risk patients
AHA guidelines for exercise testing. A report of presenting to the emergency department with chest
the American College of Cardiology/American pain. J Am Coll Cardiol 2002;40:2516.
Heart Association task force on practice guidelines [46] Lewis WL, Amsterdam EA, Turnipseed S, Kirk JD.
(committee on exercise testing). J Am Coll Cardiol Immediate exercise testing of low risk patients with
1997;30:260315. known coronary artery disease presenting to the
[32] Tsakonis JS, Shesser R, Rosenthal R, et al. Safety of emergency department with chest pain. J Am Coll
immediate treadmill testing in selected emergency Cardiol 1999;33:18437.
department patients with chest pain: a preliminary [47] Diercks DB, Kirk JD, Turnipseed S, Amsterdam
report. Am J Emerg Med 1991;9:5579. EA. Exercise treadmill testing in women evaluated
516 AMSTERDAM et al

in a chest pain unit [abstract]. Acad Emerg Med [56] Diercks DB, Kirk JD, Turnipseed S, Amsterdam
2001;8:565. EA. Use of additional electrocardiographic leads in
[48] Kontos MC, Jesse RL, Schmidt KL, et al. Value low-risk patients undergoing exercise treadmill test-
of acute rest sestamibi perfusion imaging for ing [abstract]. Acad Emerg Med 2001;8:564.
evaluation of patients admitted to the emergency [57] Ganim RP, Lewis WR, Diercks DB, et al. Right
department with chest pain. J Am Coll Cardiol precordial and posterior electrocardiographic leads
1997;30:97682. do not increase detection of ischemia in low-risk
[49] Amsterdam EA, Kirk JD, Diercks DB, et al. Assess- patients presenting with chest pain. Cardiology
ment of low risk patients presenting to the emergency 2004;102:1003.
department with chest pain: immediate treadmill test [58] Sabapathy R, Bloom HL, Lewis WR, Amsterdam
or cardiac stress imaging [abstract]? J Am Coll Car- EA. Right precordial and posterior chest leads do
diol 2001;37:149A. not increase detection of positive response in electro-
[50] Senaratne MJ, Carter D, Irwin M. Adequacy of cardiogram during exercise treadmill testing. Am J
an exercise test in excluding angina on patients Cardiol 2003;91:757.
presenting to the emergency department with [59] Amsterdam EA, Kirk JD, Diercks DB, et al. Coro-
chest pain. Ann Noninvas Electrocardiol 1999;4: nary artery disease in patients with multiple emer-
40815. gency department visits and negative immediate
[51] Kirk JD, Turnipseed S, Diercks DB, et al. Interpre- exercise tests for chest pain: an important minority
tation of immediate exercise treadmill test: inter- in a total cohort of over 3,000 low risk patients pre-
reader reliability between cardiologist and senting with chest pain [abstract]. J Am Coll Cardiol
noncardiologist in a chest pain evaluation unit. 2004;43:225A.
Ann Emerg Med 2000;36:104. [60] Lewis WR, Amsterdam EA. Chest pain. In:
[52] Diercks DB, Kirk JD, Turnipseed S, Amsterdam Gershwin ME, Hamilton ME, editors. The pain
EA. Utility of immediate exercise treadmill testing management handbook: a concise guide to diagnosis
in patients taking beta blockers or calcium channel and treatment. Totowa, NJ: Humana Press; 1998. p.
blockers. Am J Cardiol 2002;90:8825. 79115.
[53] Antman EM, Cohn M, Bernin PJLM, et al. The [61] Carter C, Maddock R, Amsterdam EA, et al. Panic
TIMI risk score for unstable angina/nonST eleva- disorder and chest pain in the coronary care unit.
tion MI. JAMA 2000;284:83542. Psychosomatics 1992;33:3029.
[54] Amsterdam EA, Diercks DB, Kirk JD, et al. Multi- [62] Thurston RC, Keefe FJ, Bradley L, et al. Chest pain
ple clinical risk factors do not preclude immediate in the absence of coronary artery disease: a biopsy-
exercise testing in a chest pain evaluation unit. chosocial perspective. Pain 2001;93:95100.
J Am Coll Cardiol 2003;41:349A. [63] Amsterdam EA, Diercks D, Kirk JD, et al. Evalua-
[55] Michaelides AN, Psomdaki ZD, Dilaveris PE, et al. tion of low risk women presenting to the emergency
Improved detection of coronary artery disease by department with chest pain: is early stress testing
exercise electrocardiography with the use of right necessary for risk stratication? Circulation, in
precordial leads. N Engl J Med 1999;340:3405. press.
Cardiol Clin 23 (2005) 517530

Imaging in the Evaluation of the Patient with


Suspected Acute Coronary Syndrome
Michael C. Kontos, MDa,b,*, James L. Tatum, MDa
a
Virginia Commonwealth University, VCU Medical Center, Room 7074 North Hospital,
1300 E Marshall Street, PO Box 980051, Richmond, VA 232980051, USA
b
Room 7-074, Heart Station, North Hospital, P.O. Box 980051, Medical College of Virginia,
12th and Marshall Streets, Richmond, VA 23298-0051, USA

Over the last decade, major advances have be made with more rapid and comprehensive
been made in the treatment of acute coronary information. Advantages of rapid, accurate di-
syndromes (ACSs). However, eective implemen- agnosis are obvious and include early initiation of
tation of these treatments requires timely and appropriate therapy for those who have myocar-
accurate identication of the high-risk patient dial infarction (MI) or unstable angina; reduction
among all those presenting to the emergency in the inadvertent discharge of patients who have
department (ED) with symptoms suggestive of ongoing ischemia; and shortened length of stay
ACS. In the patient population presenting with and reduced admissions for patients who have
a diagnostic ECG or those who have typical chest noncardiac chest pain.
pain and a history of coronary disease, the initial Before implementing a new diagnostic tool
triage and treatment strategies are straight-for- or test, several questions require an armative
ward and guideline driven. However, for most answer, including whether it is accurate (ie,
patients the diagnosis is not clear-cut and further sensitive and specic), if it adds incremental
evaluation is required. The population of patients diagnostic information to that which is already
considered low risk after the initial evaluation available, and if it is cost-eective. Over the last 15
accounts for nearly two thirds of ED chest pain years, numerous studies examining the usefulness
patients [1,2], representing as many as 4 million of acute ED myocardial perfusion imaging (MPI)
patients a year in the United States. An additional have conrmed that the answer to these questions
consideration is that the opportunity for improv- is yes.
ing outcomes is time-dependent, so that early
identication of the patient who has true ACS is
Acute myocardial perfusion imaging
essential. This necessity further increases the
need for rapid triage tools, especially in the cur- Knowledge of the ischemic cascade presup-
rent setting of ED and hospital overcrowding poses that in vivo noninvasive imaging of
that has become the norm in large urban centers. myocardial perfusion should provide an optimal
In an eort to better identify patients who do technique for not only detecting (diagnosis) but
not have myocardial ischemia, ED physicians have also grading (risk stratication) ACS. This pre-
increasingly relied on technological advances, sumption is conrmed by studies dating back over
such as new bedside biomarkers and advanced 3 decades. Wackers and colleagues [3] and others
imaging techniques. As these tools have become used planar myocardial perfusion imaging with
more available in the ED, triage decisions can thallium in the 1970s to demonstrate that acute
imaging had signicant incremental diagnostic
power in patients who had nonST-elevation
* Corresponding author. ACS [4]. However, the intrinsic characteristics
E-mail address: mckontos@vcu.edu (M.C. Kontos). of thallium as a radiotracer and the inherent
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.008 cardiology.theclinics.com
518 KONTOS & TATUM

limitations of planar imaging prevented wide- In the presence of signicant ischemia, the con-
spread adoption of the technique in the clinical comitant presence of a wall motion abnormality
management of ACS. Although some of the issues in conjunction with perfusion defects on acute
related to the radionuclide could be mitigated by imaging not only adds specicity but is a addi-
performing imaging immediately in the ED [5,6], tional prognostic indicator. Recent data indicate
the limitations of planar imaging required im- this is true. A recent multicenter study by Kaul
provements in technology to overcome the low and colleagues [11] found that perfusion plus re-
sensitivity for detecting small areas of ischemia gional function provided substantially greater di-
and for detecting ischemia in the posterior dis- agnostic and prognostic value for predicting
tribution. The development of single-photon outcomes in 163 patients who had possible ACS.
emission computed tomography (SPECT) in In our study of 2286 consecutive patients who
combination with the superior image quality pro- were admitted for exclusion of ischemia following
vided by technetium-labeled myocardial perfusion acute rest MPI, we found that among the patients
agents led to observational studies followed by who had Troponin-I (TnI) elevations (4.0% ver-
clinical trials demonstrating the usefulness of sus 3.5%), creatine kinase-myocardial band (CK-
acute ED MPI. Technetium 99m (Tc-99m) sesta- MB) MI (1.7% versus 1.5%), or who underwent
mibi and tetrofosmin are taken up by the myocar- revascularization (5.6% versus 5.1%), there was
dium in proportion to blood ow, similar to no dierence in the proportion of patients who
thallium, but do not undergo signicant redistri- had perfusion defects but had normal wall motion
bution [7]. Therefore, patients can be injected in that area and those who had normal perfusion
while experiencing symptoms and imaged up to and function. In contrast, those who had perfu-
several hours later, making it possible to perform sion defects associated with abnormal wall motion
high-quality SPECT imaging outside of the ED were signicantly more likely to have TnI eleva-
setting in the absence of dedicated equipment. tions (15%), CK-MB MI (10%), and undergo re-
The favorable energy and dosimetry of Tc-99m vascularization (16%).
also allow gated acquisition and reconstruction of
dynamic functional images, thereby providing Diagnostic value
simultaneous assessment of regional and global
Sensitivity: acute myocardial infarction
ventricular function that can be correlated with
perfusion defects [8]. In the setting of acute infarc- Several studies to assess the diagnostic accu-
tion or ischemia, wall motion is typically abnor- racy of acute ED MPI have found a high sensitiv-
mal; in contrast, an apparent perfusion defect in ity for identifying patients who have MI (Table 1)
the presence of normal wall motion and thicken- [3,1218]. However, because acute MPI is per-
ing on gated SPECT usually indicates an artifact formed predominantly in low-risk patients, the
such as tissue attenuation. The ability to perform absolute numbers of patients who have MI in
simultaneous wall motion and perfusion signi- any individual study is small, resulting in an im-
cantly improves specicity [9,10] and is particularly precise measurement of sensitivity. In a study
valuable in the acute setting where serial images that included 141 patients who had CK-MB
are not available. MI, we found that sensitivity was 89% (95%

Table 1
Diagnostic accuracy of rest myocardial perfusion imaging in patients who have acute chest pain syndrome and normal or
nonischemic rest electrocardiograms
Year N Tracer Sensitivity Specicity NPV End point
Wackers et al [3] 1979 203 Tl-201 100% 72% 100% AMI
Varetto et al [12] 1993 64 Tc-mibi 100% 92% 100% CAD
Hilton et al [13] 1994 102 Tc-mibi 94% 83% 99% CAD/AMI
Tatum et al [14] 1997 438 Tc-mibi 100% 78% 100% AMI
Kontos et al [15] 1997 532 Tc-mibi 93% 71% 99% AMI
Heller et al [16] 1998 357 Tc-tetro 90% 60% 99% AMI
Kontos et al [17] 1999 620 Tc-mibi 92% 67% 99% AMI
Udelson et al [18] 2002 1215 Tc-mibi 96% NR 99% AMI
Abbreviations: AMI, acute myocardial infarction; CAD, angiographic coronary artery disease; NPV, negative predic-
tive value; NR, not reported; Tc-mibi, Tc-99m-sestamibi; Tc-tetro, Tc-99m-tetrofosmin; Tl-201, thallium 201.
MPI IN CHEST PATIENTS 519

condence interval, 83%94%), similar to prior identify patients who have unstable angina. In
reported studies. As would be expected, patients one of the rst studies, Bilodeau and colleagues [9]
who had negative MPI had small MIs as estimated imaged 45 patients who did not have a previous
by peak creatine kinase (CK) values, with an av- history of MI and had been admitted for unstable
erage CK of 313 G 227 U/L, compared with angina. The presence of a perfusion defect had
590 G 620 U/L (P ! .001) in those who had a sensitivity of 96% and specicity of 79% for
positive MPI [19]. predicting angiographic coronary disease in pa-
Another confounding issue with these studies tients injected during an episode of pain, com-
is that most used elevations in either CK or CK- pared with a sensitivity of only 65% for the
MB as the diagnostic standard for MI. However, ECG. Although specicity of acute MPI appears
because of its high sensitivity and specicity for low when MI is used as the end point, an impor-
detecting myocardial necrosis [20,21], current tant advantage is to remember that the entire con-
guidelines recommend that cardiac troponin be tinuum of ACS remains a diagnostic dilemma.
the diagnostic standard for MI [22]. Because ap- Using revascularization or signicant disease as
proximately 3% to 4% of the left ventricle must a surrogate for unstable angina, abnormal images
be ischemic to allow detection by MPI [23], the are found in a high proportion, and overall spec-
number of patients who have troponin elevations icity and predictive value are increased. For ex-
and negative MPI would be expected to be higher ample, we found that in 532 patients admitted
than the previous standard. This assumption has after acute rest MPI in the ED, acute MI as as-
been supported by several small studies that re- sessed by CK-MB elevations was present in only
ported that, although the sensitivity of MPI was 15% of patients who had positive MPI (Fig. 2)
high, it was signicantly lower than that of TnI [15]. However, most patients who have positive
or Troponin-T [17,24]. In a larger study, we ana- MPI had an end point consistent with ACS, in-
lyzed outcomes in 319 patients who were initially cluding acute infarction, subsequent revasculari-
considered low risk for MI and underwent acute zation, or signicant coronary disease (O70%
rest MPI as part of standard chest pain evaluation stenosis) on coronary angiography.
protocol and were subsequently found to have el-
evated TnI values [25], thus meeting the ACC/
Negative predictive value and prognosis
ESC denition for MI [22]. Seventy-seven patients
had negative MPI, giving a sensitivity of only A high negative predictive value is critical for
76%; much lower than when CK or CK-MB the value of acute MPI to serve as a risk strati-
was used as the diagnostic standard for MI. How- cation tool. The ability to exclude signicant
ever, among the 77 patients who had negative ischemia in more than 99% of patients allows
MPI, more than half (n 47, 61%) had a peak eective identication of those who can be placed
CK-MB of more than 8 ng/mL and therefore in lower-intensity settings other than the coronary
would not previously have been diagnosed with care unit (CCU) or can be discharged home (see
MI (Fig. 1). Patients who had negative MPI had Table 1). The predictive value of a negative acute
signicantly lower peak CK-MB values (15 G MPI also extends beyond the immediate setting in
25 ng/mL versus 45 G 78 ng/mL, P ! .001), identifying patients at low risk for short- and
had higher ejection fractions (56% G 15% versus long-term cardiac complications. For example,
47% G 13%, P ! .01), and were more likely to Hilton and colleagues [26] found that patients
have nonsignicant disease (60% versus 85%, who had normal perfusion imaging had an excel-
P ! .001) than those who had positive MPI. lent prognosis, with no late events at 90-day fol-
These ndings demonstrate that acute MPI con- low-up. Similarly, we reported that patients who
tinues to provide additional risk/prognostic infor- had negative acute MPI had a cardiac event rate
mation even with the introduction of a newer of only 3% during the subsequent year [14]. In
generation of biomarkers. our experience over the last 9 years, low-risk pa-
tients discharged from the ED after undergoing
acute ED rest MPI (n 10,775) demonstrated
Acute coronary syndrome without myocardial
a 30-day cardiac mortality of only 0.08%, with
infarction
only nine cardiac deaths during the follow-up
In the past 15 years, a large body of evidence period.
has demonstrated that acute MPI using Tc-99m An important prognostic parameter provided
myocardial perfusion agents can accurately by acute MPI is that it not only identies patients
520 KONTOS & TATUM

A 70
61 (-) MPI
60 (+) MPI

50
Patients
39
40
33 34

30
22
20 18
16 15
14
9.1
10

0
<5 ng/mL <8 ng/mL 8-16 ng/mL 16-32 ng/mL >32 ng/mL

B 60
55
(-) MPI
50 (+) MPI

40
36
Patients

33
30

21
20
15 16

9.5
10 6.5 7
2.6
0
<200 U/L 2-300 U/L 3-400 U/L 4-500 U/L >500 U/L

Fig. 1. Proportion of patients who had positive (black bars) and negative (white bars) rest MPI and levels of (A) CK-MB
and (B) CK elevations. Of 104 patients who had creatine kinase myocardial band (CK-MB) level !8 ng/mL, 61% had
negative images, whereas of 192 patients who had CK-MB R8 ng/mL, only 27% had negative images (sensitivity 83%).
(Modied from Kontos MC, Fratkin MJ, Jesse RL, et al. Sensitivity of acute rest myocardial perfusion imaging for iden-
tifying patients with myocardial infarction based on a troponin denition. J Nucl Cardiol 2004;11:129; with
permission.)

who have ACS but also provides a validated Even in the absence of ischemic ECG changes,
measurement of the ischemic risk area. The size the ischemic risk area can be large. We found that
of a perfusion defect is of signicant clinical the ischemic risk area ranged from 0% to 62%
importance, as patients who have larger defects of the left ventricle, with a mean risk area of
have a worse long-term prognosis [27,28]. The 18% G 11%. Even patients who had normal ECGs
most important determinant of infarct size is the at the time of presentation had risk areas similar
ischemic risk zone or the amount of myocardium to those of patients who had abnormal but non-
in jeopardy [29]. MPI is the only technique among ischemic ECGs (16% G 12% versus 19% G 12%,
those commonly available that can determine the P .25) (Fig. 3) [35]. One explanation for this
ischemic risk zone [30,31]. In studies in which nding is that MI in these patients is often caused
post-MI patients had MPI before discharge, by occlusion of the left circumex coronary artery,
defect size correlated well with other outcome and the myocardial territory supplied by it is
predictors, including left ventricular ejection frac- silent on a surface ECG [36]. In a small study
tion [32,33], regional wall motion index [32], end- that included patients who had ST-segment de-
systolic volume [33], and peak CK levels [34]. pression, Christian and colleagues [37] performed
MPI IN CHEST PATIENTS 521

60 predict small MIs [36]. OKeefe and colleagues [38]


53 reported that the risk area was not signicantly dif-
50 ferent in patients who had left circumex occlusion
42 associated with ST-segment elevation, left circum-
Cardiac Events,

40
ex occlusion without ST-elevation, and right
30
coronary artery occlusion. Consistent with these
results, we found that the ischemic area at risk
20 was similar in patients in whom the infarct-related
15
artery was the left circumex (18% G 10%, me-
10
4.4
6.4 dian 19%), the right coronary (18% G 13%, me-
0.6 dian 17%), or the left anterior descending artery
0
MI MI/R MI/R/S (18% G 10%, median 19%) [35].

Fig. 2. Outcomes associated with results of acute rest


Incremental diagnostic value
MPI. Patients who had positive rest MPI (dark bars)
had signicantly (P ! .0001) more MI; MI or revascular- Although a new test may oer high diagnostic
ization (MI/R); and MI, revascularization, or signicant accuracy, it may have limited value unless it
coronary artery disease (O70% stenosis) (MI/R) or sig-
provides signicant incremental value over the
nicant disease than patients who had negative rest myo-
current standard of care. Because the existing
cardial perfusion imaging (white bars). (Modied from
Kontos MC, Jesse RL, Schmidt KL, et al. Value of acute standard usually includes testing that is readily
rest sestamibi perfusion imaging for evaluation of available and less expensive, the barrier to in-
patients admitted to the emergency department with clusion of techniques such as acute MPI is that
chest pain. J Am Coll Cardiol 1997;30:97682; with they must demonstrate a clear benet.
permission.) Studies performed in lower-risk patients have
demonstrated that acute ED MPI does oer
early MPI in 14 patients who did not have ST- incremental value. In one of the rst studies,
segment elevation and later underwent coronary Hilton and colleagues [13] used Tc-99m sestamibi
angiography. The culprit vessel was the circumex SPECT imaging to study 102 patients presenting
coronary artery in six (43%) patients. In another to the ED with typical angina and either a normal
study of 79 patients presenting with a nonischemic or nondiagnostic ECG. Seventy patients had a nor-
ECG and acute MI, we found that the left circum- mal perfusion scan, only one of whom had a cardiac
ex coronary artery was the infarct-related artery event. In comparison, 2 of the 15 (13%) patients
in 42% of the cases [35]. who had equivocal scans and 12 of the 17 (71%)
In patients who have left circumex occlusions, patients who had perfusion defects had cardiac
the absence of ischemic ECG changes does not events. When equivocal scans were classied as

30

25 23
22
21
20
# Patients

15

10
10

5
5
1 1
0
0-9 10-19 20-29 >30

Fig. 3. Distribution of the risk area sizes as a percent of the left ventricle. Dark bars represents patients who had ECG
evidence of prior infarction. (Modied from Kontos MC, Kurdziel KA, Ornato JP, et al. A nonischemic electrocardio-
gram does not always predict a small myocardial infarction: Results with acute myocardial perfusion imaging. Am Heart
J 2001;141:36066; with permission.)
522 KONTOS & TATUM

abnormal, the sensitivity and specicity of an ab- routine care or ED MPI, in which patients were
normal study for predicting adverse cardiac events injected with sestamibi in the ED and subse-
were 94% and 83%, respectively. quently underwent acute imaging, with the results
Similarly, we found that abnormal MPI was called back to the ED physician [18]. All patients,
the most important independent predictor of MI whether admitted or discharged, underwent marker
or revascularization in 532 patients who under- analysis and diagnostic evaluation. There was no
went acute ED MPI [15]. Finally, Heller and col- dierence in the percentage of patients who had
leagues [16] found that that abnormal SPECT was ACS and either MI (97% versus 96%) or unstable
the most important multivariate predictor of MI angina (83% versus 81%) who were admitted,
in 357 patients who underwent acute MPI. In ad- with one patient who had MI from each group
dition, acute MPI added signicant incremental discharged from the ED. However, there was a sig-
diagnostic value after consideration of demo- nicantly lower admission rate and a higher rate
graphic, clinical, and ECG variables (Fig. 4). of direct discharge from the ED in the ED MPI
In an interesting intent-to-treat survey study, arm compared with the standardized care arm.
Knott and colleagues [39] performed acute MPI
on 120 patients in the ED. The requesting physi-
cian completed a questionnaire before imaging,
asking what the proposed management would be Other issues
had the test not been available. They found
Radiopharmaceutical
a 34% reduction in overall hospital admissions
and a 59% reduction in planned CCU admissions. Although most studies were performed with
Overall, CCU admissions were not reduced be- sestamibi, comparable results have been obtained
cause 17 patients initially considered low risk with tetrofosmin. In a multicenter study, Heller
were admitted to the CCU after MPI was found and colleagues [16] found a sensitivity of 90% in
to be abnormal. 357 patients who underwent acute MPI. Negative
Data from several observational studies were predictive value (NPV) was equally high at 99%,
conrmed in the Emergency Room Assessment of with only two patients who had small nonQ-
Sestamibi for Evaluating Chest Pain (ERASE) wave MIs having negative acute MPI. In our
study, a large prospective randomized controlled study in which 319 patients had TnI elevations,
study. This study demonstrated that acute MPI sensitivity of those who had acute MPI with sesta-
was eective when compared with standard man- mibi (75%) and tetrofosmin (80%) were similar,
agement of low-risk chest pain patients [18]. In as was the proportion who had CK-MB MI
this study, 2475 patients were randomized to (84% versus 87%) [16].

18
17.9
Incremental Chi-Square

15

12

7.8
6
5.5
3

0
Age, Gender A+>3 RF B+ECG, CP C+SPECT
(A) (B) (C)

Fig. 4. Incremental prognostic value of rest tetrofosmin SPECT imaging over clinical variables. Model A: clinical vari-
ables; Model B: A plus three or more risk factors (RF); Model C: B plus admission ECG and chest pain (CP) at the time
of tetrofosmin injection. (Modied from Heller GV, Stowers SA, Hendel RC, et al. Clinical value of acute rest techne-
tium-99m tetrofosmin tomographic myocardial perfusion imaging in patients with acute chest pain and nondiagnostic
electrocardiograms. J Am Coll Cardiol 1998;31:101117; with permission.)
MPI IN CHEST PATIENTS 523

Comparison with troponin prior MI are at higher risk for acute events and
are usually not candidates for primary triage to
Although markers of necrosis, such as tropo-
a subsequent outpatient evaluation. Sensitivity for
nin, are considered the gold standard for identi-
identifying necrosis is imperfect for MPI, as at
fying MI, rest MPI does have some important
least 3% to 5% of the left ventricle must be
advantages over using markers alone. First, myo-
ischemic for a defect to be detected. However, the
cardial markers are by denition abnormal only in
many patients who have negative rest MPI despite
patients who have necrosis, and therefore are
marker elevations have nonsignicant disease on
negative in patients who have ischemia alone.
coronary angiography [25], and therefore are at
Also, given the time required for imaging and
low risk for short-term adverse outcomes, al-
processing, acute MPI results can be available
though aggressive risk factor modication would
within 1 to 2 hours after injection. In contrast,
still be indicated. Therefore, rather than being
markers of necrosis are not detectable in the blood
seen as competitive diagnostic tools for evaluat-
until several hours after the damage has occurred.
ing patients presenting to the ED with chest
To achieve a high sensitivity, sampling must be
pain, markers and MPI should be considered
performed over an 8- to 9-hour period [40]. Thus,
complementary.
the sensitivity of MPI is signicantly higher than
that of the initial troponin (Fig. 5).
Timing of tracer injection
A second advantage of acute MPI is that it can
identify risk area, and therefore better quantitate Although it appears that sensitivity of acute
overall cardiac risk. For example, two patients MPI decreases as the symptom-free interval in-
who have similar low-peak TnI valuesdone creases, the reduction is highly variable. In the
secondary to occlusion of a small branch vessel case of thallium, Wackers and colleagues [3] per-
and the other resulting from a brief occlusion of formed thallium-201 scintigraphy in 98 patients
the proximal portion of a major vesseldhave admitted with chest pain who had MI excluded.
markedly dierent areas at risk and the potential When imaged within 6 hours of the last anginal
for markedly dierent outcomes, which can be symptoms, 57% of the patients had abnormal
readily determined using MPI. studies; however, when imaged after 12 hours,
Acute MPI has some limitations when used to only 8% had abnormal studies. Similar results
assess patients who have chest pain. Acute MI, have been reported by Van der Wieken and col-
acute ischemia, and prior MI all cause perfusion leagues [4].
defects, and dierentiation is not possible based Studies using technetium agents also dem-
on the images alone. However, patients who have onstrate higher sensitivity when injection is

100 97
92

81 82
80 75
Cardiac Events,

60 *
53

#
40
*
30 * *
# 26 # 26 #
20 * * *
10 10 10

0
MI Rev Sig MI+Sig

Fig. 5. Sensitivity of MPI (dark bars), the initial troponin I (open bars), and troponin I (light bars) for identifying end
points. Asterisk denotes P ! .001 compared with perfusion imaging; # denotes P ! .001 compared with serial troponin.
Rev, revascularization; Sig, signicant disease (O70% stenosis). (Data from Kontos MC, Jesse RL, Anderson FP, et al.
Comparison of myocardial perfusion imaging and cardiac troponin I in patients admitted to the emergency department
with chest pain. Circulation 1999;99:20738.)
524 KONTOS & TATUM

performed during or soon after pain. Bilodeau sestamibi at varying intervals after balloon ina-
and colleagues [9] found that the sensitivity of tion, although the size of the perfusion defect de-
MPI for detection of coronary artery disease creased as the interval after the procedure
was 96% in 45 patients injected with tracer during increased. One explanation is that the pharmaco-
chest pain. When the same patients were rein- dynamics of sestamibi are dependent on mem-
jected later while pain-free, the sensitivity had brane and mitochondrial functional integrity,
decreased to 65%. However, in both cases the sen- which may be depressed as a result of lingering
sitivity was signicantly higher than that of the metabolic alterations, especially in high-energy
initial ECG. In contrast, we found that when pa- metabolites that occur after transient ischemia.
tients were injected within 6 hours of symptoms, All of these ndings strongly support the
sensitivity was similar for identifying patients who conclusion that the sensitivity of acute MPI will
had MI, revascularization, or signicant coronary be dependent on the extent, duration, severity,
disease between those who were injected with and and reperfusion status of the ischemic insult. It
without symptoms [15]. Others have also reported must be kept in mind that chest pain is not the
high sensitivity despite the absence of symptoms gold standard for myocardial ischemia and that
during injection [12]. silent ischemia is common. In patients who have
One explanation relates to the dierence in the been pain-free for prolonged periods, acute MPI
underlying mechanisms causing perfusion defects will have a lower sensitivity for detection of the
in patients who have ACS as compared with those presence of coronary disease, but persistent ab-
undergoing stress testing. Rather than causing normality suggests a more complex and possibly
true ischemia, stress perfusion defects result from unstable condition associated with higher risk.
ow heterogeneity between areas supplied by
coronary arteries with and without signicant
stenoses. The perfusion tracer is injected at the
Special populations
time of maximal ow imbalance, with a rapid
return of coronary ow to baseline once the Although not normally considered candidates
patient stops exercising. In contrast, in patients for acute MPI because of their higher pretest
who have ACS, perfusion defects result from the likelihood of ischemia, rest MPIs can provide use-
combination of intermittent thrombotic occlusion ful additional diagnostic information in selected
and vasoconstriction in the setting of complex cor- subgroups of patients who have known coronary
onary morphology [41], resulting in marked de- artery disease. This group includes patients who
creased coronary blood ow [42] and persistently have a nonischemic ECG and atypical symptoms
decreased regional myocardial perfusion [41]. Be- (particularly if they are dierent from their typical
cause regional hypoperfusion is one of the rst angina), those who have had a recent negative
steps in the ischemic cascade, symptoms of chest cardiac evaluation, or those in whom the risk for
discomfort are often a late clinical manifestation, coronary angiography is increased, such as pa-
so that regional hypoperfusion will frequently be tients who have signicant renal disease. In pa-
present even in the absence of symptoms. Subop- tients who have multivessel disease, such as those
timal ow frequently persists despite prolonged who have had prior bypass surgery, the ability to
treatment with newer pharmacologic treatments, determine risk area can be used to delineate the
such as glycoprotein IIb/IIIa antagonists [43,44]. culprit lesion.
Perfusion defects may also result from distal It should be recognized that patients who have
embolization of a proximal thrombus leading to prior MI, especially those who have Q waves, are
downstream microvascular obstruction [45]. In likely to have perfusion defects, and that sub-
a study of 75 patients who underwent sestamibi sequent repeat rest-imaging after a pain-free
injection during rotational atherectomy, a proce- period is required to dierentiate new ischemia
dure in which distal embolization of micropar- from old infarction. Alternatively, if prior images
ticles is frequent, perfusion defects were present are available, they can be used for comparison to
in 65% of patients [45]. determine the signicance of perfusion defects.
Finally, a third potential mechanism was When deciding whether to pursue further invasive
reported in an interesting study of 40 patients or noninvasive treatment, it has been our experi-
who had a percutaneous intervention. Fram and ence that minor dierences in images when
colleagues [46] found that perfusion abnormali- compared with prior studies are unlikely to be of
ties persisted in patients injected with Tc-99m clinical signicance.
MPI IN CHEST PATIENTS 525

Another group of patients in whom ED rest In addition, recommendations for using MPI
MPI can be useful are those presenting with in the ED have been published [54]. The recom-
cocaine-associated chest pain. In the absence of mended patient selection criteria are similar to
ischemic ECG changes or known coronary dis- those used for admission to a chest pain evalua-
ease, the risk for ACS is low [47]. Rather than ad- tion unit. Patients should be low risk (no ischemic
mitting the patient, an alternative evaluation ECG changes or history of coronary disease) and
process is to perform rest MPI, with discharge if hemodynamically stable. The optimal use of MPI
images are negative. We found that in 216 consec- as a triage tool is in patients who will be dis-
utive patients who had chest pain after recent co- charged home and have stress testing as an outpa-
caine use who underwent ED MPI, only ve tient if imaging is negative.
(2.3%) patients had abnormal studies, including One of the rst programs to incorporate rest
two who had acute MI [48]. None of the 38 pa- MPI as a strategy was at Virginia Commonwealth
tients who had normal MPI had subsequently University Medical Center (formerly Medical
acute MI by biomarkers after admission to the College of Virginia). In contrast to most chest
CCU, and only 7% of the 67 patients undergoing pain programs, the systematic chest pain protocol
subsequent stress MPI had reversible myocardial developed and implemented is designed for all
perfusion defects. At 30-day follow-up there chest pain patients, with MPI used for evaluation
were no cardiac events in patients who had nor- of lower-risk patients (Table 2) [14]. All patients
mal rest MPI. This nding indicates that hospital presenting to the ED with chest pain or other
admission can be avoided in this subgroup of pa- symptoms consistent with myocardial ischemia
tients who have a history of cocaine use if rest undergo rapid evaluation with assignment to a tri-
MPI is negative. age level, which is based on the probability of hav-
ing MI or myocardial ischemia derived from
clinical and ECG variables. After the initial eval-
Cost-eectiveness
uation, patients thought to be at high risk (those
The ability to discharge patients directly from who have ischemic ECG changes and those who
the ED has obvious cost implications. Despite the have known coronary disease experiencing typical
application of complex and expensive technology, symptoms) and characterized as levels 1 and 2 are
ED MPI can be cost-eective if the number of admitted directly to the CCU. Patients considered
patients admitted is decreased [16,4951]. Several low to moderate risk for ACS (eg, absence of is-
observational studies have conrmed that cost re- chemic ECG changes or history of coronary dis-
ductions occur when rest MPI is used as an inte- ease) undergo further risk stratication using
gral part of patient management. Costs are acute rest MPI [14]. Level 3 patients are admitted
reduced in two ways. One obvious mechanism is as observation patients and undergo a rapid rule-
by discharging more patients directly from the in protocol. Level 4 patients are evaluated in the
ED, with an increase in the admission of more ED. If images are either negative or unchanged
appropriate patients. A second mechanism is by from previous studies, patients are discharged
more appropriate selection of diagnostic proce- home and scheduled for outpatient stress testing.
dures, as the rate of coronary angiography in If MPI is positive, they are admitted and ad-
low-risk patients is reduced [39,49]. A preliminary vanced to the level 2 treatment protocol.
analysis from the ERASE study conrms that us- It is important to appreciate the dierence in
ing ED MPI as a key part of the initial diagnostic the role of acute rest MPI between level 3 and
strategy was cost-eective, with costs reduced to level 4 patients. In level 3 patients, the presence of
an average of $70 per patient [52]. a signicant perfusion defect identies a high-risk
patient in whom early initiation of aggressive
treatment is indicated, with the potential for early
Incorporation into chest pain evaluation
intervention. Negative MPI and negative markers,
Recent updated guidelines for the clinical use on the other hand, identify patients who can safely
of cardiac radionuclide imaging indicate that undergo early stress testing and discharge. Al-
acute ED MPI has a Class IA indication for though the identication of higher-risk patients is
assessing myocardial risk in patients who have the focus of much interest, the ability to better
possible ACS and have nondiagnostic ECGs and risk-stratify intermediate-risk patients into low-
negative initial serum markers and enzymes, if risk who can be stressed safely is an important
available [53]. advantage. In contrast, the role of MPI in patients
526
Table 2
Acute chest pain diagnostic treatment pathways at Medical College of Virginia/Virginia Commonwealth University Medical Center
Primary risk Probability Probability Diagnostic Secondary risk Treatment
assignment of AMI of ischemia criteria Disposition stratication strategy
Level 1: Very high Very high Ischemic ST elevation Admit CCU Serial ECGs Fibrinolytics within 30 min
AMI (O95%) (O95%) Acute posterior MI Cardiac markers every Primary PCI within 90 min
68 h until peak
Level 2: Moderate High Ischemic ECG Admit CCU Serial ECGs ASA
Denite or (10%50%) (20%50%) Acute CHF Fast track Cardiac Markers IV UFH or SC LMWH
highly Known CAD with rule-in protocol at 0, 3, 6, and 8 h IV and/or topical NTG
probable typical symptoms If rule-in for AMI IV and/or PO Beta Blocker
ACS (markers positive) Clopidogrel
continue every 68 h GP IIb/IIIa inhibitor if TnI
until peak positive
Cath/PCI

KONTOS & TATUM


Level 3: Low Moderate Nonischemic ECG and Observation MPI ASA
Probable (1%10%) (5%20%) either: Fast track Cardiac Markers If cardiac markers or MPI
ACS Typical symptoms rule-in protocol at 0, 3, 6, and 8 h positive: treat per level 2
O30 min, no Serial ECGs protocol
CAD, or If negative: stress
Atypical symptoms
O30 min, known
CAD,
Level 4: Very low Low Nonischemic ECG and ED evaluation MPI MPI positive: admit to
Possible (!1%) (!5%) either: level 2 treatment protocol
UA Typical symptoms MPI negative: discharge;
!30 min, or schedule out patient stress
Atypical symptoms,
Level 5: Very low Very low Evaluation must ED evaluation as As appropriate for the As appropriate for the
Very low (!1%) (!1%) clearly document a deemed necessary clinical condition clinical condition
suspicion noncardiac etiology
for AMI for the symptoms
or UA
Abbreviations: ACS, acute coronary syndrome; AMI, acute myocardial infarction; ASA, aspirin; CAD, coronary artery disease; CCU, coronary care unit; CHF, conges-
tive heart failure; ED, emergency department; GP, glycoprotein; IV, intravenous; LMWH, low molecular weight heparin; MPI, myocardial perfusion imaging; NTG, nitro-
glycerin; PCI, percutaneous intervention; TnI, troponin I; UA, unstable angina; UFH, unfractionated heparin.
MPI IN CHEST PATIENTS 527

p = NS pain and other symptoms suggestive of ischemia


p < .002
35 p < .001 remain nonspecic and frequent presentations to
30 the ED. The ECG appropriately remains the rst
Cardiac Events ( )

25 Level 2 triage tool and is essential for identifying the


p = NS p < .001 Level 3 patient who requires admission; however, numer-
20 Level 4
p < .001 ous studies have demonstrated that it is inade-
15 (+) MPI quate for triaging most of the remaining patients.
(-) MPI
10 p < .001 Acute MPI has been shown to have a high NPV
p < .05
and is a powerful risk stratication tool. When
5
MPI is used in conjunction with the newer
0 biomarkers such as troponin, the combination
MI MI or Revasc
provides an impressive risk prole.
Fig. 6. Cardiac outcomes compared with the initial tri- The inherent success of acute MPI in the ACS
age level assignment and the MPI results. The incidence population is not serendipitous but rather based
of MI or MI or revascularization was signicantly dier-
on the knowledge of the pathophysiology of ACS
ent among level 2 (hatched bars), level 3 (vertical bars),
and the ischemic cascade that has been elucidated
and level 4 (diagonal bars) patients. Patients who had
positive imaging (dark bars) had an incidence of MI or through the work of many investigators. The fact
MI or revascularization similar to the level 2 patients. that one of the earliest events in the cascade is
(Adapted from Tatum JL, Jesse RL, Kontos MC, et al. a marked reduction in absolute blood ow,
Comprehensive strategy for the evaluation and triage leading to metabolic ischemia and its functional
of the chest pain patient. Ann Emerg Med 1997; consequences, led researchers decades ago to
29:11623; with permission.) attempt to noninvasively detect this early signa-
ture of ACS. However, as is often the case, these
who are considered level 4 is to diagnose un- early investigators were limited by available
suspected ACS and prevent the inadvertent dis- technology. With the development of technolog-
charge of these patients from the ED. Follow-up ical advances such as SPECT imaging and the
stress testing is used to exclude signicant coro- introduction of Tc-99mlabeled MPI agents, it
nary disease. became possible to translate basic science obser-
This simple risk stratication scheme accu- vations into meeting a recognized patient need.
rately separates patients into high-, intermediate-, Importantly, the ability to noninvasively assess
and low-risk groups. The ability of MPI to further perfusion acutely in combination with a quantita-
risk stratify lower-risk patients is also obvious, as tive measurement of function has important
outcomes in patients who have positive MPI are prognostic power. The studies that followed
similar to those in the patients considered high-risk have conrmed that when this information is
level 2 (Fig. 6) [14]. Close collaboration of the applied to the appropriate population in a timely
CCU and nuclear medicine sta has resulted in manner it has signicant clinical impact. A key
the ability to successfully triage patients who do aspect of the clinical success of nuclear cardiol-
not have known coronary disease and have large ogy techniques in this and other settings has been
perfusion defects at the time of imaging directly the ability to standardize all aspects of this
to coronary angiography and revascularization. process, from image acquisition to computer-
Patients presenting with chest pain are evalu- assisted diagnosis. As new techniques are applied
ated similarly to the Medical College of Virginia to the ACS population, the same questions posed
protocol. Patients who present without chest pain earlier will need to be answered for each new
undergo rest SPECT thallium imaging. If images technique.
are negative, subsequent stress SPECT Tc-99m
sestamibi imaging is immediately performed. If
rest images are abnormal, the patient is re-evalu-
ated for possible ACS before stress testing. References
[1] Faroghi A, Kontos MC, Jesse RL, et al. Are cardiac
risk factors predictive in low risk emergency depart-
Summary ment chest pain patients who have chest pain? J Am
Coll Cardiol 2000;35:379A.
The rapid triage of the patient who has [2] Lee TH, Goldman L. Evaluation of the patient with
suspected ACS continues to be a challenge. Chest acute chest pain. N Engl J Med 2000;342:118795.
528 KONTOS & TATUM

[3] Wackers FJ, Lie KI, Liem KL, et al. Potential value [17] Kontos MC, Jesse RL, Anderson FP, et al. Compar-
of thallium-201 scintigraphy as a means of selecting ison of myocardial perfusion imaging and cardiac
patients for the coronary care unit. Br Heart J 1979; troponin I in patients admitted to the emergency
41:1117. department with chest pain. Circulation 1999;99:
[4] van der Wieken LR, Kan G, Belfer AJ, et al. 20738.
Thallium-201 scanning to decide CCU admission [18] Udelson JE, Beshansky JR, Ballin DS, et al. Myocar-
in patients with non-diagnostic electrocardiograms. dial perfusion imaging for evaluation and triage of
Int J Cardiol 1983;4:28599. patients with suspected acute cardiac ischemia: a ran-
[5] Henneman PL, Mena IG, Rothstein RJ, et al. Eval- domized controlled trial. JAMA 2002;288:2693700.
uation of patients with chest pain and nondiagnostic [19] Kontos MC, Kurdziel K, McQueen R, et al. Com-
ECG using thallium-201 myocardial planar imaging parison of 2-dimensional echocardiography and
and technetium-99m rst-pass radionuclide angiog- myocardial perfusion imaging for diagnosing myo-
raphy in the emergency department. Ann Emerg cardial infarction in emergency department patients.
Med 1992;21:54550. Am Heart J 2002;143:65967.
[6] Mace SE. Thallium myocardial scanning in the [20] Cohen M, Demers C, Gurnkel EP, et al. A compar-
emergency department evaluation of chest pain. ison of low-molecular-weight heparin with unfrac-
Am J Emerg Med 1989;7:3218. tionated heparin for unstable coronary artery
[7] Okada RD, Glover D, Ganey T, et al. Myocardial disease. N Engl J Med 1997;337:44752.
kinetics of technetium-99m-hexakis-2-methoxy-2- [21] Lindahl B, Venge P, Wallentin L. Relation between
methylpropyl-isonitrile. Circulation 1988;77:4918. troponin T and the risk of subsequent cardiac events
[8] Mannting F, Morgan-Mannting MG. Gated in unstable coronary artery disease. The FRISC
SPECT with technetium-99m-sestamibi for as- study group. Circulation 1996;93:16517.
sessment of myocardial perfusion abnormalities. [22] Alpert JS, Thygesen K, Bassand JP, et al. Myocar-
J Nucl Med 1993;34:6018. dial infarction redeneda consensus document of
[9] Bilodeau L, Theroux P, Gregoire J, et al. Techne- the Joint European Society of Cardiology/Amer-
tium-99m sestamibi tomography in patients with ican College of Cardiology Committee for the re-
spontaneous chest pain: correlations with clinical, denition of myocardial infarction. J Am Coll
electrocardiographic and angiographic ndings. Cardiol 2000;36:95969.
J Am Coll Cardiol 1991;18:168491. [23] OConnor MK, Hammell T, Gibbons RJ. In vitro
[10] Nicholson CS, Tatum JL, Jesse RL, et al. The value validation of a simple tomographic technique for es-
of gated tomographic Tc-99m-sestamibi perfusion timation of percentage myocardium at risk using
imaging in acute ischemic syndromes. J Nucl Cardiol methoxyisobutyl isonitrile technetium 99m (sesta-
1995;2:S57. mibi). Eur J Nucl Med 1990;17:6976.
[11] Kaul S, Senior R, Firschke C, et al. Incremental [24] Duca MD, Giri S, Wu AH, et al. Comparison of
value of cardiac imaging in patients presenting to acute rest myocardial perfusion imaging and serum
the emergency department with chest pain and with- markers of myocardial injury in patients with chest
out ST-segment elevation: a multicenter study. Am pain syndromes. J Nucl Cardiol 1999;6:5706.
Heart J 2004;148:12936. [25] Kontos MC, Fratkin MJ, Jesse RL, et al. Sensitivity
[12] Varetto T, Cantalupi D, Altieri A, et al. Emergency of acute rest myocardial perfusion imaging for iden-
room technetium-99m sestamibi imaging to rule out tifying patients with myocardial infarction based on
acute myocardial ischemic events in patients with a troponin denition. J Nucl Cardiol 2004;11:129.
nondiagnostic electrocardiograms. J Am Coll Car- [26] Hilton TC, Fulmer H, Abuan T, et al. Ninety-day
diol 1993;22:18048. follow-up of patients in the emergency department
[13] Hilton TC, Thompson RC, Williams HJ, et al. Tech- with chest pain who undergo initial single-photon
netium-99m sestamibi myocardial perfusion imag- emission computed tomography perfusion scintigra-
ing in the emergency room evaluation of chest phy with technetium 99m-labeled sestamibi. J Nucl
pain. J Am Coll Cardiol 1994;23:101622. Cardiol 1996;3:30811.
[14] Tatum JL, Jesse RL, Kontos MC, et al. Comprehen- [27] Miller TD, Christian TF, Hopfenspirger MR, et al.
sive strategy for the evaluation and triage of the Infarct size after acute myocardial infarction mea-
chest pain patient. Ann Emerg Med 1997;29:11623. sured by quantitative tomographic 99mTc sestamibi
[15] Kontos MC, Jesse RL, Schmidt KL, et al. Value of imaging predicts subsequent mortality. Circulation
acute rest sestamibi perfusion imaging for evaluation 1995;92:33441.
of patients admitted to the emergency department [28] Miller TD, Hodge DO, Sutton JM, et al. Useful-
with chest pain. J Am Coll Cardiol 1997;30:97682. ness of technetium-99m sestamibi infarct size in
[16] Heller GV, Stowers SA, Hendel RC, et al. Clinical predicting posthospital mortality following acute
value of acute rest technetium-99m tetrofosmin to- myocardial infarction. Am J Cardiol 1998;81:
mographic myocardial perfusion imaging in patients 14913.
with acute chest pain and nondiagnostic electrocar- [29] Reimer KA, Jennings RB, Cobb FR, et al. Animal
diograms. J Am Coll Cardiol 1998;31:10117. models for protecting ischemic myocardium: results
MPI IN CHEST PATIENTS 529

of the NHLBI Cooperative Study. Comparison of [41] Emre A, Ersek B, Gursurer M, et al. Angiographic
unconscious and conscious dog models. Circ Res and scintigraphic (perfusion and electrocardiogram-
1985;56:65165. gated SPECT) correlates of clinical presentation in
[30] De Coster PM, Wijns W, Cauwe F, et al. Area- unstable angina. Clin Cardiol 2000;23:495500.
at-risk determination by technetium-99m-hexakis- [42] Topol EJ, Weiss JL, Brinker JA, et al. Regional wall
2-methoxyisobutyl isonitrile in experimental re- motion improvement after coronary thrombolysis
perfused myocardial infarction. Circulation 1990; with recombinant tissue plasminogen activator: im-
82:215262. portance of coronary angioplasty. J Am Coll Car-
[31] Sinusas AJ, Trautman KA, Bergin JD, et al. Quanti- diol 1985;6:42633.
cation of area at risk during coronary occlusion [43] Heeschen C, van den Brand MJ, Hamm CW, et al.
and degree of myocardial salvage after reperfusion Angiographic ndings in patients with refractory
with technetium-99m methoxyisobutyl isonitrile. unstable angina according to troponin T status.
Circulation 1990;82:142437. Circulation 1999;100:150914.
[32] Christian TF, Behrenbeck T, Gersh BJ, et al. Rela- [44] Zhao X-Q, Theroux P, Snapinn SM, et al. Intracoro-
tion of left ventricular volume and function over nary thrombus and platelet glycoprotein IIb/IIIa re-
one year after acute myocardial infarction to infarct ceptor blockade with Tiroban in unstable angina or
size determined by technetium-99m sestamibi. Am J non-Q-wave myocardial infarction. Angiographic
Cardiol 1991;68:216. results from the PRISM-PLUS trail (platelet recep-
[33] Christian TF, Behrenbeck T, Pellikka PA, et al. tor inhibition for ischemic syndrome management
Mismatch of left ventricular function and infarct in patients limited by unstable signs and symptoms).
size demonstrated by technetium-99m isonitrile Circulation 1999;100:160915.
imaging after reperfusion therapy for acute myo- [45] Koch KC, vom Dahl J, Kleinhans E, et al. Inuence
cardial infarction: identication of myocardial of a platelet GPIIb/IIIa receptor antagonist on myo-
stunning and hyperkinesia. J Am Coll Cardiol cardial hypoperfusion during rotational atherec-
1990;16:16328. tomy as assessed by myocardial Tc-99m sestamibi
[34] Behrenbeck T, Pellikka PA, Huber KC, et al. Pri- scintigraphy. J Am Coll Cardiol 1999;33:9981004.
mary angioplasty in myocardial infarction: assess- [46] Fram DB, Azar RR, Ahlberg AW, et al. Duration of
ment of improved myocardial perfusion with abnormal SPECT myocardial perfusion imaging fol-
technetium-99m isonitrile. J Am Coll Cardiol 1991; lowing resolution of acute ischemia: an angioplasty
17:36572. model. J Am Coll Cardiol 2003;41:4529.
[35] Kontos MC, Kurdziel KA, Ornato JP, et al. A non- [47] Hollander JE, Homan RS, Gennis P, et al. Prospec-
ischemic electrocardiogram does not always predict tive multicenter evaluation of cocaine-associated
a small myocardial infarction: results with acute chest pain. Cocaine Associated Chest Pain
myocardial perfusion imaging. Am Heart J 2001; (COCHPA) Study Group. Acad Emerg Med 1994;
141:3606. 1:3309.
[36] Huey BL, Beller GA, Kaiser DL, et al. A compre- [48] Kontos MC, Schmidt KL, Nicholson CS, et al.
hensive analysis of myocardial infarction due to Myocardial perfusion imaging with technetium-
left circumex artery occlusion: comparison with 99m sestamibi in patients with cocaine-associated
infarction due to right coronary artery and left chest pain. Ann Emerg Med 1999;33:63945.
anterior descending artery occlusion. J Am Coll [49] Kontos MC, Schmidt KL, McCue M, et al. A com-
Cardiol 1988;12:115666. prehensive strategy for the evaluation and triage of
[37] Christian TF, Clements IP, Gibbons RJ. Noninva- the chest pain patient: a cost comparison study.
sive identication of myocardium at risk in patients J Nucl Cardiol 2003;10:28490.
with acute myocardial infarction and nondiagnostic [50] Radensky PW, Hilton TC, Fulmer H, et al. Potential
electrocardiograms with technetium-99m-Sestamibi. cost eectiveness of initial myocardial perfusion im-
Circulation 1991;83:161520. aging for assessment of emergency department pa-
[38] OKeefe JH Jr, Sayed-Taha K, Gibson W, et al. Do tients with chest pain. Am J Cardiol 1997;79:5959.
patients with left circumex coronary artery-related [51] Weissman IA, Dickinson CZ, Dworkin HJ, et al.
acute myocardial infarction without ST-segment Cost-eectiveness of myocardial perfusion imaging
elevation benet from reperfusion therapy? Am J with SPECT in the emergency department evalua-
Cardiol 1995;75:71820. tion of patients with unexplained chest pain. Radiol-
[39] Knott JC, Baldey AC, Grigg LE, et al. Impact of ogy 1996;199:3537.
acute chest pain Tc-99m sestamibi myocardial perfu- [52] McGuire DK, Hudson MP, East MA, et al. High-
sion imaging on clinical management. J Nucl Car- lights from the American Heart Association 72nd
diol 2002;9:25762. Scientic Sessions: November 6 to 10, 1999. Am
[40] de Winter RJ, Koster RW, Sturk A, et al. Value of Heart J 2000;139:35970.
myoglobin, troponin T, and CK-MBmass in ruling [53] Klocke FJ, Baird MG, Lorell BH, et al. ACC/AHA/
out an acute myocardial infarction in the emergency ASNC guidelines for the clinical use of cardiac ra-
room. Circulation 1995;92:34017. dionuclide imagingexecutive summary: a report of
530 KONTOS & TATUM

the American College of Cardiology/American [54] Wackers FJ, Brown K, Heller G, et al. American So-
Heart Association Task Force on Practice Guide- ciety of Nuclear Cardiology position statement on
lines (ACC/AHA/ASNC Committee to Revise the radionuclide imaging in patients with suspected
1995 Guidelines for the Clinical Use of Cardiac Ra- acute ischemic syndromes in the emergency depart-
dionuclide Imaging). J Am Coll Cardiol 2003;42: ment or chest pain center. J Nucl Cardiol 2002;9:
131833. 24650.
Cardiol Clin 23 (2005) 531539

Echocardiography in the Evaluation of Patients


in Chest Pain Units
William R. Lewis, MD
4860 Y Street, Suite 2820, Sacramento, CA 95864, USA

In experimental models, impairment of left six of these reports indicated a high positive pre-
ventricular segmental wall motion is apparent dictive value (PPV) for this technique, the overall
almost immediately after the onset of myocardial event rate in these studies was approximately
ischemia, preceding symptoms and ST-segment 51%. In these high-risk populations, the positive
alterations (Fig. 1). Using percutaneous angio- predictive value of a test for ischemia/infarction
plasty to induce the ischemic cascade in the car- will be augmented compared with that in typical
diac catheterization laboratory, echocardiographic low-risk, rule-out MI (ROMI) patients. Thus, in
wall motion abnormalities have been documented the study of Sabia and colleagues [4] in which the
to precede electrocardiographic abnormalities and cardiac event rate was 17%, the PPV of a resting
angina [1]. Therefore, detection of cardiac wall wall motion abnormality was only 31%. When
motion abnormalities is potentially more sensitive comparing resting ECG with resting echocardiog-
than the history, physical examination, and ECG raphy in a similarly low-risk population, Kontos
for identication of myocardial ischemia. Echo- and colleagues [5] demonstrated similar positive
cardiography is highly reliable for assessing cardiac and negative predictive values for predicting cardiac
wall motion and, thus, it has been used for diagno- events: 60% and 88% for ECG versus 44% and
sis and risk assessment in patients presenting to the 98% for echocardiography. In another report
emergency department (ED) with symptoms sug- with a similarly low cardiac event rate, echocardi-
gestive of myocardial ischemia. In patients who ography did not add to a 3-hour evaluation con-
have acute ST-elevation myocardial infarction sisting of clinical assessment and cardiac serum
(MI), echocardiography is comparable to invasive enzymes [12]. Furthermore, in all the aforemen-
left ventriculography for detecting wall motion ab- tioned investigations, the patients who had false-
normalities [2]. However, the usefulness of echocar- negative echocardiographic ndings had nonST
diography in the low-risk population that has chest elevation MI (NSTEMI) or unstable angina, re-
pain of uncertain origin and a nondiagnostic initial ecting the limited sensitivity of wall motion ab-
presentation is less well established. normalities for identication of this population.
Table 1 depicts the results of nine studies in pa- Unstable angina (with or without small troponin
tients presenting to the ED with chest pain in elevation) is the more common event in the low-
which resting echocardiography was used to pre- risk patients who have ischemic events.
dict cardiac events during the initial presentation Typical of a high-risk ST-elevation MI pop-
by detection of resting wall motion abnormalities ulation, Mohler and colleagues [11] studied 92
[311]. The events dened by the authors ranged patients who had a 60% cardiac event rate. The
from hard events, such as MI [4], to a combination echocardiogram was abnormal in 15 of the 18 pa-
of hard and soft end points, such as MI, revascu- tients who had MI, and in 12 of the 37 patients who
larization, angiographic coronary disease, and ab- had unstable angina. Five patients who had unsta-
normal perfusion imaging studies [11]. Although ble angina had wall motion abnormalities un-
changed from prior echocardiographic studies,
and they were therefore considered to be echocar-
E-mail address: william.lewis@ucdmc.ucdavis.edu diographically negative. Two of the three patients
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.009 cardiology.theclinics.com
532 LEWIS

Table 1
Predictive accuracy of echocardiography in patients
presenting with acute chest pain
PPV NPV
N Event Event (%) (%)
Kontos 130 RWA 15 29 34
et al [3]
RWA 6 80 93
Sabia 169 RWA 27 60 31
et al [4]
RWA 2 80 98
Kontos 260 RWA 41 53 44
et al [5]
Fig. 1. The ischemic cascade as seen in the cardiac cath- RWA 4 162 98
eterization laboratory during controlled coronary Korosoglou 98 RWA 19 2 90
occlusion. LV, left ventricle; LVEDP, left ventricular et al [6]
end-diastolic pressure; WMA, wall motion abnormality. RWA 18 59 77
Saeian 60 RWA 22 3 88
who had MI that were not detected by echocardi- et al [7]
ography had been given thrombolytic therapy. RWA 2 33 94
All patients who had a new wall motion abnormal- Sasaki 46 RWA 17 1 94
ity had an event, leading to a PPV of 100%. The et al [8]
negative predictive value was 57%, demonstrating RWA 6 22 79
Horowitz 65 RWA 34 2 94
that MIs in 43% of patients who had events were
et al [9]
not detected by echocardiography. These events RWA 2 27 93
were predominantly unstable angina diagnosed Peels 35 RWA 22 4 85
by typical chest pain lasting more than 30 minutes et al [10]
without ECG or enzymatic changes. RWA 3 14 82
All of these studies were performed in the Mohler 92 RWA 27 0 100
1980s and 1990s using older echocardiographic et al [11]
machinery. For example, the study by Kontos and RWA 28 37 57
colleagues [11] used Hewlett-Packard 500, 1000, Abbreviations: N, number of patients; RWA, re-
and 1500 machines (Hewlett-Packard, Palo Alto, gional wall motion abnormality present; RWA, no re-
California). Newer equipment with improved im- gional wall motion abnormality; Event, Coronary
aging techniques could improve wall motion im- event present; Event, no coronary event; PPV, positive
aging and improve the predictive value of resting predictive value; NPV, negative predictive value.
echocardiography for detecting ischemia. Funda- Adapted from Lewis W. Evaluation of the patient
with rule out myocardial infarction. Arch Intern
mental echocardiography receives echocardio-
Med 1996;156:415; with permission.
graphic signals at the same frequency at which
they were transmitted, usually in the 2- to 3-
MHz range for adult imaging. Fundamental echo- Hickman and colleagues [13] studied 80 pa-
cardiography has diculty imaging structures tients presenting to the ED with chest pain of
nearest the transducer because of artifact gener- suspected cardiac origin that lasted at least 30
ated by the chest wall. Artifacts can also be gener- minutes within the previous 6 hours. The popula-
ated from the side of the transducer, causing side tion studied was typical of the ROMI population
lobes. As echocardiographic sound waves travel in that 13 of the 80 patients (16%) were subse-
through the heart, the tissues start to oscillate quently diagnosed as having had an event. An
and can produce harmonics. Tissue harmonic HDI 3000 (ATL, Bothell, Washington) machine
echocardiography sends signals at the 2- to 3- with tissue harmonics was used for echocardio-
MHz fundamental, but uses the 4- to 6-MHz graphic imaging. Tissue harmonic imaging was
harmonic signals received for imaging. This tech- signicantly better than fundamental imaging
nique reduces the near-eld artifact, as near-eld and allowed for adequate visualization of all seg-
signals dont usually produce harmonic signals. ments in all patients, compared with fundamental
Harmonic signals also travel in a narrower imaging in which 11% of segments could not be
beam, reducing the side lobe artifact. visualized (Fig. 2). The positive and negative
ECHOCARDIOGRAPHY IN CHEST PAIN UNITS 533

Fig. 2. Bar graph comparing quality of segmental visu-


Fig. 3. Bar graph showing incremental value of clinical,
alization with echocardiography using fundamental (FE)
ECG, and echocardiographic (THE) variables for the
and tissue harmonics (THE) imaging. (Adapted from
prediction index acute MI or a subsequent cardiac event.
Hickman M. Wall thickening assessment with tissue har-
(Adapted from Hickman M. Wall thickening assessment
monic echocardiography results in improved risk strati-
with tissue harmonic echocardiography results in im-
cation for patients with non-ST-segment elevation
proved risk stratication for patients with non-ST-
acute chest pain. Eur J Echocardiogr 2004;5(2):1428;
segment elevation acute chest pain. Eur J Echocardiogr
with permission.)
2004;5(2):1428; with permission.)
predictive value of fundamental echocardiography ROMI population who are at risk for a cardiac
for an event (elevated creatine kinase-myocardial event. The examination is highly dependent on
band [CK-MB], coronary revascularization, and skilled, experienced personnel for adequate data
cardiac death) was 21% and 90%, respectively. acquisition and reliable interpretation. According
Using tissue harmonic echocardiography, the pos- to the Task Force on Echocardiography in
itive and negative predictive values improved to Emergency Medicine of the American Society of
26% and 97%, respectively. This predictive accu- Echocardiography and Technology and Practice
racy was nearly identical to electrocardiography Executive Committees of the American College of
with a positive and negative predictive value of Cardiology, a minimum of 6 months of training
24% and 97%, respectively (Table 2). However, and performance of 300 echocardiographic stud-
tissue harmonic echocardiography provided ies are required to become a competent reader
added ability to predict future cardiac events [14]. Therefore, highly specialized physicians,
when compared with clinical history and the elec- technical personnel, and equipment are required
trocardiogram (Fig. 3). This added benet was not for dedicated use in the ED. Furthermore, al-
seen using fundamental echocardiography. though ventricular wall motion abnormalities as-
There are several factors that limit the value of sociated with ischemia may persist for prolonged
echocardiography for detecting patients in the periods, they may also resolve in as little as 2
hours [15]. Studies in animals have demonstrated
Table 2
Accuracy of clinical, ECG and echocardiographic varia- that wall motion abnormalities may not be de-
bles for prediction of myocardial infarction as the pre- tected when infarction involves less than 20% of
senting event ventricular wall thickness [16] or less than 12%
of left ventricular circumference [17]. Therefore,
PPV (%) NPV (%) P
patients who have NSTEMI and unstable angina
Diabetes 5 80 .09 may have no discernible abnormalities of wall mo-
Abnormal ECG 24 97 O.1
tion. Echocardiography provides no information
Abnormal fundamental 21 90 O.1
on the age of a wall motion abnormality, reducing
ECG
Abnormal tissue 26 97 .007 its usefulness in patients who have known CAD or
harmonic ECG other cardiac disease. Finally, wall motion abnor-
malities may be seen in patients who have left
Abbreviations: PPV, positive predictive value. NPV,
bundle branch block and right ventricular volume
negative predictive value. P, probability.
Adapted from Hickman M. Wall thickening assess- or pressure overload that complicate interpre-
ment with tissue harmonic echocardiography results in tation of the echocardiogram for ischemic wall
improved risk stratication for patients with non-ST- motion abnormalities.
segment elevation acute chest pain. Eur J Echocardiogr Newer echocardiographic approaches may
2004;5(2):1428; with permission. overcome some of these limitations. A technique
534 LEWIS

under investigation is analysis of variation in the related to blood volume in the myocardium and
scatter of acoustic signals during ischemia (ultra- can be normalized to the ventricular cavity.
sonic tissue characterization). Acute myocardial Variability in steady state opacication is related
ischemia results in increased integrated backscat- to myocardial viability. Delay in regional myocar-
ter of ultrasound signals and decreased variability dial reperfusion is related to coronary ow rates
of scatter during systole and diastole. When and coronary disease (Fig. 4) [19].
compared with wall motion analysis for the Korosoglou and colleagues, [6] studied myocar-
detection of MI in the coronary care unit, Saeian dial contrast echocardiography in 100 patients
and colleagues [7] found ultrasonic tissue charac- presenting to the ED with chest pain. Of these
terization had comparable sensitivity, specicity, patients, 37 were ultimately given the diagnosis
and accuracy. Ultrasonic tissue characterization of acute coronary syndrome, resulting in a high
had diculty detecting apical infarcts, whereas event rate of 37%; 12 were diagnosed with unsta-
wall motion analysis had diculty detecting ische- ble angina by Braunwald criteria and found to
mia in the presence of left bundle branch block. It have greater than 70% stenosis in a coronary ar-
is possible that a combination of the two tech- tery at arteriography; and 25 were diagnosed as
niques would be complementary. having NSTEMI with troponin T elevation. Two
A second strategy that may be useful for patients were excluded from analysis because of
identication of myocardial ischemia is contrast poor echocardiographic images. Of the remaining
echocardiography [18]. In standard contrast imag- 61 patients diagnosed as having noncardiac chest
ing, the ultrasonic properties of the myocardium pain, 21 underwent angiography that showed no
and ventricular cavity can be enhanced by intrave- high-grade (R75%) lesions and 40 patients under-
nous injection of myocardial contrast agents. The went stress testing that resulted in normal ndings.
current agents used are microbubbles formed by As shown in Table 3, the presence of resting echo-
encapsulated gases that are small enough to tra- cardiographic wall motion abnormalities had
verse the pulmonary capillary bed, access the sys- a 90% PPV and a 77% negative predictive value
temic circulation, and perfuse not only the left (NPV) for acute coronary syndrome. Myocardial
ventricular cavity but also the myocardium. contrast echocardiography had an 89% PPV and
When echocardiographic ultrasound strikes these a 91% NPV for acute coronary event. Using rest-
microbubbles, the bubbles vibrate in a nonlinear ing wall motion abnormalities plus myocardial
fashion and give o enhanced echocardiographic contrast echocardiography yielded an 89% PPV
signals. When using these agents to assess myocar- and a 93% NPV. Because troponin assays can be
dial perfusion, a pulse of high-intensity ultra- rapidly performed in the ED, the group of patients
sound is given that is strong enough to burst the that remain a diagnostic dilemma are those who
bubbles in the myocardium. The myocardium is have unstable angina without troponin elevation
visualized over the next 10 to 15 cycles and ana- or electrocardiographic abnormalities. Focusing
lyzed for wash-in of myocardial contrast agent on the group diagnosed as having unstable angina
(Table 3). Normally, the myocardium reperfuses without troponin elevation in this current study,
within ve to ten cardiac cycles. The intensity of the sensitivity of resting wall motion abnormality
the signal in the myocardium at steady state is was 17% and the sensitivity of myocardial contrast

Table 3
Detection of unstable angina and nonST-elevation myocardial infarction by myocardial contrast echocardiography
Visual assessment Sensitivity Specicity PPV NPV Accuracy
Wall motion (2D-echo) 51% 97% 90% 77% 80%
Perfusion (MCE) 84% 93% 89% 91% 90%
Perfusion (MCE), single-vessel CAD 71% 93% 71% 93% 89%
Perfusion (MCE), multivessel CAD 92% 93% 85% 93% 93%
Wall motion and perfusion 89% 93% 89% 93% 92%
Abbreviations: 2D-Echo, two-dimensional echocardiography; CAD, coronary artery disease; MCE, myocardial con-
trast echocardiography; NPV, negative predictive value; PPV, positive predictive value; P ! .05, myocardial perfusion
versus wall motion.
From Korosoglou G. Usefulness of real-time myocardial perfusion imaging in the evaluation of patients with rst
time chest pain. Am J Cardiol 2004;94(10):122531; with permission.
ECHOCARDIOGRAPHY IN CHEST PAIN UNITS 535

Fig. 5. Percent incremental value of tests performed in


hierarchical order after clinical, demographic, and elec-
Fig. 4. Method of assessing mean blood ow velocity trocardiographic variables are known. Values were sig-
and myocardial blood volume using myocardial contrast nicantly higher (*) (P!.05) compared with other CE
echocardiography. (From Lepper W, Belcik T, Wei K, variables, and signicantly lower (**) compared with
et al. Myocardial contrast echocardiography. Circula- other SPECT variables. E-RF, echocardiographic re-
tion 2004;109:31325; with permission.) gional function; E-PER, echocardiographic perfusion;
E-COM, combination of echocardiographic perfusion
and function; S-RF, regional function by SPECT;
echocardiography was 66%. This means that one S-PER, perfusion by SPECT; S-COM, combination
of three patients who has unstable angina without of perfusion and function by SPECT. (From Kaul S.
troponin elevation will go undetected using Incremental value of cardiac imaging in patients with
myocardial contrast echocardiography, compared chest pain and without ST-segment elevation: a multi-
with ve out of six patients undetected using only center study. Am Heart J 2004;148(1):12936; with
resting wall motion as the detection criterion. permission.)
From the above data, it is clear that myocardial
contrast echocardiography yields greater diagnos- yield further diagnostic information regarding the
tic accuracy compared with wall motion analysis in presence of signicant coronary disease that is not
the detection of acute coronary syndromes. Kaul producing active ischemia at the time of examina-
and colleagues [20] compared myocardial contrast tion. Geleijnse and colleagues [21] studied dobut-
echocardiography to resting single photon emis- amine stress echocardiography in 102 patients
sion computed tomography (SPECT) in 203 presenting to the ED with chest pain of suspected
patients. All patients underwent myocardial con- cardiac origin. Of these patients, 59 had known
trast echocardiography, but only 163 patients un- coronary artery disease and 43 did not. The pa-
derwent SPECT. Thirty-eight patients (19%) had tients who did not have known coronary artery
an event, consistent with this being a moderate disease underwent stress testing after a minimum
risk group for a coronary event. When comparing 12-hour observation period, whereas patients
myocardial contrast echocardiography to SPECT, who had known coronary artery disease underwent
myocardial contrast echocardiography perfusion a minimum 24-hour observation period. As a re-
was superior to SPECT myocardial perfusion in sult, 13 patients were diagnosed as having events
predicting events. However, SPECT myocardial re- based on abnormal resting electrocardiograms
gional function was superior to echocardiographic and serum cardiac enzymes, and 89 patients were
regional function in predicting events, resulting in then referred for dobutamine stress testing. Pa-
SPECT having a greater overall predictive value tients who had resting wall motion abnormalities
compared with myocardial contrast echocardiog- were not excluded from stress testing. Dobutamine
raphy. The incremental value of tests performed was administered in 10 mg/kg/min increments at
in hierarchical order after clinical, demographic, 3-minute intervals up to a maximum of 40 mg/kg/
and electrocardiographic variables are known is min. Atropine, up to 1 mg, was administered to pa-
shown in Fig. 5. tients failing to reach 85% maximal predicted heart
All of the studies reviewed to this point used rate with dobutamine infusion alone. The test was
resting echocardiographic studies. As with the terminated for the usual end points of attaining
ECG, resting echocardiographic studies provide target heart rate, stress-induced wall motion
information regarding only the presence of resting abnormalities, electrocardiographic changes, ar-
ischemia within a narrow time frame. When the rhythmias, severe chest pain, hypertension, or in-
resting study is normal, then a stress study may tolerable side eects. The echocardiograms were
536 LEWIS

interpreted using the standard 16-segment model echocardiography using 3-minute stages and
[22]. Three patients had poor acoustic windows dobutamine at 10 mg/kg/min increments to a maxi-
precluding stress testing, and six patients failed mum of 40 mg/kg/min. Atropine, 1 mg, was given
to reach 85% maximal predicted heart rate. Of for patients failing to reach 85% maximal pre-
the 80 remaining patients, 44 had normal stress dicted heart rate. The test was terminated for the
echocardiograms and 36 had stress-induced wall usual end points. This testing determined that 23
motion abnormalities. Results of the stress test patients had poor acoustic windows precluding
were conveyed to the attending physician and stress testing, 4 patients had resting wall motion ab-
could be used to manage the patient. Seven pa- normalities and did not undergo stress testing, 377
tients who had positive stress echocardiograms un- patients completed the dobutamine stress protocol,
derwent coronary revascularization during the and 26 patients had stress-induced wall motion ab-
index event, and seven patients had unstable angi- normalities in at least one segment of a 16-segment
na during the 6-month follow-up period, with model. The attending physicians were blinded to
three of these patients having coronary revascular- the results of the dobutamine stress tests and pa-
ization procedures. Two patients who had normal tient management was determined by use of other
stress studies had events (one MI and one unstable diagnostic protocols. Over the 6-month follow-up
angina). The positive and negative predictive accu- period, eight patients who had positive dobut-
racy for dobutamine stress echocardiography was amine stress echocardiography had events, includ-
25% and 95.5% (Table 4). Seven patients had sig- ing one cardiac death (Table 5). Of the 351 patients
nicant dobutamine stress testingrelated adverse who had negative dobutamine stress echocardiog-
events. Two patients had stress-induced nonsus- raphy studies, 14 patients had events during the
tained ventricular tachycardia, four patients had follow-up period, including one cardiac death (see
stress-induced supraventricular tachycardia, and Table 5). The PPV for stress-induced wall motion
one patient had stress-induced atrial utter. abnormalities was 31% and the NPV was 96%
In the largest study to date, Bholasingh and (see Table 4). In 75 (19.9%) patients, the protocol
colleagues [23] studied 577 patients admitted to the was not completed because of the prespecied
ED who had suspected acute coronary disease. criteria: extensive new wall motion abnormality
During a 12-hour observation period, 119 patients (1 patient), electrocardiographic changes (7 pa-
were diagnosed with acute coronary syndrome and tients), chest pain (28 patients), and intolerable
34 were given other diagnoses, such as aortic steno-
sis or atrial brillation. Following this period, 404
patients were then referred for dobutamine stress Table 5
Clinical outcomes at 6 months according to dobutamine
stress echocardiography results
Positive Negative
Table 4
DSE (n 26) DSE (n 351)
Predictive accuracy of dobutamine stress echocardiogra-
phy in patients presenting with acute chest pain and Cardiac death 1 1
good quality images Nonfatal AMI 2 0
Rehosp UA 2 6
PPV NPV Primary end point 5 7
N Event Event (%) (%)
PTCA 2 4
Geleijnse 80 RWA 9 27 25 CABG 1 3
et al [21] Combined end point 8 14
RWA 2 42 95.5
Abbreviations: AMI, acute myocardial infarction;
Bholasingh 377 RWA 8 18 31
CABG, coronary artery bypass grafting; DSE, dobut-
et al [23]
amine stress echocardiography; PTCA, percutaneous
RWA 14 337 96
transluminal coronary angioplasty; Rehosp-UA, rehospi-
Trippi 137 RWA 7 0 100
talization for unstable angina. Primary end point, cardiac
et al [24]
death, AMI rehosp-UA. Combined end point, primary
RWA 2 128 98.5
end point PTCA CABG.
Abbreviations: N, number of patients; RWA, re- From Bholasingh R, Cornel JH, Kamp O, et al.
gional wall motion abnormality present; RWA, no re- Prognostic value of predischarge obutamine stress echo-
gional wall motion abnormality; Event, coronary event cardiography in chest pain patients with a negative car-
present; Event, no coronary event; PPV, positive pre- diac troponin T. J Am Coll Cardiol 2003;41(4):596602;
dictive value; NPV, negative predictive value. with permission.
ECHOCARDIOGRAPHY IN CHEST PAIN UNITS 537

side eects (39 patients) such as arrhythmia and testing. Caution is therefore warranted when us-
severe hypertension or hypotension. ing dobutamine stress echocardiography in evalu-
Stress echocardiography requires experienced ating these patients, and none of the published
personnel to perform and interpret the test results. guidelines yet include dobutamine stress testing
Cardiologists may not be available at o hours for as part of the evaluation of patients who have
the performance or interpretation of dobutamine chest pain in the ED.
stress testing. To circumvent these limitations, The NPV is uniformly good when comparing
Trippi and colleagues [24] developed a protocol the three dobutamine stress studies, whereas the
in which dobutamine stress echocardiography PPV varies greatly. The low PPV seen in the study
was performed by trained nurses and sonography by Geleijnse and colleagues [21] may be caused by
technicians and then transmitted through teleme- the inclusion of a large number of patients who
dicine for interpretation by cardiologists. The have pre-existing disease and resting wall motion
authors studied 163 patients admitted to the hos- abnormalities. A true positive study was also de-
pital for chest pain with negative cardiac enzymes ned as one in which the patient had a hard event.
and electrocardiogram. Overall, 57% of the stud- One could argue that in patients who have known
ies were performed after regular hours or on week- coronary disease, a positive stress echocardiogram
ends. Of this population, 26 patients had resting is by denition a true positive. This assertion
wall motion abnormalities and were excluded would result in a greatly increased PPV in this
from stress testing. Dobutamine was administered study. In the study by Bholasingh and colleagues
in 3-minute stages of 5, 10, 20, 30, 40, and 50 mg/ [23], the attending physicians were blinded to the
kg/min infusions. Stress testing was terminated results of the dobutamine stress test and therefore
prematurely in 9 patients because of chest pain not all patients underwent coronary arteriogra-
or arrhythmias. The symptoms resolved in less phy. Again, a true positive test was dened as
than 10 minutes and required no specic therapy. one with a hard event in the follow-up period.
There were no major complications of dobut- The true predictive value may be much greater
amine stress testing. The study showed that 7 and the other diagnostic modalities, which were
patients had stress-induced wall motion abnor- unreported by the authors, may not have detected
malities and were found to have greater than the disease in these patients. In the study by Trippi
50% coronary stenoses on arteriography (100% and colleagues [24], all patients who had abnor-
PPV), 130 patients had normal dobutamine stress mal stress echocardiograms underwent coronary
tests, and 2 patients were subsequently found to arteriography and the true PPV was known and
have coronary disease during the 3-month fol- found to be 100% for the detection of more
low-up period (one inferior wall MI and one un- than 50% coronary stenoses.
stable angina, 98.5% NPV). Two patients were In summary, resting echocardiography may
not described as either positive or negative by lack sucient sensitivity to be of clinical useful-
the authors. In patients who had normal resting ness in the patient who has low-risk chest pain in
echocardiograms, overall PPV and NPV for com- the ED, and the role of resting echocardiography
bined rest and subsequent dobutamine stress echo- is debatable. The National Heart Attack Alert
cardiography was 51.5% and 98.5%, respectively Program Working Group concluded that false
(see Table 4). The authors have subsequently re- negative rates in the prospective studies are too
ported over 4 years of experience in 734 patients high to be safe [28]. In addition, the American
being managed by this protocol [25]. In this larger College of Cardiology/American Heart Associa-
group of patients, the test was discontinued pre- tion 2002 Guideline Update for the Management
maturely in only 3.1% of patients, 12.5% of pa- of Patients with Unstable Angina and Non-ST-
tients had abnormal stress echocardiograms, and Segment Elevation Myocardial Infarction does
70% of patients were discharged after negative not include echocardiography in the initial evalu-
stress echocardiograms. ation and management of patients [29]. In con-
Although none of the patients suered an MI trast, another viewpoint is oered by the
with dobutamine stress testing in these papers, MI American College of Cardiology/American Heart
has been reported with dobutamine stress echo- Association/American Society of Echocardiogra-
cardiography in outpatients [26]. Dobutamine has phy that states that early echocardiography is
also been shown to tremendously activate plate- particularly useful in patients with a high clinical
lets, independent of circulating catecholamines suspicion of acute myocardial infarction but
[27]. Platelet activation was not seen with exercise a non diagnostic ECG [30].
538 LEWIS

One of the aims in evaluating patients in the [8] Sasaki H, Charuzi Y, Beeder C, et al. Utility of echo-
ED who have suspected acute coronary syndrome cardiography for the early assessment of patients
is to identify patients who are at low enough risk with nondiagnostic chest pain. Am Heart J 1986;
to be discharged home and those who are at high 112:4947.
[9] Horowitz RS, Morganroth J, Parrotto C, et al. Im-
risk and require admission. Inclusion of myocar-
mediate diagnosis of acute myocardial infarction
dial contrast echocardiography or stress echo- by two-dimensional echocardiography. Circulation
cardiography can signicantly improve the 1982;65:3239.
diagnostic accuracy of echocardiography in the [10] Peels CH, Visser CA, Kupper AJ, et al. Value of two
detection of ischemia. Therefore, a rational use of dimensional echocardiography for immediate detec-
echocardiography in this schema would be to tion of myocardial ischemia in the emergency room.
evaluate patients for resting wall motion abnor- Am J Cardiol 1990;65:68791.
malities in whom serial electrocardiograms and [11] Mohler ER, Ryan T, Segar DS, et al. Clinical utility
serum cardiac markers were negative. Patients of troponin T levels and echocardiography in the
who have resting wall motion abnormalities could emergency department. Am Heart J 1998;135
(2 Pt 1):25360.
be admitted to the hospital for further evaluation.
[12] Levitt MA, Promes SB, Bullock S, et al. Combined
Patients who do not have resting wall motion cardiac marker approach with adjunct two-dimen-
abnormalities could be scheduled for exercise or sional echocardiography to diagnose acute myocar-
dobutamine stress testing, myocardial perfusion dial infarction in the emergency department. Ann
imaging, or other diagnostic modalities as part of Emerg Med 1996;27:17.
their accelerated diagnostic protocol. [13] Hickman M, Swinburn JM, Senior R. Wall
thickening assessment with tissue harmonic echo-
References cardiography results in improved risk stratica-
tion for patients with non-ST-segment elevation
[1] Hauser AM, Gangadharan V, Ramos RG, et al. acute chest pain. Eur J Echocardiogr 2004;5(2):
Sequence of mechanical, electrocardiographic and 1428.
clinical eects of repeated coronary artery occlusion [14] Stewart WJ, Douglas PS, Sagar K, et al. Echocardi-
in human beings: echocardiographic observations ography in emergency medicine: a policy statement
during coronary angioplasty. J Am Coll Cardiol by the American Society of Echocardiography and
1985;5:1937. the American College of Cardiology. J Am Soc
[2] Lundgren C, Bourdillon PD, Dillon JC, et al. Com- Echocardiogr 1999;12:824.
parison of contrast angiography and two-dimen- [15] Jeroudi MO, Cheirif J, Habib G, et al. Prolonged
sional echocardiography for the evaluation of left wall motion abnormalities after chest pain at rest
ventricular regional wall motion abnormalities after in patients with unstable angina: a possible manifes-
acute myocardial infarction. Am J Cardiol 1990;65: tation of myocardial stunning. Am Heart J 1994;127:
10717. 124150.
[3] Kontos MC, Arrowood JA, Jesse RL, et al. Com- [16] Lieberman AN, Weiss JL, Jugdutt BL, et al. Two-
parison between 2-dimensional echocardiography dimensional echocardiography and infarct size: re-
and myocardial imaging in the emergency depart- lationship of regional wall motion and thickening
ment in patients with possible myocardial ischemia. to the extent of myocardial infarction in the dog.
Am Heart J 1998;136(4 Pt 1):72433. Circulation 1981;63:739.
[4] Sabia P, Afrookteh A, Touchstone DA, et al. Value [17] Kaul S. Echocardiography in coronary artery dis-
of regional wall motion abnormality in the emer- ease. Curr Probl Cardiol 1990;15:233.
gency room diagnosis of acute myocardial infarc- [18] Sakuma T, Hayashi Y, Shimohara A, et al. Useful-
tion. A prospective study using two-dimensional ness of myocardial contrast echocardiography for
echocardiography. Circulation 1991;84:I8592. the assessment of serial changes in risk area in
[5] Kontos MC, Arrowood JA, Paulsen WH, et al. patients with acute myocardial infarction. Am J
Early echocardiography can predict cardiac events Cardiol 1996;78:12737.
in emergency department patients with chest pain. [19] Lepper W, Belcik T, Wei K, et al. Myocardial
Ann Emerg Med 1998;31(5):5507. contrast echocardiography. Circulation 2004;109:
[6] Korosoglou G, Labadze N, Hansen A, et al. Useful- 31325.
ness of real-time myocardial perfusion imaging in [20] Kaul S, Senior R, Firschke C, et al. Incremental
the evaluation of patients with rst time chest pain. value of cardiac imaging in patients with chest pain
Am J Cardiol 2004;94(10):122531. and without ST-segment elevation: a multicenter
[7] Saeian K, Rhyne TL, Sagar KB. Ultrasonic tissue study. Am Heart J 2004;148(1):12936.
characterization for diagnosis of acute myocardial [21] Geleijnse ML, Elhendy A, Kasprzak JD, et al.
infarction in the coronary care unit. Am J Cardiol Safety and prognostic value of early dobutamine-
1994;74:12115. atropine stress echocardiography in patients with
ECHOCARDIOGRAPHY IN CHEST PAIN UNITS 539

spontaneous chest pain and a non-diagnostic [27] Galloway MT, Paglieroni TG, Wun T, et al. Platelet
electrocardiogram. Eur Heart J 2000;21(5):3445. activation during dobutamine stress echocardiogra-
[22] Schiller NB, Shah PM, Crawford M, et al. Recom- phy. Am Heart J 1998;135:888900.
mendations for quantitation of the left ventricle by [28] Selker HP, Zalenski RJ, Antman EM, et al. An eval-
two-dimensional echocardiography: American Soci- uation of technologies for identifying acute cardiac
ety of Echocardiography Committee on Standards, ischemia in the emergency department: a report
Subcommittee on Quantitation of Two-Dimen- from a National Heart Attack Alert Program Work-
sional Echocardiograms. J Am Soc Echocardiogr ing Group. Ann Emerg Med 1997;29:1387.
1989;2:35867. [29] ACC/AHA 2002 Guideline Update for the Manage-
[23] Bholasingh R, Cornel JH, Kamp O, et al. Prognostic ment of Patients with Unstable Angina and Non-St-
value of predischarge dobutamine stress echocardi- Segment Elevation Myocardial Infarction. A report
ography in chest pain patients with a negative car- of the American College Of Cardiology/American
diac troponin T. J Am Coll Cardiol 2003;41(4): Heart Association Task Force on Practice Guidelines
596602. (Committee on the Management of Patients with
[24] Trippi JA, Lee KS, Kopp G, et al. Dobutamine Unstable Angina). Available at www.acc.org/
stress tele-echocardiography for the evaluation of clinical/guidelines/unstable/incorporated/index.htm.
emergency department patients with chest pain. Accessed September 15, 2005.
J Am Coll Cardiol 1997;30(3):62732. [30] Cheitlin MD, Armstrong WF, Aurigemma GP, et al.
[25] Trippi JA, Lee KS. Dobutamine stress tele-echocar- ACC/AHA/ASE 2003 Guideline Update for the
diography as a service in the emergency department Clinical Application of Echocardiography. A report
to evaluate patients with chest pain. Echocardiogra- of the American College of Cardiology/American
phy 1999;16(2):17985. Heart Association Task Force on Practice Guide-
[26] Lewis WR, Arena FJ, Galloway MT, et al. Acute lines (ACC/AHA/ASE Committee to Update the
myocardial infarction associated with dobutamine 1997 Guidelines for the Clinical Application of
stress echocardiography. J Am Soc Echocardiogr Echocardiography). J Am Coll Cardiol 2003;42:
1997;10(5):5768. 95470.
Cardiol Clin 23 (2005) 541548

Newer Imaging Methods for Triaging Patients


Presenting to the Emergency Department
with Chest Pain
James McCord, MDa,*, Ezra A. Amsterdam, MDb
a
Henry Ford Health System, Heart & Vascular Institute, 2799 West Grand Boulevard,
K-14, Detroit, MI 48202-2689, USA
b
Department of Internal Medicine, University of California School of Medicine (Davis)
and Medical Center, 4860 Y Street, Suite 200, Sacramento, CA 95817, USA

Electron beam computed tomography (EBCT), that can assess coronary calcium, and studies have
also known as ultrafast CT, is highly sensitive for shown that this technology correlates well with
detecting and quantifying coronary artery cal- EBCT for the detection of coronary calcium
cium. Coronary artery calcium is absent in the [17,18]. Many institutions employ MSCT for cor-
normal vessel and occurs almost exclusively in onary calcium scoring as this CT modality can
the arteries with atherosclerosis [13]. The degree also be used for scanning other organs, and
of coronary calcium is strongly associated with EBCT is only useful for coronary calcium quanti-
the total atherosclerotic burden [46]. Patients cation. However, most clinical studies have em-
who have more coronary artery calcium are more ployed EBCT for cardiac risk stratication. The
likely to have signicant angiographic obstruc- cost of an EBCT is approximately $400.
tions and, more importantly, are known to suer Histologic studies have shown that tissue
more adverse cardiac events [713]. These ndings, densities greater than 130 Hounseld units (a mea-
and the procedures noninvasive methodology and surement that characterizes the relative density of
low cost, have provided the rationale for the use of a substance) are associated with calcied plaque. A
EBCT for risk stratication of patients in the coronary calcium score, also called the Agatston
emergency department (ED) who have possible score, is calculated as a product of the area of
acute coronary syndrome (ACS) [1416]. calcication and a factor based on the maximal
calcium density. A composite score to indicate the
quantity of calcium detected in the entire coronary
Technology considerations
artery system is typically used: a score of 0 is
EBCT involves an electron emitter rather than considered normal, 1 to 99 is mild, 100 to 400 is
a standard radiograph that allows for rapid moderate, and more than 400 is severe. Coronary
screening times. Transaxial images are obtained calcium scores have also been age- and sex-adjusted
at 100 ms with 3-mm to 6-mm slices during one to as coronary calcium is more extensive in older and
two breath holds. Scanning is triggered near end- male patients [19]. A more recently developed calci-
diastole to minimize motion artifact. No intrave- um volume score may oer more precision as com-
nous or oral contrast is required and the entire pared with the traditional Agatston score [20,21].
scan can be completed in 10 to 15 minutes.
Multislice CT (MSCT) is another CT modality Coronary calcium score and cardiac risk
assessment

* Corresponding author. Numerous studies have shown that that the


E-mail address: jmcord1@hfhs.org (J. McCord). coronary calcium score has prognostic value
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.010 cardiology.theclinics.com
542 MCCORD & AMSTERDAM

signicance in symptomatic and asymptomatic McLaughlin and colleagues [14] evaluated 181 pa-
individuals [2231]. A study of 5365 asymptom- tients who were admitted to the hospital from the
atic individuals demonstrated 224 adverse events ED with chest pain of presumed cardiac origin.
(death, nonfatal myocardial infarction (MI), coro- Patients were included if they had a normal or
nary bypass surgery, or percutaneous interven- nondiagnostic ECG at presentation. Exclusion cri-
tion) over 3 years [23]. Individuals who have teria were history of coronary artery disease (CAD)
a score of more than zero had a signicantly (MI, coronary bypass surgery, or percutaneous
higher event rate compared with those who do intervention), diagnostic ST changes or Q waves
not have calcium (6.1% versus 0.4% in men; on electrocardiogram, weight more than 250 lb, ini-
3.3% versus 1.0% in women). The magnitude of tial elevated creatine kinase-myocardial band (CK-
cardiac risk is directly related to the extent of cal- MB), and known or suspected pregnancy. Among
cication [24]. However, controversy has arisen those evaluated, 44 patients did not meet entry cri-
concerning the additional prognostic signicance teria and 4 patients did not give consent, yielding
of coronary calcium after accounting for tradi- 133 patients in the nal analysis. An EBCT was
tional risk factors. In 2000, the American College considered positive if the score was more than
of Cardiology/American Heart Association pub- one; the results of the scan were blinded to the re-
lished a consensus document concerning EBCT sponsible clinician.
that states, Importantly, the incremental value Patients had 30-day follow-up by phone call
of EBCT over traditional multivariate risk assess- for adverse events such as MI (as dened by
ment models has yet to be established [32]. World Health Organization criteria), coronary
Some early studies suggest that coronary artery bypass surgery, percutaneous intervention,
calcium scores do not oer additional prognostic or sudden cardiac death. The population was 71%
information after accounting for the Framingham African-American, Hispanic 19%, 19% White,
risk score [9], which predicts cardiac risk based on and 1% Asian, and comprised 84 (63%) women
age; sex; low-density lipoprotein and high-density and 50 (37%) men. The mean age was 53 G 2
lipoprotein cholesterol; hypertension; cigarette years. There were 86 (64%) patients who had
smoking; and diabetes mellitus [33]. The consen- a positive scan and 48 (36%) patients who had
sus document published in 2000 was written after a negative scan. In the entire group there were
a literature review through 1998. However, since seven (5%) adverse cardiac events, including four
that time publications have shown that coronary MIs, two coronary artery bypass surgeries, and
calcium scores give additional independent prog- one percutaneous intervention. Among the 48
nostic information. A recent meta-analysis of patients who had normal coronary calcium scores,
four studies demonstrated that an elevated coro- there was one (2%) adverse event. This patient
nary calcium score was associated with an in- was a 45-year-old man who suered a MI after
creased risk for adverse cardiac events after cocaine use. Of the patients who had an abnormal
accounting for classic cardiac risk factors [26]. coronary calcium scan, there were six (12%)
Across the categories of the Framingham risk adverse events. These ndings indicated the low
score, a study of 1461 asymptomatic individuals event rate associated with the absence of elevated
showed at a median of 7 years at follow-up that coronary calcium, and the higher adverse event
coronary calcium was predictive of cardiac risk rate in patients who had elevated coronary
among patients who had a Framingham risk score calcium. However, most patients who had in-
higher than 10%, but not among patients who creased coronary calcium did not experience an
had a score less than 10% [34]. Thus, more recent adverse cardiac event.
studies suggest that coronary calcium scores are The second study by Laudon and colleagues
independently predictive of cardiac risk even after [15] evaluated 105 patients who had chest pain in
accounting for traditional risk factors. the ED. Inclusion criteria were women between 40
and 65 years and men between 30 and 55 years;
normal initial cardiac marker; and a normal or in-
Electron beam CT in patients who present to the
determinate ECG. Exclusion criteria included his-
emergency department with chest pain
tory of CAD; ECG with ST-segment elevation, Q
Between 1999 and 2001, there have been three waves, or left bundle-branch pattern; elevated car-
published studies evaluating the diagnostic use- diac marker at presentation, hemodynamic insta-
fulness of EBCT in patients evaluated in the ED bility; and pregnancy. Patients had follow-up at
who had chest pain [1416]. The rst study by 4 months and a coronary calcium score more
CHEST PAIN IMAGING METHODS IN THE ED 543

than zero was considered abnormal. The need and event, 30 (16%) who had hard events (11 car-
type of stress testing was left to the discretion of diac deaths and 19 nonfatal MIs), and 28 (15%)
the physicians who were blinded to the results of who had other cardiac events, including nine cor-
the scan. onary artery bypass surgeries, four percutaneous
Of these 105 patients, 54 (51%) were admitted interventions, 11 hospitalizations for angina, and
to an inpatient cardiology service and 51 (49%) four ischemic strokes. Patients who had higher
were admitted to the ED chest pain observation coronary artery calcium scores had signicantly
unit. The patient population consisted of 98 more adverse cardiovascular events (Fig. 1). Di-
Whites, 4 African-Americans, 1 Southeast Asian, viding patients into quartiles of coronary calcium
1 Arab, and 1 Hispanic. There were 57 (54%) men scores, there were no hard cardiac events in the
and 48 (46%) women. Mean age for men was 45 rst quartile, one in the second quartile, ten in the
years and for women 51 years. All patients un- third quartile, and 19 in the fourth quartile. In
derwent EBCT and 100 underwent other cardiac a multivariate analysis, coronary calcium scores
testing, including 58 who underwent regular stress were more predictive of adverse cardiac events
testing; 19 who underwent exercise or pharmaco- as compared with traditional cardiac risk factors.
logic radionuclide testing; 25 who underwent
coronary angiography; and 11 who underwent Considerations regarding the potential role of
rest, exercise, or pharmacologic echocardiography. electron beam CT in the emergency department
Of the 100 patients who had EBCT and other Are we ready to start using EBCT in the ED
cardiac testing, 46 patients had an abnormal EBCT for the routine evaluation of patients who have
and 54 patients had a normal EBCT. There were no possible ACS? Use of an ECG, followed by serial
patients who had a normal EBCT and abnormal cardiac markers and then a rapid, easily per-
other cardiac test. EBCT had 100% sensitivity, formed, and inexpensive EBCT is an attractive
63% specicity, 100% negative predictive value, strategy for risk stratication in this setting. For
and 30% positive predictive value for an abnormal many institutions it is a major logistic challenge to
cardiac test. At 4 months there were no adverse provide timely stress tests, which may include
events in patients who had a normal EBCT. imaging modalities, before discharge of low- to
The third study by Georgiou and colleagues intermediate-risk patients from chest pain obser-
[16] considered 192 patients who were evaluated in vation units. However, there are certain patients
the ED and were believed to require admission to who should not have an EBCT in this setting.
exclude MI. Patients were included if they were Patients who have known CAD, which was rightly
aged 30 years or older, had chest pain lasting at an exclusion criterion for all three studies, are
least 20 minutes in the last 12 hours, and had known to have high coronary calcium scores and
a nondiagnostic ECG. Patients were excluded if therefore EBCT would be unlikely to aid in risk
they had a history of CAD (coronary artery by- stratication. This exception is a relevant issue as
pass surgery or percutaneous intervention), ECG
ndings of Q waves, ST-segment elevation more
than 1 mm in two consecutive leads, T-wave inver-
sion of 5 mm or more, hemodynamic instability,
or pregnancy. There were 221 patients screened.
Of these, 13 patients were excluded because of
an acute cardiac event at presentation and 16
were excluded because of lack of follow-up data,
yielding 192 patients for nal analysis.
Of the 192 patients, 69 (36%) were White, 49
(26%) were Hispanic, 48 (25%) were African-
American, and 26 (14%) were Asian-American.
Prevalence of coronary calcium was signicantly
Fig. 1. Patients who have chest pain screened using
lower in Hispanics compared with Whites (47%
electron beam tomography in the emergency depart-
versus 73%, P .004). There were 104 (54%) men ment. (Adapted from Georgiou D, Budo MJ, Kaufer
and 88 (46%) women. The average follow-up was E, et al. Screening patients with chest pain in the emer-
50 months G 10. Of the total group, 116 (60%) gency department using electron beam tomography:
had some level of calcium detected. There were a follow-up study. J Am Coll Cardiol 2001;38:10510;
58 (30%) patients who had an adverse cardiac with permission.)
544 MCCORD & AMSTERDAM

more patients who have atypical symptoms and management? All of these patients likely would
a history of CAD are being managed in chest pain not benet from coronary angiography, as many
units [35]. One study demonstrated that 38% of patients would not have obstructive coronary dis-
patients evaluated in a chest pain unit had known ease as a cause of their symptoms. These patients
CAD (history of MI, percutaneous intervention, would at least require some form of stress testing,
or coronary artery bypass surgery) [36]. In addi- which would make the dicult process of risk
tion, older patients are known to have a higher de- stratication even more cumbersome. In addition,
gree of coronary calcium even without known the need for double-testing would likely make
coronary disease. Thus, if EBCT is to have some the strategy less cost-eective. Instead of using
role in risk stratication of patients in the ED any degree of calcium as dening an abnormal
who have chest pain, it would likely involve EBCT, certain levels of calcium may be used to as-
a younger population. sist in risk stratication that may obviate the need
Although the low adverse event rates in for holding the patient for stress testing in the
patients who have normal EBCT scans in these chest pain unit. However, the extent of coronary
three studies are impressive (Laudon, 0%; calcium is aected by age, sex, and race. The op-
McLaughlin, 0% [excluding the one patient who timal degrees of calcium distinguished by age
had cocaine-induced MI]; Georgiou, 0.6% per and gender would need to be determined to dis-
year), these studies do have some signicant criminate a low-risk patient who has chest pain
limitations. These studies include a small number in the ED. The ongoing Multi-Ethnic Study of
of patients. More unusual cases of adverse cardiac Atherosclerosis (MESA) is following 6500 individ-
events, such as the patient who has MI secondary uals over 10 years and will provide additional data
to cocaine, may become apparent if a larger concerning EBCT and cardiac risk assessment in
number of patients were studied. Also, all three a large diverse population.
of these studies only involved a single center. Finally, although coronary artery calcium is
Moreover, in two of the studies, the patients who associated with atherosclerosis, EBCT does not
were entered represented a convenience sample detect vulnerable plaques that are the most likely
that may have led to selection bias. In the study by to lead to ACS. In fact, some vulnerable plaques
Georgiou and colleagues [16], the EBCT scanner may contain no calcium. A large degree of
was available 5 days per week from 7 AM to coronary calcium is associated with more exten-
11 PM, whereas the study by Laudon and col- sive overall atherosclerotic burden, and therefore
leagues was limited at times in the enrollment more stable and unstable plaques. Although more
of patients by the availability of the EBCT scan- coronary calcium makes developing ACS more
ner [14]. The thought-provoking results of these likely, the absence of coronary calcium does not
small studies will hopefully lead the way to a larger preclude the possibility of ACS. In a study of 118
multicenter study of consecutive patients. patients who had ACS (101 who had acute MI
A further concern of these three studies is the and 17 who had unstable angina), 12 (10%) had
large proportion of positive EBCT scans: 49%, no coronary calcium detected by EBCT [37]. Pa-
64%, and 60% (Table 1). If an EBCT is positive, tients who had ACS and a normal EBCT were sig-
what should the next step be in patient nicantly younger and more likely to be active

Table 1
Electron beam CT for risk-stratication in patients who have chest pain in the emergency department
Number (%)
adverse events in
Number of Age inclusion Number (%) Number (%) patients who have
Study patients criteria in years positive test negative test Follow-up negative test
Laudon et al 105 Men aged 3055 51 (49) 54 (51) 4m 0 (0)
1999 [14] Women aged
4065
McLaughlin et al 134 All ages 86 (64) 48 (36) 30 d 0 (0)
1999 [15]
Georgion et al 192 R30 116 (60) 76 (40) 50 m 1 (2)
2001 [16]
CHEST PAIN IMAGING METHODS IN THE ED 545

cigarette smokers. Thus, there may be certain angiography in elective patients. In their evalua-
groups of patients that are more prone to develop- tion of 59 patients, Leber et al reported the follow-
ing ACS in the setting of a normal EBCT. ing sensitivities of CT angiography for detecting
coronary stenoses of various degrees: stenosis
Cardiac MRI !50%, sensitivity 79%; stenosis O50%, sensitiv-
ity 73%; stenosis O75%, sensitivity 80% [41].
Cardiac MRI has diagnostic usefulness for
Specicity was 97%, indicating the utility of the
myocardial perfusion and regional wall motion
method in identifying absence of CAD. Ra and
abnormalities. Wall motion abnormalities may
colleagues compared invasive and CT coronary
not only be detected in MI, but transient ischemia
angiography in 70 patients in whom they analyzed
in unstable angina may lead to stunning, which in
1,065 coronary artery segments and found a mean
turn can be detected by MRI. An MRI scan may
dierence in percent stenosis of 1.3 G 14.2% [42].
be more sensitive for stunning as compared with
Specicity, sensitivity, positive and negative pre-
radionuclide imaging. Kwong and colleagues [38]
dictive values of CT angiography for signicant
studied 161 consecutive patients who were evalu-
stenoses were: 86%, 95%, 66% and 98%, respec-
ated in the ED for possible ACS. Inclusion criteria
tively. The latter value again indicates the reliabil-
were 30 minutes or more of chest pain within 12
ity of a negative CT coronary angiogram in
hours of presentation, age over 21 years, and
excluding CAD. However, there have been no re-
weight less than 270 lb. Exclusion criteria were
ports of CT coronary angiography in the ED set-
ST-segment elevation MI, pregnancy, signicant
ting in patients presenting with chest pain.
heart failure such that patient could not lie at,
Visualization of the coronary arteries by CT
and metal prosthesis precluding MRI. Of the
has important advantages that include its non-
161 patients, 25 (16%) had ACS, 10 had
invasive methodology, rapidity, reliability in ex-
NSTEMI, and 15 had unstable angina. The sensi-
cluding severe CAD and simultaneous acquisition
tivity and specicity of MRI for ACS were 84%
of other chest structures such as the aorta and
and 85%, respectively. In a multivariate regres-
pulmonary arteries. However, limitations of the
sion analysis, MRI was independently associated
technique are signicant. A low heart rate (opti-
with the diagnosis of ACS and was of greater di-
mally, !70/min) is required to avert blurring of
agnostic usefulness when compared with ECG
the images of the moving coronary arteries.
and peak troponin-I. Considerable research is
Administration of a beta blocker or rate-limiting
presently in progress with MRI in identifying cor-
calcium channel blocker is often necessary for this
onary atherosclerotic lesions, including vulnerable
purpose. Regular rhythm is also necessary pre-
plaques, which may prove to be of diagnostic
cluding CT angiography in patients with atrial
value for patients in the ED who have chest
brillation or frequent ectopic beats. Even with the
pain of unclear origin [39].
currently enhanced resolution, the method cannot
precisely assess the severity of coronary stenoses,
CT coronary angiography
which has resulted in diagnostic categories such as
Major advances in noninvasive imaging of the no CAD, !50% stenosis, O50% stenosis and
coronary arteries by contrast-enhanced CT have uninterpretable due to calcication or artifact. A
stimulated considerable interest in the potential of dose of 60100 cc of iodinated contrast is required,
this technique to rapidly conrm or exclude CAD which may be a relative contraindication in
in low-intermediate risk patients presenting to the patients with renal insucieny. Finally, the radi-
ED with chest pain. Contemporary 64 slice CT ation dose is higher than that with invasive
scanners [4042] have overcome many of the lim- coronary angiography and has led to methods
itations of earlier 16 slice devices [43], resulting in for reducing the exposure of the patient [43].
recently cited gantry rotation times of 330 msec,
spatial resolution of 0.4 mm and temporal resolu-
tion of 83165 msec. Although these values are
Summary
less than those of conventional coronary angiog-
raphy (w0.15 mm and 0.33 sec), they are sucient Three small single-center studies using EBCT
to yield diagnostic data in a large proportion of in the ED to risk stratify patients who have chest
patients (Fig. 2). The accuracy of CT coronary an- pain and nondiagnostic ECGs show promising
giography with 64 slice scanners is based on limited results, but limitations of this approach are
comparative studies with invasive coronary apparent. Although patients who did not have
546 MCCORD & AMSTERDAM

Fig. 2. Noninvasive visualization of coronary artery anatomy by 64-slice multislice computed tomography (CT). (A,B) Vol-
ume rendering technique depicts a severe stenosis of the right coronary artery (RCA) distal to the marginal branch, nodular
coronary calcications mainly extrinsic to the RCA lumen and normal left coronary artery. (C,D) Maximum-
intensity projection demonstrates severe soft plaque of the RCA and supercial calcic plaques of the RCA and proximal
left circumex artery. (E,F) Invasive coronary angiography conrms the major stenosis of the RCA and absence of signicant
lesions in the left coronary artery. (From Ra GL, Gallagher MJ, ONeill WW, et al. Diagnostic accuracy of noninvasive
coronary angiography using 64-slice spiral computed tomography. J Am Coll Cardiol 2005;46:5527; with permission.)

coronary calcium detected by EBCT had very low required to better ascertain for which subset of
adverse event rates, EBCT in this setting can be patients this would be an eective strategy. This
associated with poor positive predictive value that subset would likely include a younger population
would lead to at least the need for another cardiac that does not have a history of CAD. Ideally this
test, such as stress testing, or even unnecessary would lead to a randomized diagnostic trial
hospital admission. Before EBCT is adopted as comparing EBCT in this population with some
routine practice in this patient population, a large form of stress testing. Other imaging technologies,
multicenter study of consecutive patients is including cardiac MRI and MSCT, are even less
CHEST PAIN IMAGING METHODS IN THE ED 547

well studied in this patient population, but may Heart Association Writing Group. Circulation
prove to be of value in this setting in the future. 1996;94:117592.
Limited data with 64 slice scanners indicate [11] Schwartz RS. Coronary artery calcication and
diagnostic quality coronary imaging in many prognosis: same song, second verse? J Am Coll Car-
diol 1994;24:35961.
patients and a high degree of reliability of
[12] Wong ND, Detrano RC, Abrahamson D, et al. Cor-
a negative study. However, although the tech- onary artery screening by electron beam computed
nique has important potential, signicant limita- tomography: facts, controversy, and future. Circula-
tions remain and there are currently no reported tion 1995;92:6326.
studies of its utility in the ED setting. [13] Detrano R, Hsiai T, Wang S, et al. Prognostic value
of coronary calcication and angiographic stenoses
in patients undergoing coronary angiography.
References J Am Coll Cardiol 1996;27:28590.
[14] McLaughlin VV, Balogh T, Rich S. Utility of elec-
[1] Ross R. The pathogenesis of atherosclerosis: a per- tron beam computed tomography to stratify patients
spective for the 1990s. Nature 1993;362:8019. presenting to the emergency room with chest pain.
[2] Stary HC. Composition and classication of human Am J Cardiol 1999;84:3278.
atherosclerotic lesions. Virchows Arch A Pathol [15] Laudon DA, Vukov LF, Breen JF, et al. Use of elec-
Anat Histopathol 1992;421:27790. tron-beam computer tomography in the evaluation
[3] Stary HC, Chandler AB, Dinsmore RE, et al. A def- of chest pain patients in the emergency department.
inition of advanced types of atherosclerotic lesions Ann Emerg Med 1999;33:1521.
and a histological classication of atherosclerosis: [16] Georgiou D, Budo MJ, Kauger E, et al. Screening
a report from the Committee on Vascular Lesions patients with chest pain in the emergency depart-
of the Council on Arteriosclerosis, American Heart ment using electron beam tomography. J Am Coll
Association. Circulation 1995;92:135574. Cardiol 2001;38:10510.
[4] Simons DB, Schwartz RS, Edwards WD, et al. Non- [17] Stanford W, Thompson BH, Burns TL, et al. Coro-
invasive denition of anatomic coronary artery dis- nary artery calcium quantication at multi-detector
ease by ultrafast computed tomographic scanning: row helical CT versus electron-beam tomography.
a quantitive pathologic comparison study. J Am Am J Cardiol 2001;87:210.
Coll Cardiol 1992;20:111826. [18] Carr JJ, Crouse JR, Go DC, et al. Evaluation of
[5] Mautner SL, Mautner GC, Froehlich J, et al. Coro- subsecond gated helical CT for quantication of cor-
nary artery calcication: assessment with electron onary artery calcium and comparison with electron
beam CT and histomorphometric correlation. Radi- beam CT. AJR Am J Roentgenol 2000;174:91521.
ology 1994;192:61923. [19] Rumberger JA, Brundage BH, Rader DJ, et al. Elec-
[6] Rumberger JA, Simons DB, Fitzpatrick LA, et al. tron beam computed tomographic coronary calcium
Coronary artery calcium area by electron-beam scanning: a review and guidelines for use in asymp-
computed tomography and coronary atherosclerotic tomatic persons. Mayo Clin Proc 1999;74:24352.
plaque area: a histopathologic correlative study. Cir- [20] Nasir K, Raggi P, Rumberger JA, et al. Coronary ar-
culation 1995;15:215762. tery calcium volume scores on electron beam tomog-
[7] Sangiorgi G, Rumberger JA, Severson A, et al. Arte- raphy in 12,936 asymptomatic adults. Am J Cardiol
rial calcication and not lumen stenosis is highly 2004;93:11469.
correlated with atherosclerotic plaque burden in [21] Callister TQ, Cooil B, Raya SP, et al. Coronary ar-
humans: a histologic study of 723 coronary artery tery disease: improved reproducibility of calcium
segments using non-decalcifying methodology: elec- scoring with an electron-beam CT volumetric
tron beam computed tomography and coronary method. Radiology 1998;208:807.
artery disease: scanning for coronary artery calci- [22] Achenbach S, Ropers D, Mohlenkamp S, et al.
cation. J Am Coll Cardiol 1998;31:12633. Variability of repeated coronary artery calcium
[8] Kennedy J, Shavelle R, Wang S, et al. Coronary cal- measurements by electron beam tomography. Am
cium and standard risk factors in symptomatic J Cardiol 2001;87:210.
patients referred for coronary angiography. Am [23] Kondos GT, Ho JA, Sevrukov A, et al. Electron-
Heart J 1998;135:696702. beam tomography coronary artery calcium and car-
[9] Secci A, Wong N, Tang W, et al. Electron beam diac events: a 37-month follow-up of 5635 initially
computed tomographic coronary calcium as a pre- asymptomatic low- to intermediate-risk adults. Cir-
dictor of coronary events: comparison of two proto- culation 2003;107:25716.
cols. Circulation 1997;96:11229. [24] Arad Y, Spadaro LA, Goodman K, et al. Predic-
[10] Wexler L, Brundage B, Crouse J, et al. Coronary ar- tive value of electron beam computed tomography
tery calcication: pathophysiology, epidemiology, of the coronary arteries: 19-month follow-up of
imaging methods, and clinical implications: a state- 1173 asymptomatic subjects. Circulation 1996;93:
ment for health professionals from the American 1951.
548 MCCORD & AMSTERDAM

[25] Arad Y, Spadaro LA, Goodman K, et al. Prediction [35] Amsterdam EA, Kirk JD, Dierks DB, et al. Early ex-
of coronary events with electron beam computed to- ercise testing in the management of low risk patients
mography. J Am Coll Cardiol 2000;36:1253. in chest pain centers. Prog Cardiovasc Dis 2004;46:
[26] Pletcher MJ, Tice JA, Pignone M, et al. Using the 43852.
coronary artery calcium score to predict coronary [36] McCord J, Nowak RM, McCullough PA, et al.
heart disease events: a systematic review and meta- Ninety-minute exclusion of acute myocardial infarc-
analysis. Arch Intern Med 2004;164:1285. tion by use of quantitative point-of-care testing of
[27] Park R, Detrano R, Xiang M, et al. Combined use of myoglobin and troponin I. Circulation 2001;104:
computed tomography coronary calcium scores and 14838.
C-reactive protein levels in predicting cardiovascular [37] Schmermund A, Baumgart D, Gunter G, et al. Cor-
events in nondiabetic individuals. Circulation 2002; onary artery calcium in acute coronary syndromes:
106:2073. a comparative study of electron-beam computed to-
[28] Yang T, Doherty TM, Wong ND, et al. Alcohol con- mography, coronary angiography, and intracoro-
sumption, coronary calcium, and coronary heart dis- nary ultrasound in survivors of acute myocardial
ease events. Am J Cardiol 1999;84:802. infarction and unstable angina. Circulation 1997;
[29] Raggi P, Cooil B, Callister TQ. Use of electron beam 96:14619.
tomography data to develop models for predictor of [38] Kwong RY, Schussheim AE, Rekhraj S, et al.
coronary events: comparison of two protocols. Cir- Detecting acute coronary syndrome in the emer-
culation 1997;96:1122. gency department with cardiac magnetic resonance
[30] Wong ND, Hsu JC, Detrano RC, et al. Coronary ar- imaging. Circulation 2003;107:5317.
tery calcium evaluation by electron beam computed [39] Fayad ZA. MR imaging for the noninvasive assess-
tomography and its relation to new cardiovascular ment of atherothrombotic plaques. Magn Reson Im-
events. Am J Cardiol 2000;86:495. aging Clin N Am 2003;11(1):10113.
[31] OMalley PG, Taylor AJ, Jackson JL, et al. Prognos- [40] Leber AW, Knez A, Becker A, et al. Accuracy of
tic value of coronary electron-beam computed to- multidetector spiral computed tomography in iden-
mography for coronary heart disease events in tifying and dierentiating the composition of coro-
asymptomatic populations. Am J Cardiol 2000;85: nary atherosclerotic plaques: a comparative study
945. with intracoronary ultrasound. J Am Coll Cardiol
[32] ORourke RA, Brundage BH, Froelicker VF, et al. 2004;43:12417.
American College of Cardiology/American Heart [41] Budho MJ, Cohen MC, Garcia MJ, et al. ACCF/
Association expert consensus document of electron AHA Clinical competence statement on cardiac im-
beam computed tomography for the diagnosis and aging with computed tomography and magnetic res-
prognosis of coronary artery disease. Circulation onance. J Am Coll Cardiol 2005;46:383402.
2000;102:12640. [42] Leber AW, Knez A, von Ziegler F, et al. Quantica-
[33] Wilson PWF, DAgostino RB, Levy D, et al. Predic- tion of obstructive and nonobstructive coronary
tion of coronary heart disease using risk factor cate- lesions by 64-slice computed tomography. J Am
gories. Circulation 1998;97(18):183747. Coll Cardiol 2005;46:14754.
[34] Greenland P, LaBree L, Azen SP, et al. Coronary [43] Ra GL, Gallagher MJ, ONeill WW, et al. Diag-
Artery calcium score combined with Framingham nostic accuracy of noninvasive coronary angiogra-
score for risk prediction in asymptomatic individu- phy using 64-slice spiral computed tomography.
als. JAMA 2004;291(2):2105. J Am Coll Cardiol 2005;46:5527.
Cardiol Clin 23 (2005) 549557

Chest Pain Units: Management of Special Populations


Deborah B. Diercks, MDa,*, J. Douglas Kirk, MD, FACEPa,
Ezra A. Amsterdam, MD, FACCb
a
Department of Emergency Medicine, University of California School of Medicine (Davis) and Medical Center,
4150 V Street, PSSB Suite 2100, Sacramento, CA 95817, USA
b
Department of Internal Medicine, University of California, Davis Medical Center,
2315 Stockton Boulevard, Sacramento, CA 95817, USA

Chest pain evaluation units provide a safe and by accelerated diagnostic protocols, focusing on
cost-eective means for risk stratication of the value and limitations of this strategy for
patients who present to the emergency department assessment of these unique patients.
(ED) with chest pain and no evidence of myocar-
dial infarction (MI) or ischemia on initial evalu- Women
ation. A standardized protocol is often used that
combines serial measurements of cardiac injury Cardiovascular disease is the leading cause of
markers, repeat electrocardiograms (ECG), and death in women, accounting for nearly half of all
telemetry monitoring followed by secondary risk deaths in this population. Timely diagnosis early
stratication by additional diagnostic testing for in the course of the disease is essential. Thirty-
coronary artery disease (CAD). The latter usually eight percent of women who have MI die within
comprises provocative testing with exercise or a year, compared with 25% of men. This stagger-
pharmacologic stress. Although these units rely ing variation is in part because MI occurs at
heavily on a protocol-driven approach, apprecia- a signicantly older age in women than in men. In
tion of the limitations of this method in certain addition, within 6 years of MI, as many as one
patient populations is essential. In particular, the third of women will have another heart attack,
inclusion of women, diabetics, patients with 15% will develop angina, 13% will have a stroke,
a history of CAD, and those with chest pain after and 30% will be disabled with heart failure [1].
stimulant use (eg, cocaine, methamphetamine) These alarming numbers have led to an increased
may necessitate additional consideration in the interest in cardiovascular health in women. Re-
choice of diagnostic testing. Some would suggest cent studies have revealed unique aspects of path-
that these patients should be excluded from chest ophysiology, diagnosis, prognosis, and treatment
pain unit evaluation based solely on their pre- of CAD in women. The lower age-specic risk
sumed high risk of cardiac disease and adverse of CAD and prevalence of acute coronary syn-
events. Additionally, exercise stress, the most drome (ACS) in women leads to decreased accu-
common provocative test used in chest pain units, racy of diagnostic tools commonly used in
has important limitations in these patients. How- a chest pain unit (Box 1).
ever, the clinical presentations of many of these In addition, there are important gender-specic
patients often lack high-risk features associated physiologic responses to exercise stress [2]. Chal-
with adverse cardiac events, making them appro- lenges to diagnostic testing in women include vari-
priate candidates for a chest pain unit. This article able ability to reach target heart rate and baseline
reviews the evaluation of these special populations ECG changes that may decrease accuracy of
treadmill testing [35]. Women also have labile
* Corresponding author. ST-segment alterations and body habitus attrib-
E-mail address: dbdiercks@ucdavis.edu utes that may result in false-positive exercise
(D.B. Diercks). ECG interpretation. Despite these unique
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.015 cardiology.theclinics.com
550 DIERCKS et al

these criticisms, the ED-based exercise stress test


Box 1. Unique factors in women has provided useful prognostic information for ad-
affecting exercise test results verse cardiac events in low- to moderate-risk chest
pain patients that have included women [10,11].
Decreased exercise capacity Diercks and colleagues have evaluated the use of
Baseline electrocardiographic treadmill exercise testing in 666 women who un-
abnormalities derwent evaluation in our chest pain unit. Using
Body habitus Agency for Health Care Policy and Research clas-
Lower age-specific risk of coronary sication, 38 (5.7%) women were retrospectively
artery disease identied as high risk, 136 (20.4%) were intermedi-
High rate of single vessel disease ate risk and 492 (73.9%) were low risk for CAD.
Of these, 465 (70%) had negative exercise testing,
145 (22%) had nondiagnostic tests, and 56 (8%)
physiologic issues, the chest pain unit provides an had positive tests. Thirty-day clinical follow-up
important option for the evaluation of women. or conrmatory cardiac evaluation was achieved
Their exclusion would result in the costly use of in 512 of 666 (77%) patients. The sensitivity, spec-
traditional hospital beds to evaluate these patients icity, and negative and positive predictive values
with a low rate of CAD, especially those of pre- of the treadmill test are shown for patients with
menopausal age, no diabetes, and absence of a positive or abnormal test result (positive or non-
a smoking history. Nevertheless, understanding diagnostic) in Table 1. Compared with a negative
these limitations is important for the prudent se- exercise test, the relative risk of CAD for a positive
lection of diagnostic modalities when evaluating test was 19 (95% CI 656) and for a nondiagnostic
women in a chest pain unit. test it was 8 (95% CI 322) [12]. Despite the poor
Several approaches to risk stratication in positive predictive value of exercise testing in the
women in the nonacute setting have been pub- women, positive tests were relatively infrequent
lished. A recent review of the evaluation of chest and the negative predictive value aords excellent
pain in women suggests that it is possible to risk accuracy in identifying those patients who do not
stratify them into low, moderate, and high likeli- require hospital admission. As recommended in
hood of CAD solely by the presence or absence of the American College of Cardiology/American
traditional cardiac risk factors. Based on this Heart Association guidelines, treadmill testing
scheme, recommendations for diagnostic testing should be applied as a rst-line diagnostic method
were presented to improve the cost-eectiveness of only in those women with a normal baseline ECG
the evaluation [6]. However, these recommenda- and ability to exercise [13].
tions have not been validated in any published Imaging techniques such as myocardial stress
clinical trials. Wilson and colleagues [7,8] devel- perfusion scintigraphy have higher sensitivity and
oped a risk prediction model based on blood pres- specicity than treadmill testing, but gender-
sure, cholesterol level, history of diabetes or
smoking, and age. This model was derived from Table 1
Framingham data; information on estrogen use Diagnostic accuracy of positive or abnormal exercise
was not available. The risk score was very eective treadmill testing in women evaluated in a chest pain ob-
in stratifying remaining lifetime risk for coronary servation unit
heart disease in women, with 1.5- to 3-fold higher Negative Positive
absolute risk in the highest versus lowest tertile of predictive predictive
risk score at all ages. Despite initial promise, risk Sensitivity Specicity value value
stratication based on simple tools such as these Positive 41% 91% 96% 22%
has not gained general clinical use and outcome exercise
data are limited. In part, this may be due to the treadmill
lack of utility of such models in the acute setting. testing
Exercise stress testing, a cornerstone of the (N 56)
evaluation of patients with chest pain, has a di- Abnormal 82% 70% 99% 14%
agnostic accuracy as low as 52% in women, with exercise
a false-positive rate as high as 40% [9,10]. Thus, treadmill
its utility in women has been questioned, especially testing
(N 201)
in those considered low risk for CAD. Despite
CHEST PAIN UNITS AND SPECIAL POPULATIONS 551

specic confounders still exist. Breast attenuation, These recommendations are based on the high
small left ventricular size, and a high rate of single sensitivity and specicity and prognostic utility
vessel disease may aect the diagnostic utility of of scintigraphy [23].
this modality [14]. Despite these limitations, myo-
cardial perfusion imaging has been established as History of coronary artery disease
a safe, noninvasive, cost-eective and accurate Patients with a history of CAD compose
method to identify CAD in women [15]. After in- a high-risk subgroup whose inclusion in the
jection of thallium-201, areas of decreased uptake standard chest pain unit strategy is controversial.
by the myocardium are seen as a negative scinti- These patients are often complex and frequently
graphic image, indicating regions of ischemic or have undergone prior myocardial revasculariza-
infarcted myocardium. Gated single photon emis- tion. In general, their suitability for chest pain
sion CT radionuclide imaging with thallium or unit management should primarily depend on the
technetium99 sestamibi has been used in the ED immediate risk of ACS based on symptoms, ECG
to improve the diagnostic accuracy and risk strat- features, and cardiac serum markers rather than
ication of patients presenting with chest pain and their past history and risk factors (Box 2) [24,25].
a nondiagnostic ECG [1618]. Other considerations should include review of
Cardiac stress imaging methods provide sev- prior evaluations and management goals of the
eral advantages in women. They allow for phar- patients cardiologist.
macologic stress, which alleviates the problem of The Chest Pain Evaluation in the Emergency
decreased exercise capacity and failure to reach Room (CHEER) trial examined the utility of
target heart rate that limits the value of traditional a chest pain observation unit in a group of
exercise stress testing. In addition, the use of gated patients with suspected unstable angina and in-
images allows dierentiation of breast tissue termediate risk for adverse cardiac events based
attenuation from a true perfusion defect. Studies on AHCPR guidelines [26]. This cohort of pa-
that have evaluated the benet of gated images, tients included 27% with known CAD, dened
which included 170 women, report an improved by history of prior MI or prior percutaneous cor-
specicity for the detection of CAD from 67% to onary intervention. After adjustment for comor-
91% [19,20]. bid conditions, the investigators observed no
The prognostic value of myocardial perfusion dierence in cardiac events at 6 months in those
imaging has also been evaluated in women. patients randomized to the chest pain unit and
Pooled data from more than 7500 women dem- those receiving traditional hospital admission
onstrated an annual cardiac event rate of ! 1% in [26]. Our group evaluated the use of immediate ex-
those with a normal perfusion scan [14]. These re- ercise testing of 100 patients with a history of
sults were also found in women with a high pretest CAD evaluated in a chest pain unit. These pa-
probability of CAD [21]. Hachamovitch and col- tients had a normal or nondiagnostic ECG at pre-
leagues [22] evaluated the determinants of risk sentation and were pain free at the time of
and prognosis in patients with a normal myocar- evaluation. A negative exercise treadmill test was
dial perfusion scan. Diabetic women had the high- found in 38% of patients and 39% had nondiag-
est rate of cardiac-related death or MI at 1 year nostic tests, with two-thirds discharged home im-
after adjusting for confounders. The lowest mediately from the ED. An additional 19
1-year event rate was in nondiabetic women
(! 1%). The results of these trials have led the
American Society of Nuclear Cardiology Task Box 2. Unique factors in patients with
Force on Women and Heart Disease to incorpo- a history of coronary artery disease that
rate the use of myocardial scintigraphy into an al- affect stress test results
gorithm for the initial evaluation of women with
chest pain. These guidelines recommend the use Multiple co-morbidities
of exercise or pharmacologic gated single photon Prior myocardial revascularization
emission CT as a rst line test in intermediate- Baseline left ventricular wall motion
to high-risk women with diabetes, an abnormal abnormalities
resting ECG, or inability to exercise. In addition, Baseline fixed myocardial perfusion
they recommend this modality as a second-line di- deficits
agnostic adjunct or conrmatory test in patients Use of anti-anginal medications
who have an intermediate risk exercise ECG.
552 DIERCKS et al

patients were discharged within 24 hours of pre- normal perfusion, perfusion decit consistent
sentation. Of the 23% who had a positive exercise with single vessel disease, and that consistent
test, uncomplicated non-ST segment elevation MI with multi-vessel disease, respectively. Although
was diagnosed in two patients. No adverse events these studies were not performed in patients in
occurred during the 6-month follow-up period. a chest pain unit, they support the use of scintig-
Despite a slightly higher incidence of abnormal raphy in this setting for risk stratication of pa-
tests and adverse events than the typical chest tients with history of CAD.
pain unit patient population, this strategy was A further issue complicating chest pain unit
quite eective in risk stratication and prompt management of patients with a history of CAD is
disposition of patients [27]. the high prevalence of anti-ischemic medications
Diercks and colleagues [28] have also evaluated in this group, which may confound the diagnostic
the utility of an accelerated diagnostic protocol in utility of stress testing. Beta-blockers and calcium
patients with a history of CAD evaluated in their channel-blockers limit the heart rate and blood
chest pain unit from 1/94 to 10/01. A diagnosis of pressure response during exercise testing and they
ACS was based on clinical presentation and car- can avert angina in patients with CAD [3133].
diac testing consistent with myocardial ischemia, There are conicting data regarding the eect of
MI, or the need for revascularization within 30 these drugs on the sensitivity of the exercise test
days of presentation. Of the 6839 patients evalu- in detection of CAD, and patients are often con-
ated, 1153 (17%) had a history of CAD. MI was sidered ineligible for diagnostic exercise testing if
detected in 12 patients (1%) during their chest they are taking these agents [3138]. This potential
pain unit evaluation. Initial diagnostic testing problem may exclude patients from an accelerated
was performed in 639/1153 (55%), yielding posi- assessment in a chest pain unit, cause a delay in di-
tive results in 191. Testing included 371 (58%) ex- agnosis, or require hospital admission for further
ercise treadmill tests, 208 (32%) myocardial management.
scintigrams, 54 (8%) exercise echocardiograms, Diercks and colleagues [39] have evaluated the
and 51 (7%) cardiac catheterizations. Subsequent use of beta-blockers and calcium channel-blockers
testing was done in 77 patients who had no initial in patients undergoing immediate stress testing in
diagnostic test during their chest pain unit stay. their chest pain unit. Of 176 patients on one of
During the follow-up period there were two these agents who underwent exercise testing, 50
deaths, three MIs, and revascularization was per- (28.6%) were taking a beta-blocker only, 116
formed in 46 patients. A nal diagnosis of ACS (66.6%) a calcium channel-blocker only, and 10
was made in 238 patients (22%). The data indicate (5.7%) were being treated with both classes of
that although the frequency of ACS in patients drugs. After adjusting for age, gender, cardiac
with a history of CAD is higher than that in typ- risk factors, and presenting complaint, those pa-
ical low-risk patients presenting with chest pain, tients on beta-blockers or calcium channel-block-
evaluation in a chest pain unit is appropriate. ers (study group) were more likely to have
Studies using myocardial scintigraphy in pa- nondiagnostic tests 69 of 176 (39%) (OR 2.1,
tients with a history of CAD have reported 95% CI 1.53.1) compared with patients in the co-
satisfactory diagnostic and prognostic ability. hort who were not taking these agents. The pro-
Zellweger and colleagues [29] evaluated the prog- portion of patients with a nondiagnostic exercise
nostic value of stress radionuclide imaging in pa- test was higher in the group on antianginal drugs,
tients with prior MI. In 1143 patients, they and beta-blockers were more problematic in this
reported 118 adverse events during follow-up. regard than calcium channel-blockers. However,
The annual rate of adverse events was proportional the majority of patients on one or both of these
to the size of the perfusion abnormality. In addi- drugs had a diagnostic test 107 of 176 (61%) (neg-
tion, the size of the infarction was also an inde- ative or positive), conrming the utility of exercise
pendent predictor of adverse events. Elhendy testing in these patients. Thus, despite a higher
and colleagues [30] studied the use of stress tech- rate of nondiagnostic tests, inclusion of these
netium 99 sestamibi in patients with a history of patients in a chest pain unit protocol did aord
prior MI. In the 383 patients evaluated, there timely disposition of a substantial proportion
were 48 cardiac events during follow-up. The fre- that would have otherwise been admitted to the
quency of events was again directly proportional hospital [39].
to the size of the perfusion defect with an event Taillefer and colleagues [40] evaluated the in-
rate of 0.4%, 2.4%, and 4% in patients with uence of beta-blockers in patients undergoing
CHEST PAIN UNITS AND SPECIAL POPULATIONS 553

dypridamole Tc-99 sestamibi myocardial scintig-


raphy. They randomly assigned patients with Box 3. Unique factors in diabetics that
known CAD to receive placebo, low-dose meto- affect stress test results
prolol, or high-dose metoprolol. Of the 21 pa-
tients completing the study, the sensitivity for Increased risk of cardiac events in
detection of CAD was 85.7% in the placebo group females with diabetes
versus approximately 71.0% in both the low- and Lower predictability of traditional
high-dose beta-blocker group. In addition, the cardiac risk factors
beta-blocker appeared to decrease the ability to High rate of silent myocardial ischemia
assess severity of CAD based on the size of the Limited exercise capacity
perfusion decit compared with placebo [39]. Al- Nonspecific baseline ST segment
though the rate of nondiagnostic tests is higher alterations
and sensitivity lower for detecting CAD in pa-
tients on beta-blockers, these medications should
not be considered an absolute contraindication also been an acceptable tool for detection of
to chest pain unit management. Although patients CAD in diabetic patients. A retrospective review
with a history of CAD, including those on anti- of diabetic patients who underwent myocardial
anginal medications, pose a challenge to traditional perfusion imaging followed by cardiac catheteri-
chest pain unit protocols, their inclusion in this di- zation reported a sensitivity of 97% and positive
agnostic strategy appears to be safe and eective if predictive value of 88% for CAD [43]. Kang
appropriate selection criteria are used. and colleagues [44] reported a similarly high sensi-
tivity (86%) but a specicity of 56% in a similar
patient population. These studies support a detec-
Diabetics
tion rate for CAD by stress scintigraphy, but spec-
Diabetic patients represent a group with in- icity of the method may be limited in diabetic
creased risk and a high rate of atypical symptoms patients.
for CAD. Cardiovascular disease is the leading Myocardial stress scintigraphy has also pro-
cause of death in diabetic patients, and the vided prognostic information. Berman and col-
prevalence of CAD in some diabetic populations leagues [45] evaluated 2826 patients who
is over 50% [23]. Detection of CAD and risk strat- underwent adenosine myocardial perfusion imag-
ication is often confounded by several unique ing. They noted that diabetics who had an abnor-
features, including the markedly increased risk in mal perfusion scan were at a signicantly
female diabetics, lower predictive value of LDL increased risk of death compared with nondia-
cholesterol, and high rate of silent myocardial is- betics and that perfusion imaging data was an in-
chemia. In addition, diabetic patients often have dependent predictor of events. Including diabetic
poor exercise tolerance and nonspecic ST seg- patients in an accelerated protocol that involves
ment alterations that can result in nondiagnostic stress scintigraphy is a potentially important tool
or false-positive exercise testing (Box 3) [41]. for safe, expedient diagnosis and risk stratication
Despite the foregoing confounders, the value in these often dicult to manage patients.
of exercise testing for detection of CAD in The studies in diabetics cited previously were
asymptomatic patients with type 2 diabetes was performed largely in stable outpatients. Diabetic
demonstrated by Bacci and colleagues [42]. They patients are frequently excluded from chest pain
studied 141 patients who also had peripheral arte- unit evaluation due to their presumed high risk
rial disease or atherogenic risk factors, 71 of [7]. As in patients with a prior history of CAD, the
whom underwent coronary angiography. The di- suitability of diabetics for chest pain unit manage-
agnostic accuracy of exercise treadmill testing ment is controversial, but it should primarily de-
was 79%. In patients with a positive exercise test, pend on the immediate risk of ACS rather than
20 of 27 (74%) had a stenotic lesion O 70%. the likelihood of underlying CAD. However, the
Marwick and colleagues [41] evaluated the use safety and ecacy of including diabetics in a rapid
of stress echocardiography in 937 patients with diagnostic protocol are not well documented in
diabetes and conrmed CAD or symptoms of the literature. Diercks and colleagues [46] have as-
CAD and observed that a positive stress echocar- sessed the use of immediate exercise treadmill test-
diogram was an independent predictor of death. ing in diabetic patients evaluated in a chest pain
Elective stress myocardial perfusion imaging has unit. Their ndings demonstrated that, whereas
554 DIERCKS et al

diabetics had more frequent evidence of CAD at According to recent data, 25 million people in the
30-day follow-up than nondiabetics (12% versus United States have reported using cocaine [48,49],
5%, RR 2.6, CI 1.74.1, P ! .0001) and required and it is likely that these numbers are similar for
more subsequent diagnostic studies (26% versus methamphetamine. The cardiovascular complica-
14%, RR 2.0, CI 1.52.6, P ! .001), there was tions of cocaine have been well studied. It has
no signicant dierence in acute cardiac events been reported that the risk of MI is 24 times base-
or length of stay in the chest pain unit between line risk in the hour after cocaine use [50]. In pa-
the two populations. Immediate exercise treadmill tients under the age of 45 years with acute MI,
testing was also equally sensitive in detecting approximately 25% are cocaine related [51]. Of
CAD in diabetics as in nondiabetics. Hence, exer- patients who present to the ED with cocaine-related
cise testing in diabetics with a normal baseline chest pain, 6% have MI and the rate of cardiac
ECG and negative cardiac injury markers is arrhythmia has been reported to be over 2%
a safe, ecacious, and cost-eective method for [52]. Because there are no prospective studies
risk stratication and improves quality of care in in the methamphetamine patient population,
this patient population. data for cocaine are often extrapolated to the
Further evidence in support of the safety of methamphetamine patients. A small study of pa-
this strategy was presented by Sanchez and tients admitted to the hospital with a positive
colleagues [47] who evaluated the 30-day out- screen for methamphetamine reported an in-hos-
comes, resource use, and severity of CAD in pa- pital event rate of 8%, and in this selected
tients with and without diabetes admitted to a group of patients ACS was diagnosed in 25%
chest pain unit, who comprised a subgroup of the [53]. Because this rate of ACS is similar to
Chest pain Evaluation and Creatinine Kinase- that in the typical chest pain unit population,
MB, Myoglobin, and Troponin I (CHECK- it is important to examine the potential utility
MATE) study. The study included 772 patients, of this strategy in patients with chest pain dur-
of whom 109 had diabetes. All patients were as- ing stimulant use. Important considerations in-
signed to a chest pain unit or evaluated in the clude dening the optimal time of observation
ED for more than 6 hours. In accord with the au- and the need for specic diagnostic testing after
thors aforementioned ndings, the investigators an initial negative evaluation that included car-
reported an increase in adverse cardiac events diac injury markers and ECG on presentation
(death/MI) in the diabetic patients (8.3% versus to ED.
3.2%, P .027). In addition, they reported higher Weber and colleagues [48] performed a pro-
rates of the following events in the diabetic group: spective study to determine the safety of a 9- to
admission from the chest pain unit, incidence of 12-hour period of observation in patients who
elevated cardiac markers, and rate of multivessel were at low to moderate risk for coronary events.
disease, especially in patients with O 2 cardiac The patient cohort consisted of 347 patients pre-
risk factors. However, there were no safety issues senting with acute chest pain associated with co-
related to assignment to the accelerated diagnostic caine use. A total of 42 patients (12%) were
protocol. Indeed, even the higher adverse event admitted to the hospital, of whom 20 had a nal
rate in the diabetic group was well within that ex- diagnosis of ACS, three had congestive heart fail-
pected in a typical chest pain unit population. ure, and two had ventricular tachycardia. Of the
These studies indicate that chest pain unit remaining 305 patients discharged home after
observation is appropriate in the foregoing pop- a negative 9- to 12-hour observation period, in-
ulations if it is predicated on recognition of the cluding exercise testing, there were four nonfatal
higher prevalence of CAD in these patients and MIs (1.6%) in patients who continued to use co-
appreciation of the limitations of the diagnostic caine. The investigators concluded that a 9- to
tools. Rigorous patient selection and rm adher- 12-hour period of observation that included cardiac
ence to protocol standards cannot be overempha- monitoring, serial cardiac biomarkers, and stress
sized to enhance safe evaluation of in the chest testing based on physician discretion was safe.
pain unit. These results support prior retrospective stud-
ies that evaluated the use of an observation unit
for the management of cocaine related chest pain.
Chest pain related to stimulant use
Kushman and colleagues [54] reported a similarly
The use of stimulants such as cocaine and low rate of complications in a retrospective study
methamphetamines is of epidemic proportions. of 197 consecutive patients admitted to a chest
CHEST PAIN UNITS AND SPECIAL POPULATIONS 555

pain unit. Their protocol included cardiac bio- women in chronic heart failure secondary to ische-
markers at hours 0, 3, and 6 and consultation by mic or idiopathic dilated cardiomyopathy. Am J
a cardiology fellow with selective use of stress Cardiol 1997;80:858.
evaluation by treadmill testing. Of the entire [3] Cummings GR, Dufrense C, Kich L, Sann J. Exer-
cise electrocardiogram patterns in normal women.
group, 22 (11.2%) were admitted to the hospital
Br Heart J 1973;35:105561.
from the observation unit, one of whom had [4] Weiner DA, Ryan TJ, Parsons L, et al. Long-term
MI. In the remaining patients who were dis- prognostic value of exercise testing in men and
charged, there was one (0.6%) cardiac complica- women from the Coronary Artery Surgery Study
tion during follow-up. However, the implications (CASS) registry. Am J Cardiol 1995;75:86570.
of provocative testing could not be assessed, be- [5] Barolsky SM, Gilbert CA, Faruqui A, et al. Dier-
cause it was performed in only 40.6% of patients. ences in electrocardiographic response to exercise
Current studies on the inclusion of patients of women and men: a non-Bayesian factor. Circula-
with cocaine-related chest pain in a chest pain unit tion 1979;60:10217.
have concluded that a 6- to12-hour observation [6] Douglas PS, Ginsburg GS. The evaluation of chest
pain in women. N Engl J Med 1996;334:13115.
period is safe. In both studies mentioned pre-
[7] Wilson PW, DAgostino RB, Levy D, et al. Predic-
viously, approximately 10% of patients were tion of coronary heart disease using risk factor cate-
admitted to the hospital from the observation gories. Circulation 1998;97:183747.
unit. This result substantiates the utility of in- [8] Lloyd-Jones, Wilson PW, Larson MG, et al. Fra-
clusion of these patients in a chest pain unit, mingham risk score and prediction of lifetime risk
despite the overall low risk of ACS and arrhyth- for coronary heart disease. Am J Cardiol 2004;94:
mia. Because of the general lack of provocative 204.
testing in these studies, it is dicult to determine [9] Bartel AG, Behar VS, Peter RH, et al. Graded ex-
the overall cardiac risk of these patients and ercise stress tests in angiographically documented
therefore dicult to compare the prevalence of coronary artery disease. Circulation 1974;49:
34856.
CAD in these patients with that in a more typical
[10] Diercks DB, Gibler WB, Liu T, et al. Identication
chest pain unit population. of patients at risk by graded exercise testing in an
emergency department chest pain center. Am J Car-
diol 2000;86:28992.
Summary [11] Amsterdam EA, Kirk JD, Diercks DB, et al. Imme-
diate exercise testing to evaluate low risk patients
Chest pain units provide an important alter- presenting to the emergency department with chest
native to traditional hospital admission for pa- pain. J Am Coll Cardiol 2002;40:2516.
tients who present to the ED with symptoms [12] Diercks DB, Kirk JD, Turnipseed SD, Amsterdam
compatible with ACS and a normal or inconclu- EA. Exercise treadmill testing in women evaluated
sive initial evaluation. Although patient sub- in a chest pain evaluation unit [abstract]. Acad
groups such as women, diabetics, those with Emerg Med 2000;8(5):565.
established CAD, and those with symptoms re- [13] Gibbons RJ, Balady GJ, Beasley JW. ACC/AHA
guidelines for exercise testing: executive summary.
lated to stimulant use present unique challenges,
A report of the American College of Cardiology/
management in a chest pain unit appears to be American Heart Association Task Force on Practice
appropriate in these populations. Judicious appli- Guidelines (Committee on Exercise Testing). J Am
cation of accelerated diagnostic protocols and Coll Cardiol 1997;30:260311.
current testing methods can promote safe, accu- [14] Mieres JH, Rosman DR, Shaw LJ. The role of myo-
rate, and cost-eective risk stratication of special cardial perfusion imaging in special populations:
populations to identify patients who can be safely women, diabetics, and heart failure. Semin Nucl
discharged and patients who require hospital Med 2005;35:5281.
admission for further evaluation. [15] Radensky PW, Hilton TC, Fulmer H, et al. Potential
cost eectiveness of initial myocardial perfusion im-
aging for assessment of emergency department
References patients with chest pain. Am J Cardiol 1997;79:
5959.
[1] American Heart Association. Heart disease and [16] Fesmire FM, Hughes AD, Stout PK, et al. Selective
stroke statisticsd2005 update. Dallas, TX: Ameri- dual nuclear scanning in low-risk patients with
can Heart Association; 2005. chest pain to reliably identify and exclude acute
[2] Daida H, Allison TG, Johnson BD, et al. Compari- coronary syndromes. Ann Emerg Med 2001;38:
son of peak exercise oxygen uptake in men versus 20715.
556 DIERCKS et al

[17] Kontos MC, Jesse RL, Schmidt KL, et al. Value of [29] Zellweger MJ, Dubois EA, Lai S, et al. Risk strat-
acute rest sestamibi perfusion imaging for evalua- ication in patients with remote prior myocardial
tion of patients admitted to the emergency depart- infarction using rest-stress myocardial perfusion
ment with chest pain. J Am Coll Cardiol 1997;30: SPECT: prognostic value and impact on referral
97682. to early catheterization. J Nucl Cardiol 2002;9:
[18] Kontos MC, Jesse RL, Anderson FP, et al. Compar- 2332.
ison of myocardial perfusion imaging and cardiac [30] Elhendy A, Mahoney DW, Khandheria BK. Prog-
troponin I in patients admitted to the emergency nostic signicance of impairment of heart rate re-
department with chest pain. Circulation 1999;99: sponse to exercise: impact of left ventricular
20738. function and myocardial ischemia. J Am Coll Car-
[19] Santana-Boado C, Candell-Riera J, Castell-Conesa diol 2003;42:82330.
J. Diagnostic accuracy of technetium-99m-MIBI [31] Wander GS, Pasricha S, Aslam N, et al. Should beta-
myocardial SPECT in women and men. J Nucl blockers be withdrawn in post-myocardial infarction
Med 1998;39:7515. patients before treadmill test? Indian Heart J 1997;
[20] Taillefer R, DePuey EG, Udelson JE. Comparative 49:5036.
diagnostic accuracy of Tl-201 and Tc-99m sestamibi [32] Ades PA, Thomas JD, Hanson JS, et al. Eect of
SPECT imaging (perfusion and ECG-gated SPECT) metoprolol on the submaximal stress test performed
in detecting coronary artery disease in women. J Am early after acute myocardial infarction. Am J Car-
Coll Cardiol 1997;29:6977. diol 1987;60:9636.
[21] Hachamovitch R, Berman DS, Kiat H, et al. Eec- [33] Ashmore RC, Corkadel LK, Green CL, Horwitz
tive risk stratication using exercise myocardial per- LD. Verapamil but not nifedipine impairs left ven-
fusion SPECT in women: gender-related dierences tricular function during exercise in hypertensive
in prognostic nuclear testing. J Am Coll Cardiol patients. Am Heart J 1990;119:63641.
1996;28:3444. [34] Ho SW, McComish MJ, Taylor RR. Eect of beta-
[22] Hachamovitch R, Hayes S, Friedman JD, et al. adrenergic blockade on the results of exercise testing
Determinants of risk and its temporal variation in related to the extent of coronary artery disease. Am J
patients with normal stress myocardial perfusion Cardiol 1985;55:25862.
scans: what is the warranty period of a normal [35] Pellinen TJ, Virtanen KS, Valle M, Frick MH.
scan? J Am Coll Cardiol 2003;41:132940. Studies on ergometer exercise testing. II. Eect of
[23] Mieres JH, Shaw LJ, Hendel RC, et al. Ameri- previous myocardial infarction, digoxin, and beta-
can Society of Nuclear Cardiology consensus blockade on exercise electrocardiography. Clin Car-
statement: Task Force on Women and Coronary diol 1986;9:499507.
Artery Diseasedthe role of myocardial perfusion [36] Herbert WG, Dubach P, Lehmann KG, Froelicher
imaging in the clinical evaluation of coronary ar- VF. Eect of beta-blockade on the interpretation
tery disease in women [correction]. J Nucl Car- of the exercise ECG: ST level versus delta ST/HR in-
diol 2003;10:95101. dex. Am Heart J 1991;122:9931000.
[24] Brush JE Jr, Brand DA, Acampora D, et al. Use of [37] Martin GJ, Henkin RE, Scanlon PJ. Beta blockers
the initial electrocardiogram to predict in-hospital and the sensitivity of the thallium treadmill test.
complications of acute myocardial infarction. Chest 1987;92:4867.
N Engl J Med 1985;312:113741. [38] Sweeney ME, Fletcher BJ, Fletcher GF. Exercise
[25] Goldman L, Cook EF, Brand DA, et al. A computer testing and training with beta-adrenergic block-
protocol to predict myocardial infarction in emer- ade: role of the drug washout period in unmask-
gency department patients with chest pain. N Engl ing a training eect. Am Heart J 1989;118:
J Med 1988;318:797803. 9416.
[26] Farkouh ME, Smars PA, Reeder GS, et al. A clinical [39] Diercks DB, Kirk JD, Turnipseed SD, Amsterdam
trial of a chest-pain observation unit for patients EA. Utility of immediate exercise treadmill testing
with unstable angina. Chest Pain Evaluation in the in patients taking beta-blockers or calcium channel
Emergency Room (CHEER) Investigators. N Engl blockers. Am J Cardiol 2002;90:8825.
J Med 1998;339:18828. [40] Taillefer R, Ahlberg AW, Masood Y, et al. Acute
[27] Lewis WR, Amsterdam EA, Turnipseed S, Kirk JD. beta-blockade reduces the extent and severity of
Immediate exercise testing of low risk patients with myocardial perfusion defects with dipyridamole
known coronary artery disease presenting to the Tc-99m sestamibi SPECT imaging. J Am Coll Car-
emergency department with chest pain. J Am Coll diol 2003;42:147583.
Cardiol 1999;33:18437. [41] Marwick TH, Case C, Sawada S. Use of stress echo-
[28] Diercks DB, Kirk JD, Amsterdam EA. Evaluations cardiography to predict mortality in patients with di-
of patients with a history of coronary artery disease abetes and known or suspected coronary artery
in a chest pain unit [abstract]. 5th National Con- disease. Diabetes Care 2002;25:10428.
gress, Society of Chest Pain Centers and Providers. [42] Bacci S, Villella M, Villella A, et al. Screening for si-
New Haven, Connecticut, 2002. lent myocardial ischemia in type 2 diabetic patients
CHEST PAIN UNITS AND SPECIAL POPULATIONS 557

with additional atherogenic risk factors: applicability [48] Weber JE, Shofer FS, Larkin GL, et al. Validation
and accuracy of the exercise stress test. Eur J Endo- of a brief observation period for patients with co-
crinol 2002;147:64954. caine-associated chest pain. N Engl J Med 2003;
[43] Bell DS, Yumuk VD. Low incidence of false-positive 348:5107.
exercise thallium-201 scintigraphy in a diabetic pop- [49] Lange RA, Hillis RD. Cardiovascular complications
ulation. Diabetes Care 1996;19:1836. of cocaine use. N Engl J Med 2001;345:3518 [Erra-
[44] Kang X, Berman DS, Lewin H, et al. Comparative tum: N Engl J Med 2001;345:1432].
ability of myocardial perfusion single-photon emis- [50] Mittleman MA, Mintzer D, Maclure M, et al. Trig-
sion computed tomography to detect coronary ar- gering of myocardial infarction by cocaine. Circula-
tery disease in patients with and without diabetes. tion 1999;99:273741.
Am Heart J 1999;137:94957. [51] Quershi AI, Suri MF, Guterman LR, et al. Cocaine
[45] Berman DS, Kang X, Hayes SW, et al. Adenosine use and the likelihood of nonfatal myocardial infarc-
myocardial perfusion single photon emission com- tion and stroke: data from the Third National
puted tomography in women compared to men. Im- Health and Nutrition Examination Survey. Circula-
pact of diabetes in incremental prognostic value and tion 2001;103:5026.
eect on patient management. J Am Coll Cardiol [52] Weber JE, Chudnofsky CR, Boczar M, et al.
2003;41:112533. Cocaine-associated chest pain: how common is
[46] Diercks DB, Kirk JD, Onesko N, Amsterdam EA. myocardial infarction? Acad Emerg Med 2000;7:
Patients with diabetes can undergo immediate exer- 8737.
cise treadmill testing in a chest pain evaluation [53] Turnipseed SD, Richards JR, Kirk JD, et al. Fre-
unit. Acad Emerg Med 2003;10:540A. quency of acute coronary syndrome in patients pre-
[47] Sanchez CD, Newby LK, Hasselblad V, et al. Com- senting to the emergency department with chest pain
parison of 30-day outcome, resource use, and coro- after methamphetamine use. J Emerg Med 2003;24:
nary artery disease severity in patients with 36973.
suspected coronary artery disease with and without [54] Kushman SO, Storrow AB, Liu T, Gibler WB, et al.
diabetes mellitus assigned to chest pain units. Am J Cocaine-associated chest pain in a chest pain center.
Cardiol 2003;91:122830. Am J Cardiol 2000;85:3946.
Cardiol Clin 23 (2005) 559568

Management of the Patient with Chest Pain


and a Normal Coronary Angiogram
Eric H. Yang, MD, Amir Lerman, MD*
The Center of Coronary Physiology and Imaging, Division of Cardiovascular Diseases,
Mayo College of Medicine, 200 First Street Southwest, Rochester, Minnesota 55905, USA

Approximately 20% to 30% of patients un- [5]. The endothelium plays an important role in
dergoing coronary angiography for symptomatic the regulation of vasomotor tone and CBF. This
chest pain are found to have normal epicardial regulation occurs by way of the production and
coronary arteries [1,2]. When compared with pa- release of vasoactive factors. The important vaso-
tients who present with obstructive coronary artery dilators are:
disease, these patients are more likely to be women
1. Nitric oxide: Nitric oxide (NO) is a potent
and tend to be younger [3,4]. Several conditions
vasodilator. Its eects are mediated by the
can result in chest pain with a normal coronary an-
secondary messenger cyclic 3#5-guanosine
giogram (NOCAD) and a proper diagnosis of the
monophosphate [6]. It has a short half-life
etiology is essential in managing these patients.
and is synthesized from the oxidation of the
This article discusses the pathophysiology of
N-terminal of arginine [7]. This reaction is
NOCAD and provides a systematic diagnostic ap-
catalyzed by the enzyme endothelial nitric
proach to these patients. Potential therapeutic op-
oxide synthase, which is produced by endo-
tions and prognosis are also reviewed.
thelial cells [7].
2. Prostacyclin I: Prostacyclin I2 causes coronary
Pathophysiology vasodilation by increasing the production
of cyclic 3#5-adenosine monophosphate in
Angina occurs when there is a mismatch be- platelets and smooth muscle cells [8]. It is de-
tween myocardial oxygen supply and demand. In rived from arachidonic acid by the sequential
the absence of signicant coronary artery stenosis, reactions of cyclooxygenase and prostacyclin
coronary blood ow (CBF) is regulated and synthase [9].
limited by two main factors: coronary endothelial
function and microvascular function. The most potent vasoconstricting agents pro-
Endothelial function duced by the endothelium are:

Traditionally, the coronary endothelium has 1. Endothelin-1: Endothelin-1 is produced in


been thought of as a monolayer of endothelial endothelial cells from the cleavage of a pro-
cells that line the lumen of the vascular bed; peptide by endothelin converting enzymes
however, it is now known that the endothelium I and II [1012]. Vasoconstriction occurs
also extends into the vascular wall and adventitia when endothelin-1 binds to the ETA-receptor
located on smooth muscle cells [13].
2. Thromboxane A2: Thromboxane A2 is de-
Dr. Lerman is supported by NIH grants R01 HL
rived from arachidonic acid by way of a
63911 and K24 HL 69840. cyclooxygenase catalyzed reaction. It stimu-
* Corresponding author. lates vasoconstriction by binding to the
E-mail address: lerman.amir@mayo.edu thromboxane receptor located on vascular
(A. Lerman). smooth muscle cells [14].
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.002 cardiology.theclinics.com
560 YANG & LERMAN

If the endothelium does not function properly, intracoronary Doppler wire or by way of trans-
NOCAD may occur as a result of a mismatch thoracic echocardiography. Patients who present
between myocardial oxygen supply and demand. with endothelial-independent microvascular dys-
In addition to NOCAD, endothelial dysfunction function have a CFR less than or equal to 2.5.
also promotes atherosclerosis and the eventual
development of coronary artery stenosis.
Dierential diagnosis
The dierential diagnosis of NOCAD can be
Microvascular function classied anatomically into three categories: (1)
The coronary microcirculation consists of the epicardial disease, (2) coronary microvascular dys-
small arteriolar vessels that are less than 300 mm function, and (3) noncoronary disease (Table 1).
in diameter. These vessels determine 80% of the Epicardial disease resulting in NOCAD includes
coronary resistance and therefore play an impor- endothelial dysfunction, coronary artery spasm,
tant role in the regulation of CBF. Four factors and coronary artery bridging. Microvascular dys-
regulate the microcirculation: function may be secondary to endothelial dysfunc-
tion, hypertension, cardiomyopathy, inltrative
1. Myogenic control: The pressure exerted from disease, valvular disease, or idiopathic. Noncoro-
the surrounding myocardium inuences the nary artery disease involving other organs systems,
intraluminal area of the microcirculation. such as the pulmonary, gastrointestinal, or muscu-
Myogenic control of the microcirculation al- loskeletal systems, can also result in NOCAD.
lows for the optimal transport of metabolic
substances across capillary membranes by
maintaining intraluminal pressure within Epicardial disease
a physiologic range [15].
2. Flow-mediated control: This process is en- Endothelial dysfunction
dothelium-dependent allowing for auto- Proper endothelial function is necessary to
regulation of microvascular blood ow [16]. insure that CBF is adequate to meet myocardial
Flow-mediated control helps to maintain oxygen demand and endothelial dysfunction can
intracoronary pressure and prevent shear result in NOCAD. Intracoronary acetylcholine
stress-mediated injury. can be used to assess endothelial function [18,19].
3. Metabolic control: The metabolic state of the In patients who have normal endothelial function,
myocardium plays an important role in the the smooth muscle contraction mediated by ace-
regulation of blood ow. Oxygen consump- tylcholine is counterbalanced by its stimulation
tion is the main regulating factor and me-
diators, such as adenosine, prostacyclin,
Table 1
norepinephrine, and carbon dioxide are Dierential diagnosis of chest pain in the setting of a nor-
involved [17]. mal coronary angiogram
4. Neurohormonal control: Sympathetic and
I. Coronary disease
parasympathetic innervation of the microcir-
A. Epicardial disease
culation allows for neural regulation of CBF. 1. Endothelial dysfunction
Neurotransmitters, such as acetylcholine, 2. Coronary spasm
norepinephrine, and neuropeptide Y, help 3. Coronary bridging
mediate the response [17]. B. Microvascular dysfunction
1. Microvascular endothelial dysfunction
If these control mechanisms do not function
2. Hypertension
properly, microvascular dysfunction occurs. A 3. Cardiomyopathy
reduction in the coronary ow reserve (CFR) in 4. Inltrative disease
the absence of epicardial coronary artery stenosis 5. Valvular disease
indicates the presence of endothelial-independent 6. Idiopathic
microvascular dysfunction. CFR is the ratio of the II. Noncoronary disease
average peak velocity (APV) during maximal A. Gastrointestinal
hyperemia to the APV at baseline. Intracoronary B. Pulmonary
or intravenous adenosine can be used to achieve C. Musculoskeletal
D. Psychologic
hyperemia, and APV can be measured with an
CHEST PAIN & NOCAD MANAGEMENT 561

of NO production in endothelial cells. According spasm also can occur at the site of myocardial
to previous studies, epicardial endothelial dys- bridging and may result in myocardial ischemia
function is dened as a reduction in coronary ar- and coronary artery thrombosis [33].
tery diameter greater than 20% in response to
acetylcholine [20,21].
Microvascular disease
Coronary spasm
Microvascular endothelial dysfunction
Angina caused by coronary spasm was rst
Impaired microvascular endothelial function
described in 1959 by Prinzmetal and colleagues
can result in NOCAD by disrupting autoregula-
[22]. The chest pain induced by coronary spasm is
tion of CBF. Previous studies have dened
similar to the pain caused by coronary artery dis-
microvascular endothelial dysfunction as an in-
ease but tends to occur at rest and between mid-
crease in CBF less than 50% in response to
night and 8:00 AM [23]. Each episode can last
acetylcholine [20,21].
from 30 to 60 minutes and may be associated
with ischemic changes on the ECG. Although the
mechanism of disease is not completely known, it Hypertension and microvascular dysfunction
seems to involve endothelial dysfunction and an in-
Patients who present with hypertension can
creased response to vasoconstrictor agents, such as
present with NOCAD. The mechanism of chest
catecholamines and thromboxane A2 [24,25]. Evi-
pain in these patients may involve a decrease in
dence also exists that an increase in autonomic
CBF caused by an increase in microvascular
tone may play a role in coronary spasm [26].
resistance [34,35]. The increase in resistance may
Intracoronary acetylcholine can be used as
be a result of an increase in myogenic tone caused
a provocative test for the diagnosis of coronary
by elevated diastolic pressures or compression of
spasm [27]. Other direct vasoconstricting agents,
the microcirculation by the hypertrophied myo-
such as ergonovine maleate, also can be used
cardium. A previous study suggests that ventricu-
[28]. Although many patients have some degree
lar hypertrophy plays an important role in
of vasoconstriction to these agents, spasm is de-
microvascular dysfunction because hypertensive
ned as a greater than 50% focal reduction in lu-
patients without ventricular hypertrophy did not
men caliber that is reversed with intracoronary
have a reduction in CFR [36].
nitroglycerin [29].

Myocardial bridging Cardiomyopathy and microvascular dysfunction


Systolic compression of a coronary artery by the NOCAD occurs in approximately 50% of pa-
surrounding myocardium is referred to as myocar- tients who present with dilated cardiomyopathy
dial bridging. It was rst described in 1737 by [37] and studies have shown that these patients
Reyman [30] and occurs when a segment of the cor- have microvascular dysfunction and a reduced
onary artery is tunneled in the myocardium. Bridg- CFR [38,39]. The mechanism of microvascular
ing most frequently occurs in the left anterior dysfunction most likely involves myogenic com-
descending artery, but reports have shown that it pression secondary to elevated lling pressures
may involve any of the major epicardial coronary or microvascular endothelial dysfunction. Patients
arteries. The prevalence of myocardial bridging in who present with hypertrophic cardiomyopathy
autopsy studies has varied from 5% to 86%, but may also have a reduced CFR and NOCAD
angiographic studies have shown a lower rate of [40,41]. Histologic studies suggest that microvas-
0.5% to 33% [31]. Although the mechanism is cular dysfunction in these patients may be caused
not clear, myocardial bridging can cause NOCAD. by a reduction in the number of small arterioles
Intravascular ultrasound studies have shown that and a narrowing of the luminal area of the micro-
compression can extend into early diastole and re- circulation [42,43]. An alternative mechanism for
duce CBF [32]. Intracoronary hemodynamic stud- the reduction in CFR may be that the microcircu-
ies have also shown that bridging results in lation is already near maximally dilated in the
a greater dependence on diastolic CBF [32]. Tachy- basal state to meet the increase in oxygen demand
cardia can therefore decrease the diastolic time pe- [44]. Because the vessels are already dilated, their
riod and cause a decrease in CBF and ischemia. ability to further dilate is limited and the CFR is
Endothelial dysfunction and coronary artery reduced.
562 YANG & LERMAN

Inltrative disease and microvascular dysfunction


Patients who present with cardiac amyloidosis
can present with NOCAD [45] caused by endothe-
lial dysfunction [46] and amyloid deposits in the
tunica media [47]. These deposits cause a reduction
in the luminal area of the microcirculation [4850]
and the subsequent ischemia may be one of the
factors responsible for the sudden cardiac death
that occurs in cardiac amyloidosis.

Valvular disease and microvascular dysfunction


Fig. 1. Functional angiogram protocol to assess coro-
NOCAD can occur in patients who develop
nary endothelial and microvascular function. Ach, ace-
aortic stenosis, aortic regurgitation, and mitral tylcholine; IC, intracoronary; NTG, nitroglycerin.
regurgitation [5154]. Previous studies have
shown that these patients have a reduction in has been described previously [20,6264] and a
CFR that may be caused by an increase in lling summary is as follows:
pressures and wall stress [52,5557]. These factors After coronary angiography, a guiding cathe-
increase the myogenic tone and the amount of mi- ter is placed into the left main coronary artery. A
crovascular compression. Surgical treatment of 0.014-inch Doppler guide wire is placed within
the underlying valvular disease seems to correct a 2.2-F coronary infusion catheter and the system
the microvascular dysfunction in these patients is advanced through the guiding catheter into the
[52,58,59]. middle portion of the left anterior descending
coronary artery. The Doppler wire is then posi-
Idiopathic microvascular dysfunction tioned 2 to 3 mm distal to the tip of the infusion
Patients who present with no apparent etiology catheter, and a baseline APV is obtained. Coro-
for microvascular dysfunction have idiopathic nary artery diameter is measured 5 mm distal to
disease. Syndrome X has been used to describe the tip of the Doppler wire and baseline CBF is
these patients, but the term is not specic and calculated with the following equation: CBF p
probably should not be used. The mechanism of (coronary artery diameter/2) [2]  (APV/2). Mi-
disease is not known but may involve an impaired crovascular function is assessed with the use of
vasodilator response as a result of smooth muscle an intracoronary adenosine (1842 mg) bolus.
cell dysfunction and primary microvascular endo- CFR is calculated as the ratio of the APV during
thelial dysfunction [60]. Increased activity of the maximal hyperemia to the APV at baseline. The
sodium-hydrogen ion channel also may play authors dene a normal CFR as greater than
a role in idiopathic microvascular disease [61]. 2.5. Endothelial function is then assessed with
the use of acetylcholine. Intracoronary acetylcho-
Noncoronary etiologies line is infused through the infusion catheter at
three dierent doses: 106, 105, and 104 M
The noncoronary etiologies of NOCAD are for 3 minutes each to achieve intracoronary con-
shown in Table 1. Although these causes of NO- centrations of 108, 107, and 106 respectively.
CAD are not discussed in this review, the clinician Two hundred micrograms of intracoronary nitro-
should be aware of noncardiac disease when eval- glycerin is then given at the end of the procedure.
uating patients who present with NOCAD. APV and coronary diameter is measured before
and after each infusion. Based on prior studies,
Systematic approach to diagnosing the etiology an abnormal response to acetylcholine is an in-
of normal coronary angiogram crease in CBF less than or equal to 50% or a re-
duction in epicardial coronary artery diameter
Functional angiogram
greater than or equal to 20% [20,62].
Determination of the etiology of NOCAD is
Interpretation of results
essential in its management. A functional angio-
gram (Fig. 1) involving the invasive assessment Epicardial stenosis and myocardial bridging can
of coronary physiology allows for a systematic di- be ruled out by carefully reviewing the diagnostic
agnostic approach to these patients. The protocol angiogram. The functional angiogram then can be
CHEST PAIN & NOCAD MANAGEMENT 563

used to classify patients into one of four groups shown to be associated with an increase in myo-
(Fig. 2). Those patients who experience an abnor- cardial infarction, coronary revascualrization, and
mal response to acetylcholine but a normal CFR cardiac death [20]. Fortunately, endothelial dys-
have abnormal endothelial function. If there is a fo- function is reversible if the proper therapy is initi-
cal greater than 50% reduction in luminal caliber ated. Risk factor modication is the cornerstone
during the infusion of acetylcholine then coronary of therapy and previous studies have shown im-
spasm is present. Patients who present with a nor- provement in endothelial function with exercise,
mal response to acetylcholine but an abnormal weight loss, and smoking cessation [65,66]. Blood
CFR have endothelial-independent microvascular pressure and cholesterol control are also essential.
dysfunction. These patients probably should un- Angiotensin-converting enzyme inhibitors (Table 2)
dergo echocardiography to help determine the eti- have been shown to improve endothelial function
ology of their microvascular dysfunction. An [6770]. The mechanisms are unknown but seem
abnormal response to acetylcholine and a reduced to be independent of their blood pressure eects.
CFR indicates the presence of endothelial dysfunc- The benets of 3-hydroxy-3-methylglutaryl coen-
tion and endothelial-independent microvascular zyme A reductase inhibitors (statins) on endo-
dysfunction. Finally, those patients who experience thelial function seem to be independent of their
a normal response to acetylcholine and a normal cholesterol lowering eects and may involve their
CFR are most likely to have a noncoronary etiol- antioxidant and anti-inammatory properties
ogy of NOCAD. These patients should be evalu- [71]. Initial studies with peroxisome-activated re-
ated for other etiologies of chest pain, such as ceptor-g agonists, such as rosiglitazone, have
gastrointestinal, pulmonary, or musculoskeletal shown some benecial eects on endothelial func-
disease. The distribution of ndings from 820 func- tion [72]. Finally, other therapies, such as L-argi-
tional angiograms performed at the Mayo Clinic nine and folic acid, have also been shown to
are summarized in Fig. 3. improve endothelial function [73,74].
Coronary spasm
Prognosis and treatment
During the initial active phase (rst 6 months)
Endothelial dysfunction
patients experience frequent episodes of angina
Coronary endothelial dysfunction is a marker and are at increased risk for adverse cardiac
of early coronary atherosclerosis and has been events; however, long-term studies have shown

IC Adenosine

CFR > 2.5 _ 2.5


CFR <

ACH ACH
abnormal response normal response normal response abnormal response

Endothelial
Non-Endothelial Dysfunction and
Endothelial Non-Cardiac
Microvascular Non-Endothelial
Dysfunction Chest Pain
Dysfunction Microvascular
Dysfunction

Fig. 2. Algorithm for a systematic diagnostic approach to patients who present with chest pain and NOCAD. An abnormal
response to acetylcholine is dened as an increase in CBF less than 50% (microvascular endothelial dysfunction) or less than
a 20% increase in coronary artery diameter (epicardial endothelial dysfunction). Ach, acetylcholine; CFR, coronary ow
reserve; IC, intracoronary. (Data from Al Suwaidi J, Higano ST, Holmes DR Jr, et al. Pathophysiology, diagnosis, and cur-
rent management strategies for chest pain in patients with normal ndings on angiography. Mayo Clin Proc 2001;76:813.)
564 YANG & LERMAN

Fig. 3. Distribution of results from 820 functional angiograms performed at the Mayo Clinic.

that after the active phase, the 5-year survival is Myocardial bridging
excellent and there is no increase in cardiac events
Patients who present with myocardial bridging
[75,76]. The rst line of therapy for patients who
have been shown to have a good long-term
present with coronary spasm (see Table 2) is nondihy-
prognosis [83]. Pediatric patients who present
dropyridine calcium channel blockers and long-
with hypertrophic obstructive cardiomyopathy
acting nitrates [77]. Nonselective beta-blocking
and myocardial bridging, however, may have an
agents, such as propranolol, should be avoided
increased risk for death and adverse cardiac
because of their potential to exacerbate spasm
events [84]. This increased risk does not seem to
by way of unopposed alpha-1 action [78]. Aspirin
occur in adult patients who have developed hyper-
also should be used with caution because of its po-
trophic obstructive cardiomyopathy and myocar-
tential to inhibit the production of vasodilating
dial bridging [85].
prostacyclin derivatives [79]. Additional therapy
All patients who develop NOCAD as a result
with nifedipine and alpha-blocking agents can be
of myocardial bridging should be treated with
used in patients refractory to rst-line therapy
beta-blockers [86]. Those patients who do not
[80,81]. Finally, when medical therapy becomes
improve with medical therapy can be treated
unsuccessful, placement of coronary artery stents
with more invasive procedures, such as surgical
or bypass surgery can be considered [82].

Table 2
Treatment strategies for patients with chest pain and a normal coronary angiogram
Etiology First Line Therapy Second Line Therapy Comments
Endothelial Dysfunction Life style modication, L-arginine, folate PPAR-g agents may oer
ACEI, Statins benet
Coronary Spasm Non-dihydropyridine Nifedipine, alpha-blockers, Non-selective beta blockers
calcium channel blockers, coronary artery stenting, should be avoided,
long-acting nitrates bypass surgery aspirin may exacerbate
spasm
Myocardial Bridging Beta-blockers Coronary artery stenting, Higher rates of in-stent
surgical myotomy, bypass restenosis occurs in
surgery bridging segment
Endothelial-Independent Treatment of underlying Beta-blockers, Statins Imipramine may oer
Microvascular etiology of microvascular benets in idiopathic
Dysfunction dysfunction microvascular
dysfunction
Abbreviations: ACEI, angiotensin converting enzyme inhibitor; PPAR-g, peroxisome-activated receptor-g agonist.
CHEST PAIN & NOCAD MANAGEMENT 565

myotomy, coronary artery stenting, and bypass to assess endothelial and microvascular function.
surgery [8789]. Stents placed in coronary seg- Once the etiology of chest pain is determined, the
ments with bridging have a higher restenosis rate appropriate therapy can be initiated.
than those placed in segments without bridging
[90].

Microvascular dysfunction References


Little is known about the long-term prognosis [1] Kemp H Jr, Vokonas P, Cohn P, et al. The anginal
of patients who experience microvascular dysfunc- syndrome associated with normal coronary arterio-
tion and otherwise normal coronary arteries. grams. Report of a six year experience. Am J Med
Patients who develop hypertrophic cardiomyopa- 1973;54:73542.
thy and microvascular dysfunction seem to have [2] Proudt W, Shirey E, Sones FJ. Selective cine coro-
an increase in adverse clinical events [91]. The in- nary arteriography. Correlation with clinical nd-
teraction between microvascular dysfunction and ings in 1,000 patients. Circulation 1966;33:90110.
[3] Hasdai DH, Higano DR Jr, Burnett ST, et al. Prev-
dilated cardiomyopathy is not known. Treatment
alence of coronary blood ow reserve abnormalities
of the underlying etiology of microvascular dys- among patients with nonobstructive coronary artery
function is the focus of therapy. Blood pressure disease and chest pain. Mayo Clin Proc 1998;73:
control in those with hypertension and reduction 113340.
of lling pressures and wall stress in patients [4] Papanicolaou MN, Cali RM, Hlatky MA, et al.
who present with dilated cardiomyopathy im- Prognostic implications of angiographically normal
proves microvascular function. Valvular replace- and insignicantly narrowed coronary arteries. Am
ment or repair has been shown to improve J Cardiol 1986;58:11817.
microvascular function in patients who have val- [5] Herrmann J, Lerman LO, Rodriguez-Porcel M, et al.
vular disease [52,58,59]. In patients who develop Coronary vasa vasorum neovascularization precedes
epicardial endothelial dysfunction in experimental
hypertrophic obstructive cardiomyopathy, the ef-
hypercholesterolemia. Cardiovasc Res 2001;51:
fect of myectomy and septal ablation on micro- 7626.
vascular function is not known. [6] Rapoport RM, Draznin MB, Murad F. Endo-
Numerous therapies have been investigated for thelium-dependent relaxation in rat aorta may be
the treatment of patients who present with idio- mediated through cyclic GMP-dependent protein
pathic microvascular dysfunction (see Table 2). phosphorylation. Nature 1983;306:1746.
The ecacy of these therapies is dicult to deter- [7] Palmer RMJ, Ashton DS, Moncada S. Vascular en-
mine because of the heterogeneous population of dothelial cells synthesize nitric oxide from L-argi-
patients. Beta-blockade seems to reduce angina nine. Nature 1988;333:6646.
symptoms, but no large randomized studies have [8] Moncada S, Vane VR. Pharmacology and endoge-
nous roles of prostaglandin endoperoxides, throm-
been performed. Treatment with angiotensin con-
boxane A2 and prostacyclin. Pharmacol Rev 1979;
verting enzyme inhibitors and HMG-CoA reduc- 30:293.
tase inhibitors also result in some benet [92,93]. [9] Vane JR, Botting RM. Pharmacodynamic prole of
Imipramine, a tricyclic antidepressant, has been prostacyclin. Am J Cardiol 1995;75:3A10A.
shown to reduce anginal pain in patients who de- [10] Yanagisawa M, Kurihara H, Kimura S, et al. A
velop idiopathic microvascular dysfunction [94]. novel potent vasoconstrictor peptide produced by
The mechanism is not understood completely but vascular endothelial cells. Nature 1988;332:4115.
may involve inhibition of pain-modulating neu- [11] Ohnaka K, Takayanagi R, Nishikawa M, et al. Pu-
rons by blockade of norepinephrine reuptake [95]. rication and characterization of a phosphora-
midon-sensitive endothelin-converting enzyme in
porcine aortic endothelium. J Biol Chem 1993;268:
Summary 2675966.
Approximately 20% to 30% of patients who [12] Xu D, Emoto N, Giaid A, et al. ECE-1: a membrane-
experience chest pain who undergo coronary bound metalloprotease that catalyzes the proteolytic
activation of big endothelin-1. Cell 1994;78:47385.
angiography are found to have normal epicardial
[13] Arai H, Hori S, Aramori I, et al. Cloning and expres-
coronary arteries. The management of these sion of a cDNA encoding an endothelin receptor.
patients can be challenging and a correct diagno- Nature 1990;348:7302.
sis of the etiology is essential. A systematic [14] Luscher TF, Noll G. The pathogenesis of cardiovas-
approach to diagnosing the cause of chest pain cular disease: role of the endothelium as a target and
can be accomplished with a functional angiogram mediator. Atherosclerosis 1995;118:S8190.
566 YANG & LERMAN

[15] Davis MJ. Microvascular control of capillary pres- of the American College of Cardiology/American
sure during increases in local arterial and venous Heart Association Task Force on Practice Guide-
pressure. Am J Physiol 1988;254:H77284. lines (Committee on Coronary Angiography) devel-
[16] Pohl U, Holtz J, Busse R, et al. Crucial role of endo- oped in collaboration with the Society for Cardiac
thelium in the vasodilator response to increased ow Angiography and Interventions. J Am Coll Cardiol
in vivo. Hypertension 1986;8:3744. 1999;33:1756824.
[17] Komaru T, Kanatsuka H, Shirato K. Coronary mi- [30] Reyman H. Disertatio de vasis cordis propriis. Med
crocirculation: physiology and pharmacology. Phar- Diss Univ Gottingen 1737;7:132.
macol Ther 2000;86:21761. [31] Mohlenkamp S, Hort W, Ge J, et al. Update on
[18] el-Tamimi H, Mansour M, Wargovich T, et al. Con- myocardial bridging. Circulation 2002;106:261622.
strictor and dilator responses to intracoronary ace- [32] Ge J, Jeremias A, Rupp A, et al. New signs charac-
tylcholine in adjacent segments of the same teristic of myocardial bridging demonstrated by
coronary artery in patients with coronary artery dis- intracoronary ultrasound and Doppler. Eur Heart
ease. Endothelial function revisited. Circulation J 1999;20:170716.
1994;89:4551. [33] Herrmann J, Higano ST, Lenon RJ, et al. Myocar-
[19] Egashira KIT, Hirooka Y, Yamada A, et al. Im- dial bridging is associated with alteration in coro-
paired coronary blood ow response to acetylcho- nary vasoreactivity. Eur Heart J 2004;25:213442.
line in patients with coronary risk factors and [34] Opherk D, Mall G, Zebe H, et al. Reduction of cor-
proximal atherosclerotic lesions. J Clin Invest 1993; onary reserve: a mechanism for angina pectoris in
91:2937. patients with arterial hypertension and normal coro-
[20] Suwaidi JA, Hamasaki S, Higano ST, et al. Long- nary arteries. Circulation 1984;69:17.
term follow-up of patients with mild coronary artery [35] Brush J, Cannon R, Schenke W, et al. Angina due
disease and endothelial dysfunction. Circulation to coronary microvascular disease in hypertensive
2000;101:94854. patients without left ventricular hypertrophy. N
[21] Hasdai D, Holmes DR Jr, Higano ST, et al. Preva- Engl J Med 1988;319:13027.
lence of coronary blood ow reserve abnormalities [36] Hamasaki S, Al Suwaidi J, Higano ST, et al. Atten-
among patients with nonobstructive coronary artery uated coronary ow reserve and vascular remodel-
disease and chest pain [see comment]. Mayo Clin ing in patients with hypertension and left
Proc 1998;73:113340. ventricular hypertrophy. J Am Coll Cardiol 2000;
[22] Prinzmetal M, Kennamer R, Merliss R, et al. Angina 35:165460.
pectoris. I. A variant form of angina pectoris: pre- [37] Abelmann WH, Lorell BH. The challenge of cardio-
liminary report. Am J Med 1959;27:375. myopathy. J Am Coll Cardiol 1989;13:121939.
[23] Ogawa H, Yasue H, Oshima S, et al. Circadian var- [38] Pasternac A, Noble J, Streulens Y, et al. Pathophys-
iation of plasma brinopeptide A level in patients iology of chest pain in patients with cardiomyopa-
with variant angina. Circulation 1989;80:161726. thies and normal coronary arteries. Circulation
[24] Cox ID, Kaski JC, Clague JR. Endothelial dysfunc- 1982;65:77889.
tion in the absence of coronary atheroma causing [39] Cannon RO III, Cunnion RE, Parrillo JE, et al. Dy-
Prinzmetals angina. Heart 1997;77:584. namic limitation of coronary vasodilator reserve in
[25] Maseri A, Crea F, Lanza GA. Coronary vasocon- patients with dilated cardiomyopathy and chest
striction: where do we stand in 1999. An important, pain. J Am Coll Cardiol 1987;10:1190200.
multifaceted but elusive role. Cardiologia 1999;44: [40] Maron B, Epstein S, Roberts W. Hypertrophic car-
1158. diomyopathy and transmural myocardial infarction
[26] Lanza GA, Pedrotti P, Pasceri V, et al. Autonomic without signicant atherosclerosis of the extramu-
changes associated with spontaneous coronary ral coronary arteries. Am J Cardiol 1979;43:
spasm in patients with variant angina. J Am Coll 1086102.
Cardiol 1996;28:124956. [41] Cannon RO III, Rosing DR, Maron BJ, Leon MB,
[27] Okumura K, Yasue H, Matsuyama K, et al. Diuse et al. Myocardial ischemia in patients with hypertro-
disorder of coronary artery vasomotility in patients phic cardiomyopathy: contribution of inadequate
with coronary spastic angina. Hyperreactivity to vasodilator reserve and elevated left ventricular ll-
the constrictor eects of acetylcholine and the dila- ing pressures. Circulation 1985;71:23443.
tor eects of nitroglycerin. J Am Coll Cardiol [42] Schwartzkop B, Mundhenke M, Strauer B. Altera-
1996;27:4552. tions of the architecture of subendocardial arterioles
[28] Previtali M, Ardissino D, Storti C, et al. Hyperven- in patients with hypertrophic cardiomyopathy and
tilation and ergonovine tests in Prinzmetals variant impaired coronary vasodilator reserve: a possible
angina: comparative sensitivity and relation with cause for myocardial ischemia. J Am Coll Cardiol
the activity of the disease. Eur Heart J 1989;10: 1998;31:108996.
1014. [43] Tanaka M, Fujiwara H, Onodera T, et al. Quantita-
[29] Scanlon PJ, Faxon DP, Audet AM, et al. ACC/ tive analysis of narrowings of intramyocardial small
AHA guidelines for coronary angiography: a report arteries in normal hearts, hypertensive hearts, and
CHEST PAIN & NOCAD MANAGEMENT 567

hearts with hypertrophic cardiomyopathy. Circula- [58] Hildick-Smith DJR, Shapiro LM. Coronary ow re-
tion 1987;75:11309. serve improves after aortic valve replacement for
[44] Yang EH, Yeo TC, Higano ST, et al. Coronary he- aortic stenosis: an adenosine transthoracic echocar-
modynamics in patients with symptomatic hyper- diography study. J Am Coll Cardiol 2000;36:
trophic cardiomyopathy. Am J Cardiol 2004;94: 188996.
6857. [59] Nemes A, Forster T, Kovacs Z, et al. The eect of
[45] Brandt KS, Cathcart ES, Cohen A. A clinical anal- aortic valve replacement on coronary ow reserve
ysis of the course and prognosis of forty-two in patients with a normal coronary angiogram.
patients with amyloidosis. Am J Med 1968;44: Herz 2002;27:7804.
95569. [60] Chauhan AMP, Taylor G, Petch MC, et al. Both en-
[46] Suwaidi JA, Velianou JL, Gertz MA, et al. Systemic dothelium-dependent and endothelium-independent
amyloidosis presenting with angina pectoris. Ann function is impaired in patients with angina pectoris
Intern Med 1999;131:83841. and normal coronary angiograms. Eur Heart J 1997;
[47] Hongo M, Yamamoto H, Kohda T, et al. Compar- 18:608.
ison of electrocardiographic ndings in patients with [61] Gaspardone A, Ferri C, Crea F, et al. Enhanced ac-
AL (primary) amyloidosis and in familial amyloid tivity of sodium-lithium countertransport in patients
polyneuropathy and anginal pain and their relation with cardiac syndrome X: a potential link between
to histopathologic ndings. Am J Cardiol 2000;85: cardiac and metabolic syndrome X. J Am Coll Car-
84953. diol 1998;32:20314.
[48] Roberts WC, Waller BF. Cardiac amyloidosis caus- [62] Hasdai D, Gibbons RJ, Holmes DR Jr, et al. Coro-
ing cardiac dysfunction: analysis of 54 necropsy nary endothelial dysfunction in humans is associated
patients. Am J Cardiol 1983;52:13746. with myocardial perfusion defects. Circulation 1997;
[49] Satz JE, Sazama K, Roberts WC. Amyloidosis 96:33905.
limited to small arteries causing angina pectoris [63] Vita J, Treasure C, Nabel E, et al. Coronary vasomo-
and sudden death. Am J Cardiol 1983;51:12345. tor response to acetylcholine relates to risk factors
[50] Smith RRL, Hutchins GM. Ischemic heart disease for coronary artery disease. Circulation 1990;81:
secondary to amyloidosis of intramyocardial 4917.
arteries. Am J Cardiol 1979;44:4137. [64] Zeiher A, Drexler H, Wollschlager H, et al. Modula-
[51] Basta LL, Raines D, Najjar S, et al. Clinical, haemo- tion of coronary vasomotor tone in humans. Pro-
dynamic, and coronary angiographic correlates of gressive endothelial dysfunction with dierent early
angina pectoris in patients with severe aortic valve stages of coronary atherosclerosis. Circulation
disease. Br Heart J 1975;37:1507. 1991;83:391401.
[52] Akasaka T, Yoshida K, Hozumi T, et al. Restricted [65] Hambrecht R, Wolf A, Gielen S, et al. Eect of exer-
coronary ow reserve in patients with mitral regurgi- cise on coronary endothelial function in patients
tation improves after mitral reconstructive surgery. with coronary artery disease. N Engl J Med 2000;
J Am Coll Cardiol 1998;32:192330. 342:45460.
[53] Alexopoulos D, Kolovou G, Kyriakidis M, et al. [66] Celermajer D, Sorensen K, Georgakopoulos D,
Angina and coronary artery disease in patients et al. Cigarette smoking is associated with dose-
with aortic valve disease. Angiology 1993;44: related and potentially reversible impairment of
70711. endothelium-dependent dilation in healthy young
[54] Hakki AH, Kimbiris D, Iskandrian AS, et al. Angina adults. Circulation 1993;88:214955.
pectoris and coronary artery disease in patients with [67] ODriscoll G, Green D, Rankin J, et al. Improve-
severe aortic valvular disease. Am Heart J 1980;100: ment in endothelial function by angiotensin convert-
4419. ing enzyme inhibition in insulin-dependent diabetes
[55] Nitenberg A, Foult JM, Antony I, et al. Coronary mellitus. J Clin Invest 1997;100:67884.
ow and resistance reserve in patients with chronic [68] ODriscoll G, Green D, Maiorana A, et al. Improve-
aortic regurgitation, angina pectoris and normal ment in endothelial function by angiotensin-convert-
coronary arteries. J Am Coll Cardiol 1988;11: ing enzyme inhibition in non-insulin-dependent
47886. diabetes mellitus. J Am Coll Cardiol 1999;33:
[56] Marcus ML, Doty DB, Hiratzka LF, et al. De- 150611.
creased coronary reserve: a mechanism for angina [69] Mancini GBJ, Henry GC, Macaya C, et al. Angio-
pectoris in patients with aortic stenosis and normal tensin-converting enzyme inhibition with quinapril
coronary arteries. N Engl J Med 1982;307:13626. improves endothelial vasomotor dysfunction in
[57] Julius BK, Spillmann M, Vassalli G, et al. Angina patients with coronary artery disease: the TREND
pectoris in patients with aortic stenosis and normal (Trial on Reversing Endothelial Dysfunction) Study.
coronary arteries. Mechanisms and pathophysiolog- Circulation 1996;94:25865.
ical concepts [see comment]. Circulation 1997;95: [70] Prasad A, Narayanan S, Husain S, et al. Insertion-
8928. deletion polymorphism of the ace gene modulates
568 YANG & LERMAN

reversibility of endothelial dysfunction with ACE cardiomyopathyda risk factor for sudden death.
inhibition. Circulation 2000;102:3541. N Engl J Med 1998;339:12019.
[71] Bonetti PO, Lerman LO, Napoli C, et al. Statin [85] Sorajja P, Ommen SR, Nishimura RA, et al. Myo-
eects beyond lipid loweringdare they clinically rel- cardial bridging in adult patients with hypertrophic
evant? Eur Heart J 2003;24:22548. cardiomyopathy. J Am Coll Cardiol 2003;42:88994.
[72] Reusch JE, Regensteiner JG, Watson PA. Novel [86] Schwarz ER, Klues HG, vom Dahl J, et al. Func-
actions of thiazolidinediones on vascular function tional, angiographic and intracoronary doppler
and exercise capacity. Am J Med 2003;115:69S74S. ow characteristics in symptomatic patients with
[73] Lerman A, Burnett JC Jr, Higano ST, et al. Long- myocardial bridging: eect of short-term intrave-
term L-arginine supplementation improves small- nous beta-blocker medication. J Am Coll Cardiol
vessel coronary endothelial function in humans. Cir- 1996;27:163745.
culation 1998;97:21238. [87] Klues HG, Schwarz ER, vom Dahl J, et al. Dis-
[74] Doshi SN, McDowell IFW, Moat SJ, et al. Folate turbed intracoronary hemodynamics in myocardial
improves endothelial function in coronary artery bridging: early normalization by intracoronary stent
disease: an eect mediated by reduction of intracel- placement. Circulation 1997;96:290513.
lular superoxide? Arterioscler Thromb Vasc Biol [88] Hillman ND, Mavroudis C, Backer CL, et al. Supra-
2001;21:1196202. arterial decompression myotomy for myocardial
[75] Bory M, Pierron F, Panagides D, et al. Coronary ar- bridging in a child. Ann Thorac Surg 1999;68:2446.
tery spasm in patients with normal or near normal [89] Iversen SHU, Mayer E, Erbel R, et al. Surgical treat-
coronary arteries. Long-term follow-up of 277 ment of myocardial bridging causing coronary ar-
patients [see comment]. Eur Heart J 1996;17: tery obstruction. Scand J Thorac Cardiovasc Surg
101521. 1992;26:10711.
[76] Mark DB, Cali RM, Morris KG, et al. Clinical [90] Haager PK, Schwarz ER, Dahl JV, et al. Long term
characteristics and long-term survival of patients angiographic and clinical follow up in patients with
with variant angina. Circulation 1984;69:8808. stent implantation for symptomatic myocardial
[77] Mayer S, Hillis LD. Prinzmetals variant angina. bridging. Heart 2000;84:4038.
Clin Cardiol 1998;21:2436. [91] Cecchi F, Olivotto I, Gistri R, et al. Coronary micro-
[78] Robertson R, Wood A, Vaughn W, et al. Exacerba- vascular dysfunction and prognosis in hypertrophic
tion of vasotonic angina pectoris by propranolol. cardiomyopathy. N Engl J Med 2003;349:102735.
Circulation 1982;65:2815. [92] Chen JW, Hsu NW, Wu TC, et al. Long-term angio-
[79] Miwa K, Kambara H, Kawai C. Eect of aspirin in tensin-converting enzyme inhibition reduces plasma
large doses on attacks of variant angina. Am Heart J asymmetric dimethylarginine and improves endo-
1983;105:3515. thelial nitric oxide bioavailability and coronary mi-
[80] Goldberg S, Reichek N, Wilson J, et al. Nifedipine in crovascular function in patients with syndrome X.
the treatment of Prinzmetals (variant) angina. Am J Am J Cardiol 2002;90:97482.
Cardiol 1979;44:80410. [93] Pizzi C, Manfrini O, Fontana F, et al. Angiotensin-
[81] Taylor SH. Alpha- and beta-blockade in angina pec- converting enzyme inhibitors and 3-hydroxy-3-
toris. Drugs 1984;28:6987. methylglutaryl coenzyme a reductase in cardiac
[82] Kultursay H, Can L, Payzin S, et al. A rare indica- Syndrome X: role of superoxide dismutase activity.
tion for stenting: persistent coronary artery spasm. Circulation 2004;109:538.
Heart Vessels 1996;11:1658. [94] Cannon RO, Quyyumi AA, Mincemoyer R, et al.
[83] Kramer KH Jr, Proudt WL, Sones FM Jr. Clinical Imipramine in patients with chest pain despite nor-
signicance of isolated coronary bridges: benign and mal coronary angiograms. N Engl J Med 1994;330:
frequent condition involving the left anterior 14117.
descending artery. Am Heart J 1992;103:2838. [95] Dubner R, Bennett GJ. Spinal and trigeminal mech-
[84] Yetman AT, McCrindle BW, MacDonald C, et al. anisms of nociception. Annu Rev Neurosci 1983;6:
Myocardial bridging in children with hypertrophic 381418.
Cardiol Clin 23 (2005) 569588

Using the Emergency Department Clinical Decision


Unit for Acute Decompensated Heart Failure
W. Frank Peacock, MD, FACEP
Department of Emergency Medicine, The Cleveland Clinic, 9500 Euclid Avenue, Cleveland, OH 44195, USA

In the United States, heart failure (HF) man- prognosis. After becoming symptomatic, 2-year
agement costs exceed all other single pathologic mortality is about 35%, and rises to 80% in
entities. Because HF is predominately a disease of men and 65% for women over the next 6 years.
the elderly, and demographic trends are expected After a diagnosis of pulmonary edema, only
to double the at-risk cohort over the next 30 years, 50% survive at 1 year. Up to 85% of those with
the consequences of HF will only increase. More cardiogenic shock are dead within 1 week. Unless
ecacious strategies for early diagnosis, better new strategies are devised, the HF epidemic will
treatment options, and improved clinical out- markedly worsen. Management in an OU repre-
comes are urgently needed. This article reviews sents an opportunity to improve the quality of
the newest options for HF management with care for HF patients.
particular emphasis on care in the emergency HF is a disease of recidivism, characterized by
department (ED) observation unit (OU). frequent clinical exacerbations that prompt ED
visits. Although it may have a pattern of de-
terioration and improvement, the overall course is
Epidemiology and economics that of a steadily deteriorating clinical pattern. As
Disproportionately aecting the elderly, HF is HF slowly progresses, quality of life is eroded by
the only cardiovascular pathology that is increas- frequent ED visits, hospitalizations, and increas-
ing in both incidence and prevalence [1,2]. If under ing disability until death. The frequency of re-
50 years of age, less than 1% of the U.S. popula- cidivism is reected in the ED HF population:
tion is diagnosed with HF, but by age 80, this num- only 21% of patients represent a de novo pre-
ber rises to 10%. This has huge cost implications sentation. The majority of the ED HF population
for our national health system. Inpatient HF costs already carries the HF diagnosis when they arrive
are estimated as high as $23.1 billion and outpa- with acute dyspnea.
tient costs at $14.7 billion, annually. Annual HF Once a patient presents to the ED with acute
hospitalization costs exceed the combination of decompensated HF (ADHF), inpatient admission
costs for breast and lung cancer [3]. is the rule, rather than the exception. An ED visit
In the United States, patients older than 65 is indicative of either severe disease or the
have health insurance provided by the Centers for consequence of inadequate social support. In the
Medical Studies (CMS). Because most HF pa- best of situations, denitive resolution of either is
tients are elderly, statistics from the CMS drive dicult, particularly in the ED environment.
many of HF management decisions. According to Consequently, 80% of HF presentations result
CMS, HF is the most common cause of hospital- in hospitalization.
ization, and it is the most common reason for As discussed previously, CMS provides health
hospital readmission in patients over 65 years of insurance to patients older than 65 in the United
age [3]. Unfortunately, HF has an extremely poor States. Because most HF patients are older than
65, this creates an unusual economic situation
where a specic disease is predominately covered
E-mail address: peacocw@ccf.org by a single third-party payor, and a hospitals
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.014 cardiology.theclinics.com
570 PEACOCK

economic health may be greatly impacted by their of consumption of the DRG payment. In this
reimbursement rules. Inpatient care of HF falls manner, the areas with the greatest cost of care
within the diagnostic related group (DRG) sys- (eg, the ICU) must be carefully used to prevent
tem, with DRG 127 (HF) as the single most economic loss.
expensive diagnosis. In this year alone, there will When global resource use is considered, di-
be more than 1 million HF hospital discharges in agnostic and therapeutic interventions represent
the United States. a small percentage of the cost of HF management.
CMS aects care by reimbursement induce- Rather, LOS, 30-day revisit frequency, and the
ments, where three main parameters drive HF intensity of service occurring within the hospital
management. These include length of stay (LOS), are the main determinants of institutional nan-
revisit frequency, and intensity of service. By this cial success.
method, when a hospital submits patient data,
a diagnosis-based algorithm is used to calculate
Role of the observation unit
reimbursement and a regionally standardized sum
of money is paid. The actual amount received by In 2002 CMS began an ambulatory patient
the hospital is determined by the diagnosis and is classication (APC) code for the OU. The OU is
independent of the actual cost of care. Therefore, dened as an outpatient environment providing
the nancial margin on which the hospital must a longer period of management, far exceeding the
survive is the dierence between reimbursement usual capabilities of the ED, but not considered
and the true cost of providing care. hospitalization. This new code allows a non-DRG
The single greatest contributor to cost is LOS. reimbursement, independent of the 30-day revisit
Therefore, signicant pressure exists for the in- penalty, for treatment with a length of stay up to
stitution to control LOS so that costs do not 48 hours. OU admissions may occur as many
exceed the CMS payment. Ultimately, an exces- times as is required, and the hospital is reimbursed
sive average LOS results in an institutional loss. for each individual event. Not only is this eco-
With regard to HF, the breakeven point occurs at nomically benecial to the hospital, but by avoid-
approximately 5 days. The average LOS must be ing inpatient hospitalization, limited intensive
maintained closer to 4 days to ensure that a prot management could provide quality-of-life
margin exists to cover the losses incurred by improvements.
outliers. In reality, the average hospital in the The OU is an option for HF care in an
United States is unable to control cost and environment where a period of short intensive
sustains a net loss of $1,288 for every HF therapy, monitoring, and aggressive management
admission [3]. can occur. This denition is important for its
If the only economic incentive were to shorten success. The OU is not a preadmission testing
LOS, undue discharge pressure on hospitals could unit, nor is it a holding area for patients with
adversely impact care. Theoretically, patients unclear treatment goals, undened diagnostic
could be sent home from the hospital every 3 endpoints, or vague disposition plans. Successful
days, only to immediately return for readmission. HF management is complicated and dicult.
Therefore, to balance the early discharge incen- Physician involvement, and a dedicated nursing
tive, the 30-day revisit penalty was created. After sta armed with clear management protocols, can
an initial hospitalization, readmission within 30 signicantly improve the care in the HF patient. A
days for the same DRG may be considered part of well-orchestrated OU HF management plan has
the original visit. In this situation, there is no been shown to decreases revisits, hospitalizations,
additional reimbursement, and the hospital re- and LOS [4]. Even if subsequent hospitalization is
ceives no additional money for the second admis- required, overall LOS, inclusive of the OU visit,
sion. This creates an economic pressure to prevent decreases [4].
premature discharge and eliminate readmissions To qualify for the APC code, there are several
within 30 days. requirements. First, care must be given for at least
The third pressure impacting HF care comes 8 hours in the OU environment. Second, there are
from the resources required to manage the pa- several clinical parameters that must be met; these
tient. This represents the true cost of care and is include the documentation of pulse oximetry
proportionate to the intensity of service provided. measurement, the performance of a chest radio-
Therefore, the more sta or physical resources graph, and obtaining an electrocardiogram
required to manage a patient, the greater the rate (ECG). Finally, the APC allows for care that
HEART FAILURE MANAGEMENT AND THE ED OBSERVATION UNIT 571

may not exceed 48 hours. However, the length of patient. If the average acuity increases, the case
stay rule is tempered by an important caveat: mix multiplier is increased proportionally. By
although the APC structure allows for hospitali- admitting short-stay HF patients to the hospital
zation for up to 48 hours, CMS ceases reimburse- and claiming the DRG, the hospital will realize
ment at 24 hours. Therefore, most units plan for a short-term gain as they receive the higher acuity
disposition within 1 day at which time patients are DRG, up until time for the case mix multiplier
either admitted to the hospital and converted to adjustment. After their case mix multiplier is
a DRG or discharged home. decreased, they will receive a decreased DRG
Although the APC thresholds are easily met, it reimbursement for all their DRG 127 (HF)
is important that the OU does not admit every ED patients. This becomes trading short-term gain
patient with HF. Because the requirement of the for long-term loss and provides the incentive for
OU is discharge within 24 hours, populating the accurate DRG and APC coding.
unit with patients who exceed the LOS parameter
results in decreasing its eciency. Furthermore,
nurse-to-patient ratios are usually lower in the Pathophysiology
OU than in the ICU (three to four patients per
nurse), so that admissions to an OU environment Wall tension is a product of pressure (after-
should match its ability to manage appropriately load) and ventricular radius. Increasing wall
selected patients. tension is a stimulus for cardiac remodeling.
An OU HF management program can provide With increasing tension, cardiac myocytes either
benet to the institution and the patient. The hypertrophy or die (apoptosis) to form scar tissue.
hospital benet is two-fold: (1) by providing care The dominant response determines HF type.
outside of the DRG revisit rule, the average per Many dierent pathologies may ultimately lead
patient reimbursement rate is higher, and (2) the to the clinical presentation of HF (Box 1). The
intensity of service costs are lower, compared with dominating pathway determines the type of HF
the inpatient unit. For the patient, the OU is able that results.
to provide short, intensive therapy without the HF is divided into systolic dysfunction or pre-
necessity of several days of hospitalization and served systolic function (PSF) types by the EF. A
allows early return to the home environment. normal EF is dened as 60%, with systolic
Additionally, the OU can facilitate the perfor-
mance of procedures that may be dicult. For
instance, ejection fraction (EF) measurement is Box 1. Causes of heart failure
a standard for HF treatment and is requirement of
the JCAHO [5]. Transportation limitations in  Acute mitral regurgitation (papillary
nonambulatory patients (eg, a nursing home resi- rupture)
dent) can delay or prevent this evaluation. Like-  Acute pulmonary embolus
wise, multiple outpatient visits required for  Anemia
initiation and optimization of HF medication  Arteriovenous fistula (eg, dialysis)
(eg, beta-blockers or angiotensin-converting en-  Cardiac free wall rupture
zyme inhibitors) can be accomplished.  Constrictive pericarditis
As a last point of discussion regarding the  Complications of MI
duciary structure options for a hospital, an  Coronary artery disease
institution may benet nancially to a greater  Hyperkinetic states
extent by the admission of all short-stay HF  Idiopathic cardiomyopathy
patients to the higher reimbursed inpatient  Myocarditis
DRG. Although this is accurate in the short-  Pericardial disease/tamponade/
term, in the long-term this behavior is penalized effusion
by the case mix multiplier. The case mix multiplier  Poorly controlled hypertension
is the tool by which CMS adjusts DRG reim-  Postpartum cardiomyopathy
bursements at the local level. It is based on the  Sustained cardiac arrhythmia/
average acuity for a given DRG. When a hospitals tachycardia
average acuity of illness is decreased, the case mix  Thyrotoxicosis
multiplier is decreased proportionally so that the  Valvular rupture or disease
hospital receives less average DRG payment per
572 PEACOCK

dysfunction dened as an EF ! 40%. Systolic upper quadrant pain, and HJR, without pulmo-
dysfunction is most commonly the result of ische- nary symptoms. Because the cardiovascular sys-
mic heart disease, although many other causes tem is mechanistically closed and abnormal
exist. The pathologic feature of systolic dysfunc- pressure and chamber volumes are eventually
tion is a ventricle that has diculty ejecting blood. reected to the contralateral side, this distinction
Impaired contractility leads to increased intracar- has greatest applicability when there is suspicion
diac volumes and pressure, and increasing after- of valvular heart disease.
load sensitivity. Consequently, these patients are
sensitive to hypertension, and maintaining blood
Role of neurohormones
pressure (BP) to as low as is tolerated becomes an
important management goal. Before the natriuretic peptides (NPs) were
PSF HF is dened by preserved mechanical identied, extracellular uid regulation was be-
contractile function. When measured, the EF is lieved to be controlled by the kidneys, adrenal
normal or higher. The pathologic decit is a ven- glands, and sympathetic nervous system via the
tricle with impaired relaxation, which results in an reninangiotensin system and other neuroendo-
abnormal diastolic pressure-volume relationship. crine mechanisms [7]. When arterial BP declines,
In this situation, the left ventricle (LV) has renin is released by the kidneys. Renin splits hepati-
diculty in receiving blood. Decreased LV com- cally synthesized angiotensinogen to form angio-
pliance necessitates higher atrial pressures to tensin I. Angiotensin I is a biologically inactive
ensure adequate diastolic LV lling. The hemody- decapeptide that is cleaved by angiotensin-convert-
namic consequence of a stiened noncompliant ing enzyme (ACE) to form active angiotensin II
ventricle is preload sensitivity, where excessively (AII). AII, a potent vasoconstrictor, increases pe-
lowered preload may result in hypotension be- ripheral vascular resistance and causes an increase
cause of a lack of ventricular lling. The fre- in systolic BP. AII also has direct kidney eects
quency of diastolic dysfunction increases with age. that result in salt and water retention and stimu-
Chronic hypertension and LV hypertrophy are lates adrenal aldosterone release. Increased aldo-
often responsible for this syndrome. Coronary sterone causes renal tubular absorption of
artery disease (CAD) also contributes, and di- sodium and results in water retention. Ultimately,
astolic dysfunction is an early event in the ischemic extracellular volume and BP increase [7].
cascade. It has been reported that as many as 30% Natriuretic peptides are important in both BP
to 50% [6] of HF patients have circulatory conges- and uid balance. Physiologically, atrial NP (ANP)
tion on the basis of diastolic dysfunction. and B-type NP (BNP) function as a counter-
The pathologic distinctions based on EF are regulatory arm to the reninangiotensin system
less important in the acute care setting of the ED in regard to BP and volume maintenance. Three
and OU. In these environments, volume overload types of natriuretic peptides are recognized.
with excessive lling pressures are the most ANP is primarily secreted from the atria. BNP
common ED presentation. Irrespective of the is secreted mainly from the cardiac ventricle.
EF, the treatment approach is therefore similar. Finally, C-type natriuretic peptide (CNP) is
However, once hemodynamics are stabilized, and localized in the endothelium. The clinical eects
volume status approaches euvolemia, recognition of NPs are vasodilation, natriuresis, decreasing
of the underlying EF and the etiology of HF levels of endothelin, and inhibition of both the
should be considered. In patients with PSF, reninangiotensinaldosterone system and the
excessive diuresis or venodilation may exacerbate sympathetic nervous system. BNP is synthesized
the underlying decit in ventricular lling and as a prohormone, which is cleaved to inactive
result in hypotension. N-terminal pro-BNP, with a half-life of approx-
Determining HF type is dicult using the imately 2 hours, and physiologically active BNP
history and physical examination; consequently, with a half-life of about 20 minutes [8].
an ECG becomes necessary. Some dierentiate Although BNP was named brain natriuretic
between left- and right-sided HF. Left-sided HF peptide because it was rst identied in porcine
has dyspnea, fatigue, weakness, cough, paroxys- brain [9], in humans the dominant source is myo-
mal nocturnal dyspnea, and orthopnea in the cardial. BNP is secreted and stored in cardiac
absence of peripheral edema, jugular venous ventricular membrane granule [9]. BNP is contin-
distention (JVD), or hepatojugular reux (HJR). uously released from the heart in response to both
Right-sided HF has peripheral edema, JVD, right volume expansion and pressure overload [10].
HEART FAILURE MANAGEMENT AND THE ED OBSERVATION UNIT 573

BNP is cleared by three pathways: a protein recep- hormone elevations are critical and correlate
tor, neutral endopeptidases, and to a lesser extent, directly with mortality in HF patients.
the kidney [11]. Both ANP and BNP have natri- Neurohormonal activation results in both
uretic and diuretic characteristics that increase so- sodium and water retention and an increase in
dium and water excretion by increasing systemic vascular resistance. Although these re-
glomerular ltration rate and inhibiting renal so- exes are initially compensatory and function to
dium resorption [12]. They also decrease aldoste- maintain systemic BP and perfusion, they occur at
rone and renin secretion, causing both a cost of increased myocardial workload and
a reduction in blood pressure and extracellular cardiac wall tension. HF can be asymptomatic
uid volume [812]. Circulating BNP levels in- through these initial neurohormonal and hemo-
crease in direct proportion to HF severity, as dynamic perturbations. However, these reexes
based on the New York Heart Association establish the mechanism that initiates the second-
(NYHA) classication, and BNP is detectable ary pathologic process of cardiac remodeling.
even with minimal clinical symptoms. Physiologi- Neurohormonal activation portends a worse
cally, there is a correlation between BNP concen- prognosis in HF. Its attenuation forms the theo-
trations and LV end diastolic pressure (LVEDP). retical basis for nearly all treatments that decrease
This suggests that the natriuretic eects, coupled morbidity and mortality. This includes ACE
with neurohormonal antagonism, serve to coun- inhibitors, angiotensin-receptor blockers, aldoste-
terbalance uid overload and elevated ventricular rone antagonists, beta-blockers, and nesiritide.
wall tension [12]. There is also an inverse correla-
tion between BNP and LV function after acute Dierential diagnosis
MI. Elevated BNP occurs in the setting of raised
Many diseases mimic HF (Box 2). Because
atrial or pulmonary wedge pressures, or MI [12].
treatment omissions prevent optimal response,
Ultimately, BNP measurement oers an indepen-
and misdirected therapy may have adverse conse-
dent assessment of ventricular function without
quences, an accurate diagnosis is important. Acute
the use of intravascular pressure monitoring.
MI must always be considered as the cause of a HF
visit. As many as 14% of ED HF presentations will
Clinical features of decompensated heart failure have a troponin diagnostic for MI [13]. Further-
more, ADHF patients with elevations in troponin
HF may present after myocardial infarction as
have markedly worse acute outcomes [14].
the result of acute pump dysfunction. This is the
Additionally, because shortness of breath is the
result of the loss of a critical amount of myocar-
most common presenting symptom, other dys-
dial contractile ability, the consequence of which
pneic conditions must be considered. A common
is immediate symptoms. If there is symptomatic
confounder is coexisting chronic obstructive
hypotension accompanied by ndings of symp-
toms of inadequate perfusion (eg, mental status
change, decreased urine output), cardiogenic
shock is diagnosed. Patients with cardiogenic Box 2. Differential diagnosis of acute
shock require hemodynamic monitoring and ar- heart failure
rangements for emergency revascularization. They
are therefore inappropriate OU candidates.  Acute MI or myocardial ischemia
HF can present precipitously, as acute pulmo-  Aortic dissection
nary edema (APE), and also insidiously as the  Chronic obstructive pulmonary disease
nal consequence of a cascade of pathologic exacerbation
events initiated by myocardial injury or stress.  Hypoproteinemias (nephrotic
After a threat to cardiac output, a cascade of syndrome, liver failure)
neurohormonally mediated reexes occurs. These  Pericardial effusion
include activation of both the reninangiotensin  Pneumonia
aldosterone system and the sympathetic nervous  Pneumothorax
system. Consequently, the levels of these neuro-  Pulmonary embolism
hormones increase and include norepinephrine,  Renal failure
vasopressin, endothelin (the most potent vasocon-  Superior vena cava syndrome
strictor known), and TNF-alpha. Although not  Thyroid disease
available in routine clinical practice, these
574 PEACOCK

pulmonary disease. Severe hypertension and pe- ndings are jugular venous distention and ab-
ripheral vasoconstriction may suggest acute HF, dominal jugular reux (AJR), but their overall
even with audible wheezing. Pneumonia or pul- accuracy is only 81% for predicting a pulmonary
monary embolus can mimic or exacerbate HF. capillary wedge pressure (PCWP) O 18 mm Hg.
Finally, edema is common in HF, but is non- Alone, AJR has a specicity of 94%, but its
specic because it is found in many hypoprotei- sensitivity is only 24% [24]. Rales are common,
nemic states, including hepatic or renal failure, but their negative predictive value is only 35%
and vascular diseases. [23]. Lung sounds, peripheral edema, jugular ve-
nous distention, AJR, and the presence of extra
heart sounds help to detect uid overload, but
Diagnostic evaluation
they are either insensitive or nonspecic in those
Eective HF care must begin with accurate at risk for HF. Unfortunately, the easily available
diagnosis and be followed by expeditious treat- diagnostic tests (laboratory studies, ECG, and ra-
ment. Failure with either aspect will have adverse diographs) are not suciently accurate to reliably
eects on outcome. Despite being common in the make a correct diagnosis [16,17].
ED, HF is frequently misdiagnosed. This is due to Many physicians rely on the chest radiograph
the fact that the history and physical, ECG, and (CXR) for diagnosis; however, it is an insensitive
radiograph ndings are either nonspecic or in- tool. In chronic HF, CXR signs of congestion
sensitive [6,1525]. The ED misdiagnosis rate has have poor sensitivity and specicity for detecting
been reported as 12%. Of these, half are patients a high PCWP [19]. A radiographically enlarged
diagnosed with HF, although their symptoms cardiac silhouette can help, but 20% of echocar-
are actually the result of other pathology, and diographically proven cardiomegaly is undetect-
the remainder are HF patients given a dierent di- able on radiograph [20]. Pleural eusions can
agnosis (eg, chronic obstructive pulmonary dis- also be missed by CXR, especially if the patient
ease) [26]. This occurs not only because of the is intubated and the radiograph performed supine.
limitations noted previously but also because the The sensitivity, specicity, and accuracy of the su-
dierential diagnosis of the at-risk population is pine CXR are reported as 67%, 70%, and 67%,
complicated by many other conditions. In individ- respectively [21], with the portable CXR having
ual patients, even experienced physicians disagree even less sensitivity [22].
on the diagnosis of HF, especially if the patient The determination of EF has been termed the
presents early in the disease course [10]. single most important measurement in HF [1]
HF has historically been dened as a syndrome, and represents the standard for noninvasive ven-
but even with the advent of the BNP assay, tricular function assessment. It is indicated in
clinical data are still required for an accurate those without an established diagnosis of systolic
diagnosis. This presents a diagnostic challenge, dysfunction, unless performed within the previous
and various strategies have been constructed to year [1,2830]. Although dening HF cause and
improve diagnostic accuracy. Diagnostic criteria, type is important, there is no correlation between
such as the Framingham HF scoring systems [27], symptoms and EF. Consequently, EF measure-
help address the challenge. However, they depend ment is usually unnecessary in the ED. EF may
on the presence of symptoms, and so they are in- be useful in the OU to help determine treatment
sensitive if the patient is asymptomatic. Conse- strategies at discharge.
quently, the severity of illness is important when
attempting to determine the presence of an HF di-
agnosis. In a study of patients presenting to a pri-
B-type natriuretic peptide assays
mary care environment, the rst diagnosis of HF
was falsely positive in more than 50% of cases. Although a number of companies manufacture
This misdiagnosis rate was attributed to obesity, central lab BNP assay, and Roche Diagnostics
unsuspected cardiac ischemia, and pulmonary dis- produces a central lab NT-pro-BNP assay, there is
ease [15]. Diagnostic accuracy was also aected by currently only one point-of-care BNP assay avail-
gender, and the rate of a correct HF diagnosis was able (Triage BNP, Biosite Diagnostics, San Diego,
surprisingly low at the rst encounterd18% for California). This has ramications in the ED,
women and 36% for men [15]. where time to diagnosis has consequences for
With regard to the physical diagnosis of HF, disposition decisions, as well as general ED
the clinical examination is poor. The best physical operational eciency. The only point-of-care
HEART FAILURE MANAGEMENT AND THE ED OBSERVATION UNIT 575

assay is a uorescent immunoassay that quanti- there is coexistent LV hypertrophy. BNP is also
tatively measures whole blood BNP, or plasma elevated in a number of other conditions. The el-
specimens in which EDTA is the anticoagulant. It derly can have elevated BNP, and there is good
is rated as a moderately complex assay per correlation between BNP, age, and LV mass index
Clinical Laboratory Improvement Amendment [38]. In a study of 72 healthy 85 year olds [37],
(1988) regulations. To perform the assay, a sample compared with 105 healthy 40-year-old men, the
of whole blood is placed in the device, and within older cohort had a mean BNP level of 24.8
15 minutes it displays the BNP concentration. pg/mL, compared with a level of 4 pg/mL in the
Testing should not be delayed more than 30 mi- younger group (P ! .05). BNP may be elevated
nutes after the blood has been placed in the de- in the elderly because of greater ventricular mass
vice. The analytic sensitivity is 5 pg/mL (95% CI as compared with the young [39].
0.2 to 4.8 pg/mL) (per package insert). When BNP is also elevated in renal failure. It is
this system was tested against more than 50 com- unclear if this results from volume overload
monly used cardiac medications (eg, digoxin, war- related BNP elevation or a decrease in BNP
farin, nitroglycerin, furosemide) and cardiac clearance. In a study of 32 hemodialysis patients
neurohormones (eg, renin, aldosterone, angioten- without overt HF, BNP was markedly increased.
sin I, and II, ANP), the assay demonstrated no The predialysis BNP was 688 pg/mL, decreasing
signicant measurement interference or cross-re- to 617 pg/mL after dialysis, although mean BNP
activity (per package insert). levels were higher if the EF was less than 60%
[40]. Studies controlling for creatinine clearance
have shown that BNP has predictive utility for
B-type natriuretic peptide versus lling pressures
the diagnosis of HF, and some have suggested us-
In a study of 72 symptomatic LV dysfunction ing a higher cutpoint of 200 pg/mL for BNP indi-
patients with EF ! 50%, BNP was an indepen- cating HF [41].
dent predictor of increased LVEDP, and BNP Other causes of non-HF BNP elevation include
varied directly with changes in LVEDP [31]. In conditions that result from predominately right
a second study, the sensitivity and specicity for ventricular dysfunction. In these situations, BNP
predicting LVEDP O 18 mm Hg were 81% and can be elevated and may be helpful for determin-
85%, respectively [32]. Others support that an el- ing both diagnosis and prognosis. BNP appears to
evated BNP is predictive of elevated end-diastolic be elevated in ventricular dysfunction, irrespective
pressures [31,33]. In decompensated HF patients, if left or right, and independent of the cause of the
with BNP levels obtained every 2 hours during dysfunction. Although there are limited data to
treatment, there was a correlation between date, this also suggests that BNP may be elevated
PCWP and BNP changes (r 0.79, P ! .05). in other causes of right-sided heart strain (eg,
BNP levels dropped in parallel with a falling a large pulmonary embolus).
PCWP in response to treatment [34]. These studies Recent literature has addressed the potential
suggest that BNP is an indicator of elevated intra- for false-negative BNP levels. Some have reported
cardiac pressures and responds dynamically to a negative relationship between body mass index
ventricular volume and pressure changes. and BNP levels [42]. In the ambulatory care set-
ting, both symptomatic and asymptomatic pa-
tients with chronic, stable systolic HF may
B-type natriuretic peptide confounders
present with a wide range of plasma BNP levels
If the BNP assay is construed as only a test for [43]. Clinical impression, in addition to conrma-
HF, a number of confounders exist. In one report, tory testing, is necessary to interpret the results of
approximate median BNP levels were increased BNP testing accurately.
two-fold in medically treated essential hyperten-
sion, three-fold in cirrhosis, and 25-fold in dialysis
Diagnosis
patients, although all measurements included
signicant ranges around the median [35]. Al- BNP is markedly elevated over baseline in
though some describe increased BNP with hyper- symptomatic HF. Data from several studies
tension [36], others have not duplicated this [10,44,45] indicate that its sensitivity for diagnos-
nding unless there was coexistent LV hypertro- ing HF is from 85% to 97%, with a specicity of
phy [37]. Therefore, isolated hypertension is prob- 84% to 92%. The positive predictive value is
ably not associated with elevated BNP unless 70% to 95% [10,26], and the negative predictive
576 PEACOCK

value consistently exceeds 90% [26,46]. In 250 Vet- after receiving inpatient treatment for HF, 52%
erans Administration dyspneic urgent care pa- had an increasing BNP during hospitalization.
tients, BNP was an accurate predictor of the This compares to the group without readmission
presence of HF. In this trial, the mean BNP in HF or death, in which 84% had a declining BNP
was 1076 G 138 pg/mL, versus 38 G 4 pg/mL during hospitalization [42]. This suggests that
among nonHF patients. Using a cut point of therapy-induced BNP changes can predict out-
100 pg/mL, BNP provided sensitivity, specicity, come in decompensated HF.
and positive and negative predictive values of In the outpatient setting, BNP can reect HF
94%, 94%, 92%, and 96%, respectively [26]. In severity and prognosis [51]. In 290 NYHA class I
this study, 30 patients were misdiagnosed clini- or II HF patients, with a mean EF of 37%, fol-
cally. In 15 patients diagnosed as having HF, al- lowed for 812 days, an initial BNP O 56 pg/mL
though later proven to have another diagnosis, was an independent predictor of HF progression
the mean BNP was 46 G 13 pg/mL. In the 15 erro- and mortality [52]. Others corroborate BNP pre-
neously given nonHF diagnoses, their mean BNP dicting cardiovascular mortality. In a 1-year
was 732 G 337 pg/mL. BNP levels also were indic- study, an elevated BNP was a better predictor of
ative of being hospitalized; those requiring hospi- cardiovascular mortality than age, ANP, EF,
talization had a mean levels of 700 pg/mL, PCWP, gender, HF etiology, or NYHA class
compared with those who were discharged who [53]. An elevated BNP predicts greater mortality
had a mean BNP of 254 pg/mL [26]. and morbidity for HF patients, and this relation
Similar results were found in the larger Breath- is independent of underlying CAD.
ing Not Proper trial, which studied 1586 ED
patients short of breath. Using a cut point of
N-terminal pro-B-type natriuretic peptide
100 pg/mL, BNP had a diagnostic accuracy for
HF of 83.4%, and the negative predictive value N-terminal pro-BNP (pro-BNP) is a synthetic
was 96.0% at a cut point of 50 pg/mL [47]. BNP precursor. Like BNP, pro-BNP originates
Although excellent for excluding the diagnosis primarily from the ventricular myocardium and is
of HF, BNP is only moderately accurate at released as a result of ventricular stress from
determining HF type. It predicts systolic dysfunc- either pressure or volume overload. On a molecu-
tion with a sensitivity and specicity of 83% and lar basis, pro-BNP is about twice the size of BNP
77%, respectively, compared with PSF, for which and has a half-life of 1 to 2 hours. According to
the sensitivity and specicity were 85% and 70%, manufacturer (Roche Diagnostics, San Diego,
respectively [32]. Compared with PSF, systolic HF California) recommendations, the pro-BNP assay
had higher levels; 362 pg/mL, as compared with has two diagnostic cut points based on age, 125
137 pg/mL (P .03) [48]. The receiver operating pg/mL if !75 years and 450 pg/mL if O75 years
characteristic (ROC) AUC for BNP detecting of age. Only recently available, there are relatively
HF was 0.92, (P ! .0001). Although BNP is few data guiding the clinician in the clinical use of
a good predictor of the presence of HF, it does pro-BNP. One trial directly compared pro-BNP
not accurately predict EF. and BNP for predicting EF [54]. They found that
Once a baseline BNP is established, serial while both assays had similar performance, nei-
measures may suggest therapeutic response. In ther was adequately sensitive or specic for clini-
a study of HF patients receiving carvedilol, there cal EF prediction. In evaluating the development
was a correlation (r 0.698, P ! .01) between im- of HF in post-MI patients, pro-BNP had a 97%
proving EF and a declining BNP [49]. In a report sensitivity for predicting an EF ! 45%, if levels
of malignant hypertension, therapy-associated were determined between 3 and 5 days post-MI.
LV hypertrophy regression was associated with Accuracy of predicting post-MI ventricular dys-
decreasing BNP levels [50]. Sequential BNP mea- function is similar between pro-BNP and BNP,
surements may have an application for monitor- but the level must be obtained later after presenta-
ing therapeutic response. tion if using pre-BNP.
Dierences between the two assays include
that, unlike BNP, biologically inert pro-BNP is
Prognosis
not confounded by concurrent nesiritide infusion
In decompensated HF, serial BNP measure- and, therefore, pro-BNP measurement can be
ment may predict outcomes. In a group who died performed accurately during its infusion. Because
or was readmitted to the hospital within 30 days four half-lives are needed to reach steady state,
HEART FAILURE MANAGEMENT AND THE ED OBSERVATION UNIT 577

pro-BNP levels reect hemodynamics from 6 to 8


hours prior. To make the same determination Box 3. Interpretation of B-type
with BNP, because of its much shorter half-life, natriuretic peptide (BNP) assays
a nesiritide infusion must be withheld for 90
minutes so that endogenous BNP levels reect  Low BNP (< 100 pg/mL)
the hemodynamic state rather than infused BNP. - Presenting symptoms are unlikely the
Finally, while BNP is available on a point-of-care result of HF, an alternative diagnosis
platform, pro-BNP is only currently available on should be considered (eg, chronic
a central lab platform. obstructive pulmonary disease).
 Intermediate BNP (100 to 500 pg/mL)
- Consider other diagnoses (Cor
Clinical use of B-type natriuretic peptide assays
pulmonale, pulmonary embolism,
In ED HF patients, an elevated BNP suggests primary pulmonary hypertension).
future adverse cardiac events [55] and, when con- - Compare with prior baseline BNP
sidered with clinical impression, may help in se- levels.
lecting candidates for OU HF therapy. BNP is  High BNP (> 500 pg/mL)
most useful for excluding HF in those conditions - HF is the likely diagnosis, although
where the dierential would normally suggest confounders should be considered.
HF. This includes undiagnosed patients with any
of the classic signs or symptoms of HF (shortness
of breath dyspnea on exertion, orthopnea, depen- Supplemental oxygen can be given based on pulse
dent edema, and physical examination ndings of oximetry. As the consequences of hypoxia are of
a cardiac S3, jugular venous distention, or basilar greater signicance than the potential for hyper-
rales). In this scenario, a normal BNP should carbia, O2 is not withheld based on CO2 retention
prompt the consideration of an alternative concerns. Arterial blood gases may be helpful in
diagnosis. the critically ill or if CO2 retention is likely.
Because nonHF conditions can result in an Because patients with HF do not hemodynam-
elevated BNP, the clinical context of a positive ically tolerate hypertension, it is desirable to
BNP must be considered. A positive BNP suggests maintain the BP as low as is consistent with the
the need for routine tests to conrm the diagnosis, ability to mentate, urinate, and ambulate. Chronic
as well as evaluation of the cause and denition of systolic BPs in the 90 mm Hg range are usually
the type of HF (eg, ECG, CXR, and echocardio- well tolerated by the HF patient. In the hyperten-
gram). See Box 3 for suggested approach to the sive acute pulmonary edema patient, a controlled
use of the BNP assay. lowering of BP may result in a profound improve-
BNP is also used to monitor chronic HF, ment of symptoms and dyspnea.
because levels correlate with treatment ecacy. Noninvasive positive pressure ventilation
After HF exacerbation, a declining BNP indicates (NPPV) is controversial. Some report it may be
a good response to therapy and portends a more used to prevent endotracheal intubation in the
favorable outcome. A rising BNP suggests a greater properly selected patient while awaiting hemody-
risk of adverse outcome, and a more aggressive namic interventions to become eective. Biphasic
treatment strategy may be warranted. positive pressure ventilation (BiPAP) requires the
delivery of separately controlled inspiratory and
expiratory pressures via facemask. Continuous
Emergency department management positive airway pressure (CPAP) provides con-
stant pressure throughout the respiratory cycle.
General support
For a trial of NPPV, close cardiac monitoring,
The initial approach is based on the acuity of relative hemodynamic stability, and patient co-
presentation, volume status, and systemic perfu- operation are needed. NPPV may provide mor-
sion. In critical patients, airway management tality benet over invasive mechanical ventilation
overrides all other interventions. This is in con- in chronic obstructive pulmonary disease, but the
trast to minimally symptomatic patient whose data are controversial in APE. In APE the use of
evaluation may occur in lieu of stabilization NPPV may decrease the rate of endotracheal
procedures. Initial stabilization is aimed at main- intubation, but mortality is unchanged and
taining airway control and adequate ventilation. CPAP patients may have higher rates of MI
578 PEACOCK

than those treated with BiPAP [56]. Patients re- may be used to monitor uid status or if urine
quiring PPV or more than 2 L per nasal cannula output is sucient to interfere with the patients
supplemental oxygen are not good OU ability to rest.
candidates.
Intravenous (IV) access is needed in all with
Benets of early therapy
HF exacerbation. This is because electrolyte
abnormalities may occur from aggressive diuretic All patients not being discharged from the ED
therapy, and HF patients are at risk for ventric- should receive therapy while still in the ED. The
ular arrhythmia. In both scenarios, prompt ther- reasons for this are two-fold. Treatments delayed
apy can be needed. Limitations notwithstanding, until inpatient arrival are not received by the
all suspected HF patients should have a chest patient for many hours Second, optimal ED HF
radiograph to help exclude other confounding outcomes are directly related to time to treatment.
diagnoses and provide conrmatory evidence of The Acute Decompensated HF National Reg-
HF. Underlying coronary artery disease is the istry (ADHERE) is a multicenter database that
most common cause of HF in the United States; records data from episodes of hospitalizations in
therefore, until stability is determined, all sus- patients discharged after an inpatient stay for
pected HF patients need initial continuous ECG acutely decompensated HF. Using data from
monitoring, a 12-lead ECG, and cardiac bio- 46,599 hospitalizations from this registry [57], pa-
marker testing. Additionally, a search for the tients were stratied as to where IV vasoactive
other HF precipitants is needed (Box 4). therapy was started. Vasoactive therapy was de-
Diuresis causes abnormalities in K, Na, ned as the receipt of an intravenous agent that
blood urea nitrogen, and creatinine, undetectable would be administered to eect a change in hemo-
by history or examination, so these should be dynamics (eg, dopamine, dobutamine, nitroglyc-
evaluated. A complete blood count is needed to erin, nesiritide). Of the cohort receiving
check for anemia. With hepatomegaly resulting vasoactives, 4096 received them in the ED, com-
from passive congestion, liver enzymes can ex- pared with 3499 whose treatment was delayed un-
clude other pathologies. Elevated lactate levels til arrival on the inpatient unit. Patients treated in
may identify unsuspected cardiogenic shock, and the ED received the vasoactive agent much sooner
testing for drug levels (eg, digoxin) is guided by (1.1 versus 22.2 hrs), had reduced lengths of stay
presentation. Occasionally ethanol and drug (4.5 versus 7.0 days), and lower mortality (4.3 ver-
screening may be needed. Assessment of the Mg2 sus 10.9%) than those whose treatment was de-
level is considered when there is cardiac arrhyth- layed until arrival on the inpatient unit.
mia and if severe or treatment-resistant hypokale- An experienced emergency physician usually
mia occurs. Finally, selective foley catheterization knows within minutes which patients will need
inhospital admission.. When the requirement for
HF admission is identied, the early use of
Box 4. Common causes of heart failure vasoactive therapy is indicated. In the minority
decompensation of HF patients for whom ED treatment may result
in discharge home, a limited trial (2 to 3 hours) of
 Atrial fibrillation intermittent diuretic bolus therapy may be
 Acute MI appropriate.
 Chronic nonsteroidal anti-
inflammatory drug use
Identifying candidates for the observation unit
 Excessive alcohol
 Endocrine abnormalities (eg, diabetes, The OU is an eective treatment option for
hyperthyroidism) many diseases requiring a short period of in-
 Infection tensive therapy or diagnostic evaluation. Patients
 Negative inotropic medications with a HF exacerbation usually require longer
 Noncompliance (diet, medication) than the short treatment course possible in the
 Suboptimal pharmacologic ED; consequently, many are OU candidates for
management additional therapy. Relief of congestion is the
 Obesity most common rate-limiting step preventing dis-
 Uncontrolled hypertension charge. In the past, most HF patients were simply
admitted for inpatient therapy. Because most will
HEART FAILURE MANAGEMENT AND THE ED OBSERVATION UNIT 579

have insucient diuresis for relief of congestion in Like acute coronary syndromes, patients with
the ED, the OU oers an opportunity for longer decompensated HF represent a spectrum of pre-
therapies and may prevent the necessity of an sentation that must be sorted to determine
inpatient admission [58]. appropriate OU admission. Although there are
The OU provides safe and eective therapy for very few analyses to dene the best OU candidate,
the appropriately selected HF patients. In a retro- the exclusion of patients with excess risk is
spective study of decompensated HF [59], post- important. An analysis of the ADHERE registry
discharge revisit rates for OU patients treated [60] provides several mortality predictors that
for 24 hours, compared with inpatients discharged should exclude patients from consideration of
within 24 hours, were superior in the OU cohort. OU admission. The most of important of these
In the OU group, there were no return visits with- predictors is the BUN. Patients with a BUN ex-
in 1 week of admission compared with a revisit ceeding 43 mg/dL have markedly increased inpa-
rate of 8% in the hospitalized group. By 1 month, tient mortality (8.98%) compared with those with
only 8% of the OU patients had a revisit, versus a lower BUN (2.68%). Further rening the risk
16% of the inpatient group. There were no mor- stratication process, in the cohort of patients
tality dierences between the groups (P O .05).
This study suggested that OU treatment is at least
as safe and eective as a similar period of inpa- Box 6. Heart failure observation unit:
tient hospitalization. exclusion criteriaa
Patient selection before OU consideration is
important. If a patient has a high likelihood of  Unstable vital signs (despite
HF, with pulmonary or systemic congestion, and emergency department therapy):
meets the entry criteria listed in Box 5, OU admis- -Systolic BP > 220, Diastolic BP > 120
sion may be appropriate. Because OU treatment is mm Hg
temporally restricted, with lower nursingpatient -Respiratory rate > 25 /min
stang ratios, and has limited invasive monitoring -Heart rate > 130 beats/min
capability, careful patient selection is necessary to -Temperature > 38.5 C
ensure that admissions are appropriate for the
available level of care. Exclusion criteria, designed  Electrocardiogram or serum markers
to prevent admissions in those with needs exceed- diagnostic of myocardial ischemia or
ing OU resources, are listed in Box 6 [58]. Patients infarction
with airway instability, a high probability of ad-  Unstable airway or oxygen
verse outcome (eg, acute MI), and those with he- requirement > 4 L/min by nasal
modynamics suggestive of critical underlying cannula
pathology should be excluded.  Evidence of cardiogenic shock or signs
of end-organ hypoperfusion
 Require continuous titration of
vasoactive medication (eg,
nitroglycerin)
Box 5. Heart failure observation unit:  Clinically significant cardiac
inclusion criteria arrhythmia
 Altered mental status
 Adequate systemic perfusion  Severe electrolyte imbalances
 Evidence of hemodynamic stability  Chronic renal failure requiring dialysis
-Heart rate > 50 and < 130 beats/min  Peak expiratory flow rate < 50% of
-Systolic BP > 90 and < 175 mm Hg predicted
-Oxygen saturation > 90%  Chest radiograph suggestive of
 No evidence of acute ischemia/ pneumonia
infarction by electrocardiogram or
cardiac markers a
These criteria are meant to discourage
 Chest radiograph findings compatible entry by patients not likely to benefit from an
with the diagnosis of HF aggressive diuresis and vasodilation manage-
 B-type natriuretic peptide > 100 pg/mL ment protocol.
580 PEACOCK

with a BUN less than 43, adding systolic BP pro- The only validated OU HF protocol published
vides additional prognostic data. In this group, to date includes an aggressive diuretic algorithm,
the addition of a systolic BP ! 115 mm Hg was initiated in the ED, and continued throughout the
associated with a mortality rate of 5.49% com- OU stay [4,58]. In this protocol, additional diuretic
pared with those with a higher BP whose mortal- use was driven by the patients urine output. If the
ity was 2.89%. urine output was inadequate, inpatient admission
OU admission candidates should have a BUN for invasive monitoring was suggested. Addition-
! 43 and careful consideration if admitted with ally, ACE inhibitor algorithms encouraged physi-
a systolic BP ! 115 mm Hg. cian initiation and up-titration toward target
Other studies have specically examined OU levels, provided there were no renal function con-
outcome predictors for HF. In 499 patients, traindications, systolic blood pressure was ade-
Diercks and colleagues [61] reported that a nega- quate, and there was no history of ACE inhibitor
tive troponin and an initial systolic BP O 160 intolerance. Unless there are signicant contrain-
mm Hg identied a cohort who were successfully dications (eg, anaphylaxis), all HF patients should
discharged within 24 hours of admission and be discharged on an ACE inhibitor [1].
had no death or re-hospitalization within the sub- The OU provides an opportunity for more
sequent 30 days. Burkhardt and colleagues [62] re- extensive evaluation than can be performed in
ported in 385 OU patients that successful most EDs. EF measurement may be determined in
discharge from the OU within 24 hours of an ad- those without an established diagnosis of systolic
mission for decompensated HF was predicted by HF. This should be repeated if PSF HF was
an admission BUN ! 30 mg/dL. These parame- previously diagnosed, and it has been more than 1
ters may also be considered to assist in the selec- year since the last assessment of ventricular
tion of the appropriate OU candidate. function. The OU environment also oers the
Once admitted to the OU, many dierent option of elective multidisciplinary consultations,
aspects of medical management are needed to not available in a busy ED, for those who may
assure optimal outcomes and discharge rates. have transportation diculties in getting to an
Attention to the many details required for HF outpatient appointment. While in the OU, the
management [63,64] can be daunting for a physi- option of HF cardiology specialist consultation
cian already running a busy ED. This not only in- may help to evaluate discharge medication dos-
cludes medication intervention and titration, but ages, and screen candidates for heart transplan-
the diagnostic evaluation, patient education, and tation listing. Other ancillary care sta may
the discharge planning required by regulatory consult also. This includes dietetics and home
agencies. HF protocols drive treatment algorithms health care. Social workers can ensure that all
and provide superior outcomes, as compared with patients have the ability to obtain their medicines
standard therapy, and ensure that required inter- and can arrange a home environment assessment
ventions are accomplished [4,58]. to assess the potential of other psychosocial,
ED OU protocol-driven management has been cultural, or economic factors that could prevent
shown to result in a signicant improvement in therapeutic compliance. A home health care con-
outcomes, as compared with independent physi- sultation serves to ensure post-discharge follow-
cian-driven care. In a before and after study of up care and can help arrange visiting nurse
154 decompensated OU HF patients, a prespe- services for home bound or nonambulatory
cied management protocol resulted in signicant patients.
outcome improvements [4]. Use of the protocol re- Because noncompliance causes up to 50% of
sulted in 90-day ED HF revisit rates declining by HF rehospitalizations [63,64,66], patient educa-
56% (0.90 to 0.51, P .0000) compared with pre- tion is a critical facet in the treatment program.
protocol management. Similarly, 90-day HF reho- Bedside teaching videotapes on HF may provide
spitalizations decreased by 64% (0.77 to 0.50, P detailed education during a teachable moment
.007). Lastly, 90-day mortality and OU HF read- for the patient. Finally, patients should be pro-
missions decreased from 4% to 1% (P .096), vided with HF literature, medication informa-
and 18% to 11% (P .099), respectively. From tion, and lifestyle modication suggestions at
a cost perspective, during the same time period, discharge.
annualized hospital costs declined by nearly OU HF management not only impacts the OU,
$100,000, predominately the result of 30-day read- it results in changes in the inpatient HF popula-
mission avoidance [65]. tion as well. By treating selected patients in the
HEART FAILURE MANAGEMENT AND THE ED OBSERVATION UNIT 581

OU, rather than the inpatient unit, patients are


diverted to less intensive levels of care. After Box 7. Heart failure observation unit:
implementation of an OU HF management pro- management protocol
tocol, inpatient acuity as indexed by the mean
number of billable procedures per HF patient,  Standardized orders (algorithm)
increased by 11% [65]. This provides improved re- - Diuretics
source matching between patients requiring inten- - Neurohormonal antagonist
sive monitoring environments and those who (angiotensin-converting enzym
could benet from less costly OU care. inhibitors or nesiritide)

 Aggressive fluid management


Therapeutic agents
monitoring
HF management is a complicated task. It - Admission weights, input/output
requires the successful interaction of many dier- - Fluid restriction, low-sodium diet
ent medications, individual titration regimens,
patient factors (eg, education and compliance  Diagnostic testing
issues), and multidisciplinary services. Clearly - Standing orders for measurement and
dened patient care management programs ad- correction of potassium and
dressing these issues have demonstrated clinical magnesium
and nancial success in both the OU and out- - Electrocardiogram and cardiac injury
patient environment [4,58,6367]. A validated markers
management protocol is provided in Box 7. - Echocardiography
Very large trials provide evidence-based data
for managing stable, systolic HF. The principal  Patient education
drugs are beta-blockers, ACE inhibitors, angio- - Educational videos (living with HF,
tensin-receptor blockers (ARBs), hydralazine/ni- smoking cessation)
trates, diuretics, digoxin, and spironolactone. In - Personalized educational material
general, the strategy focuses on maintaining the - Nursing and physician bedside
lowest possible BP that allows mentation, ambu- teaching
lation, and urination [1]. All HF patients without - Dietetic consultations
contraindications should be on an ACE inhibitor
and beta-blocker, even in the setting of stable dis-  Discharge planning
ease with minimal symptoms. In most diseases, - Social worker, home health nurse
therapy is driven by continuing symptoms; this consultation
is not the situation in HF because of the unique - HF specialist consultation
neurohormonal antagonism requirements for the
treatment of HF.
The emphasis on neurohormonal antagonism
in HF represents a major management shift. The of 180 mg to promote urine output. If a patient is
relief of congestion by the use of diuretics has not currently on a diuretic, 40 mg of furosemide is
been the main thrust of ED therapy. Although usually adequate [4,58,59,65,68]. Urine output
diuretics are important for acute symptomatic and serum electrolytes are monitored to track di-
congestion relief, they do not improve mortality. uresis volume and to screen for treatment-induced
Very large studies evaluating the role of ACE hypokalemia. If target urine outputs are not met,
inhibitors, beta-blockers, and other agents show the diuretic dose is doubled and repeated at 2
neurohormonal antagonism is required for the hours. Adequate output should exceed 500 cc
greatest mortality improvement. within 2 hours, unless the creatinine exceeds 2.5
mg/dL. With elevated creatinine, 2-hour urine
Diuretics output goals are halved. Failure to meet output
Initial OU HF therapeutic goals are directed at goals suggests inpatient hospitalization may be
the relief of congestion by the use of IV diuretics. needed [4,58,59,65,68].
Recommended dosing strategies are to use up to Diuretics prevent hospitalization and provide
twice the daily dose of furosemide (or its equiv- symptomatic relief. They are indicated for all
alent) administered as an IV bolus, to a maximum patients with congestive ndings, but not as
582 PEACOCK

chronic monotherapy (diuretics should be com- signicant intolerance or a contraindication to


bined with ACE inhibitors and beta-blockers) the ACE inhibitor exists [1,29].
because alone they provide no mortality benet Improvement of hemodynamic parameters in
[1,5,27,29]. With aggressive diuresis, potassium the acutely decompensated HF patient, reected
supplementation is frequently required. However, by improvements in vital signs and in the clinical
if loop diuretics are used in concert with nesiritide, presentation of dyspnea, may be obtained with the
potassium supplements are less often needed. use of intravenous vasodilators. The appropriate
Standing orders for oral or IV potassium supple- candidate usually has a blood pressure in excess of
mentation with the goal of maintaining a K be- 90 mm Hg and no contraindications to vasodila-
tween 4.0 and 5.0 meq/dL may be useful. tion. Vasodilation can result in hypotension in
Magnesium is an important cofactor for myo- selected populations, including those conditions
cardial function and should be supplemented if where there is either an impediment to outow
the patient is decient. Give magnesium if the (eg, aortic stenosis), situations where cardiac out-
creatinine is ! 1.5 mg/dL; otherwise, therapy put depends on adequate or elevated pre-load (eg,
should be individualized. Magnesium may be right ventricular infarct, pericardial tamponade),
given as 140 mg magnesium oxide orally once. or when preload is already abnormally decreased
Limited data suggest that combining loop (eg, volume depletion).
diuretics treatment with IV vasoactive agents has Because their optimal use in HF requires
benet. In 1442 patients from 42 hospitals, invasive monitoring, nitroprusside and nitroglyc-
patients receiving intermittent IV bolus therapy erin are generally discouraged from use in the OU
(eg, furosemide bolus) of any type while in the ED environment. Conversely, nesiritide, which can be
had a mean hospitalization LOS of 9.9 days. This used safely and ecaciously without invasive
compared with an LOS of 6.6 days (P .004) for hemodynamic monitoring, is permissible in
those treated with IV infusion therapy (eg, nesiri- the OU.
tide or nitroglycerin) while still in the ED. In this Nesiritide is an IV medication for the treat-
registry analysis, early infusion therapy with a va- ment of decompensated HF. It provides hemody-
soactive agent, as opposed to intermittent bolus namic [73,74] and clinical benets [7577] from
therapy, was associated with a signicantly the combination of vasodilation [78], natriuresis,
shorter hospital LOS [57]. and neurohormonal antagonism [78]. It has been
shown to decrease costs and hospital readmissions
[79] and lower 6-month mortality as compared
with dobutamine [75,76]. Compared with nitro-
Vasodilators glycerin, nesiritide improves dyspnea and hemo-
Certain vasodilators have the potential to dynamics more rapidly and to a greater extent
provide both symptomatic improvement and [77]. Nesiritide is appropriate for use in the ED
mortality reduction in HF. Although hemody- OU. Once stabilized in the ED, patients on the
namic improvements are the direct result of the recommended xed dose of nesiritide (2 mg/kg
eect on the vascular tree, mortality improve- IV bolus, then a 0.01 mg/kg infusion) are candi-
ments in long-term outcomes are not a universal dates for OU admission.
feature of all vasodilators. For mortality reduc- The Proaction Trial was a multicenter, double-
tion to occur, the vasodilator must also have the blind, randomized, placebo-controlled safety and
characteristic of neurohormonal antagonism di- ecacy trial of standard OU therapy, with and
rected at the pathologic hormonal excesses of HF without nesiritide [80]. In this trial, OU patients
[6972]. With regard to specic agents, ACE in- with acute decompensated HF and receiving stan-
hibitors, with both vasodilation and neurohor- dard therapy were treated with at least 12 hours of
monal antagonistic eects, represent a class of either standard dose-blinded nesiritide or placebo.
medications with a mortality reduction benet of In the safety analysis, there were no dierences in
such magnitude that all HF patients deserve adverse outcomes between the standard care and
a therapeutic trial [1,29,5,6972]. ARBs, predicted nesiritide cohorts. NYHA class III or IV HF pa-
to have similar physiologic eects as ACE inhibi- tients receiving nesiritide had a 29% decrease in re-
tors, may have fewer side eects than ACE inhib- visit rates (P .057). Days in the hospital in the
itors. However, as the preponderance of the month after study entry were markedly lower in
current mortality reduction data uses ACEI, patients receiving nesiritide compared with stan-
they should be replaced by ARBs only if dard therapy (2.5 versus 6.5, respectively; P .03).
HEART FAILURE MANAGEMENT AND THE ED OBSERVATION UNIT 583

Digoxin DVT prevention in hospitalized, bedridden HF


Similar to diuretics, digoxin decreases hospital- patients must be balanced against the relatively
izations but does not alter mortality. It is recom- low risk of complications associated with antico-
mended for LV systolic dysfunction and rate agulant prophylaxis. The precise value of prophy-
control in atrial brillation [1,5,29]. Toxicity man- lactic anticoagulation has never been reported in
ifests as cardiac arrhythmia (heart block, ectopy, ED OU HF management. However, as it is antic-
or re-entrant rhythms), gastrointestinal symp- ipated that this cohort of patients will be dis-
toms, or neurologic complaints (eg, visual distur- charged in a relatively limited time frame,
bances and confusion). Serum levels can suggest anticoagulation is usually not performed.
toxicity if they exceed 2.0 ng/mL, but toxicity
can also occur at lower levels if there is coexistent Other agents
hypokalemia or hypomagnesemia. Digoxin Calcium channel blockers (CCBs) are not
should be used at a dose of 0.125 to 0.250 mg routinely recommended in HF [1,90]. This is be-
daily [1,5,29]. cause short-term use may result in pulmonary
edema and cardiogenic shock, whereas in the
Beta-blockers long-term, they may increase the risk of worsening
Beta-blockers prolong life in HF patients [81 HF and death [9194]. These adverse eects have
85] but should only be initiated if the patient is he- been attributed to the negative inotropic eects of
modynamically stable [1,29]. They should not be CCBs. If necessary, amlodipine, the CCB with no
started in decompensated HF; consequently, this clear adverse mortality eect, may be used for
generally precludes OU initiation. Conversely, compelling clinical reasons (eg, as an anti-anginal
abruptly stopping a beta-blocker has the potential agent despite maximal therapy with nitrates and
to worsen hemodynamics. Recommended therapy beta-blockers).
for decompensated HF patients presenting to the Nonsteroidal anti-inammatory drugs should
ED while on a beta-blocker is to hold the dose be avoided in HF [1,90]. They inhibit the eects of
or continue it at one dosing level lower than the diuretics and ACE inhibitors and can worsen car-
maintenance dose. If inotropes are required, the diac and renal function [1].
beta-blocker may be withheld [1,29]. Patients at HF is an important risk factor for sudden
this stage of their disease are not usually ideal cardiac death, and its likelihood increases in
OU candidates. proportion to the decrease in EF and HF severity
[95]. Premature ventricular contractions occur in
Aldosterone antagonists 95% of dilated cardiomyopathy patients, and
The aldosterone antagonist spironolactone de- nonsustained VT may be seen in up to 30% to
creases the relative risk of mortality in end stage 40% of cases. Prophylactic administration of
HF [86]. These patients should receive 12.5 to 25.0 anti-arrhythmics is not eective and may actually
mg qid [1,29]. It is not recommend if the creatinine increase mortality [96]. Therefore, their routine
exceeds 2.5 mg/dL or the K is O 5.0 mEq/L. If use to suppress asymptomatic ventricular arrhyth-
a patient is already on an aldosterone antagonist, mias is not warranted.
it should be continued in OU patients.

Observation unit disposition


Anticoagulants
The risk of thromboembolism in the clinically OU patients may be discharged at any time,
stable HF outpatient is low, estimated at 1% to once there has been a good therapeutic response.
3% per year, and is greatest patients with the Although there are very few studies that have
lowest EF [87,88]. However, hospitalized HF pa- determined predictors of successful ED discharge,
tients are at signicant risk for sustaining deep a net volume output of greater than 1 L is
vein thrombosis (DVT) and its complications. In associated with a higher rate of successful dis-
the MEDENOX (prophylaxis in MEDical pa- charge from the ED OU [68].
tients with ENOXaparin) study of hospitalized Discharge criteria may aid in selecting candi-
patients, subcutaneous enoxaparin decreased ve- dates who may be sent home (Box 8). Most im-
nographically documented DVT, from 14.9% in portant in the disposition determination is the
the placebo group to 5.5% in the 40-mg every clinical assessment, but a post-treatment BNP
day enoxaparin prophylaxis group [89]. Empiric level can help to guide this decision. If nesiritide
584 PEACOCK

nurses. Because noncompliance is estimated to


Box 8. Heart failure observation unit: cause 50% of HF re-hospitalizations [60,63,
discharge guidelinesa 64,66], patient education is a critical facet to be
addressed during OU treatment.
 Patient reports subjective Disposition from the OU is predicated on the
improvement. relief of dyspnea, improvement of congestion
 If normally ambulatory, able to do so without long-suering orthostasis, and discharge
without significant orthostasis. to an adequate outpatient environment. If these
 Resting heart rate < 100 beats/min; goals cannot be met, inpatient hospitalization or
Systolic BP > 80 mm Hg. placement in an assisted-living facility should be
 Total urine output > 1 L, and urine considered. Approximately 25% [4,58,59,65] of
output > 30 cc/h (or > 0.5 cc/kg/h). HF patients will require inpatient hospitalization
 Room air oxygen saturation > 90% after a 24-hour OU admission. Patients not meet-
(unless on home oxygen). ing discharge criteria by the 24-hour OU LOS lim-
 No angina. it require inpatient admission. A lower threshold
 No electrocardiogram or cardiac for admission is appropriate in the very elderly,
marker evidence of myocardial those with poor social situations, and those with
ischemia/infarction. multiple co-morbidities. Even patients requiring
 No new clinically significant admission after an OU stay derive a measurable
arrhythmia. benet from its use. In patients admitted from
 Normal electrolyte profile without the OU after failure of therapy, the mean hospital-
increasing azotemia. ization LOS, inclusive of their OU time, was 0.8
days less than patients admitted directly from
a the ED to the inpatient unit [4].
Patients not meeting all of the guidelines
should be considered for inpatient treatment
(except as appropriate in the end-stage pallia-
tive care cohort).
References
[1] Packer M, Cohn JN. Consensus recommendations
for the management of chronic heart failure. Am J
has been used, it will be necessary to stop its infu- Cardiol 1999;83(2A):1A38A.
sion at least 4 half-lives (90 minutes) before its [2] Massie BM, Shah NB. Evolving trends in epidemi-
measurement. ologic factors of heart failure: rationale for preven-
Successful discharge of HF patients after an tative strategies and comprehensive disease
OU treatment course requires the coordination of management. Am Heart J 1997;133:70312.
a multimedication regimen that includes diuretics, [3] CMS 2001 in-patient discharge data base: MED-
digoxin, ACE inhibitors (or ARBs), and possibly PAR, DRG 427.
spironolactone, and beta-blockers. Optimally, [4] Peacock WF 4th, Remer EE, Aponte J, et al. Eec-
tive observation unit treatment of decompensated
medications are adjusted to meet target-dosing
heart failure. Congest Heart Fail 2002;8:6873.
recommendations [1,5,29]. Therefore, close con- [5] USDHHS: Heart failure: evaluation and care of
sultation with the physician who will ultimately patients with left-ventricular systolic dysfunction
manage the outpatient course is necessary to pro- (Clinical Practice Guideline, Number 11, AHCPR
vide the best outcomes. An aggressive manage- publication No. 940612). US Department of
ment protocol may anticipate the discharge of Health and Human Services, Washington, DC,
approximately 75% of OU HF patients June, 1994.
[4,58,59,65]. [6] Marantz PR, Tobin JN, Wassertheil-Smoller S,
Discharge planning is critical in HF. A multi- et al. The relationship between left ventricular sys-
disciplinary team approach can produce signi- tolic function and congestive heart failure diag-
nosed by clinical criteria. Circulation 1988;77:
cant improvements in HF outcomes by decreasing
60712.
revisit rates, lowering inpatient length of stay, [7] Hobbs RE, Czerska MD. Congestive heart failure.
and reducing hospitalization costs [4,58,59,65,67, Current and future strategies to decrease mortality.
97100]. The team approach should provide the Postgrad Med 1994;96:16772.
option of outpatient consultations with social [8] Clemens LE, Almirez RG, Baudouin KA, et al.
work, dietetics, cardiology, and advance practice Pharmacokinetics and biological actions of
HEART FAILURE MANAGEMENT AND THE ED OBSERVATION UNIT 585

subcutaneously administered human brain natri- [24] Marantz PR, Kaplan MC, Alderman MH. Clinical
uretic peptide. J Pharmacol Exp Ther 1998;87: diagnosis of congestive heart failure in patients
6771. with acute dyspnea. Circulation 1993;88:10715.
[9] Nakao K, Mukoyama M, Hosoda K, et al. Biosyn- [25] Peacock WF, Kies P, Albert NM, et al. Bioimpe-
thesis, secretion, and receptor selectivity of human dance monitoring for detecting pulmonary uid in
brain natriuretic peptide. Can J Physiol Pharmacol heart failure: equal to chest radiography? Congest
1991;59:15006. Heart Fail 2000;6:869.
[10] Cowie MR, Struthers AD, Wood DA, et al. Value [26] Dao Q, Krishnaswamy P, Kazanegra R, et al. Util-
of natriuretic peptides in assessment of patients ity of B-type natriuretic peptide in the diagnosis of
with possible new heart failure in primary care. congestive heart failure in an urgent-care setting.
Lancet 1997;350:134953. J Am Coll Cardiol 2001;37:37985.
[11] Mair J, Thomas S, Puschendorf B. Natriuretic pep- [27] McKee PA, Castelli WP, McNamara PM, et al.
tides in assessment of left-ventricular dysfunction. The natural history of congestive heart failure:
Scan J Clin Invest 1999;9(Suppl 230):13242. The Framingham Study. N Engl J Med 1971;285:
[12] Struthers AD. Ten years of natriuretic peptide re- 14416.
search: a new dawn for their diagnostic and thera- [28] Francis CM, Caruana L, Kearney P, et al. Open ac-
peutic use? BMJ 1994;308:16159. cess echocardiography in management of heart fail-
[13] Peacock WF, Emerman CL, Doleh M, et al. The in- ure in the community. BMJ 1995;310:6346.
cidence of elevated cardiac enzymes in decompen- [29] ACC/AHA. ACC/AHA guidelines for the evalua-
sated heart failure. Acad Emerg Med 2001;5:552. tion and management of chronic heart failure
[14] Peacock WF, DeMarco T, Emerman CL, Wynne J. in the adult. [Cited March 18, 2002]. Avail-
Heart failure with an elevated troponin is associated able from: http://www.acc.org/clinical/guidelines/
with increased morbidity and mortality: an AD- failure/I_introduction.
HERE registry analysis. Ann Emerg Med 2004; [30] Wheeldon NM, MacDonald TM, Flucker CJ, et al.
44(4):S72. Echocardiography in chronic heart failure in the
[15] Remes J, Miettinen H, Reunanen A, et al. Validity community. Q J Med 1993;86:25561.
of clinical diagnosis of heart failure in primary [31] Maeda K, Tsutamoto T, Wada A, et al. Plasma
health care. Eur Heart J 1991;12:31521. brain natriuretic peptide as a biochemical marker
[16] Stevenson LW. The limited availability of physical of high left ventricular end-diastolic pressure in
signs for estimating hemodynamics in chronic heart patients with symptomatic left ventricular dysfunc-
failure. JAMA 1998;261:8848. tion. Am Heart J 1998;135:82532.
[17] Davis AP, Francis CM, Love MP, et al. Value of [32] Yamamoto K, Burnett JC, Jougasaki M, et al. Su-
the electrocardiogram in identifying heart failure periority of brain natriuretic peptide as a humoral
due to left ventricular systolic dysfunction. Br marker of ventricular systolic and diastolic dys-
Med J 1996;312:222. function and ventricular hypertrophy. Hyperten-
[18] Marantz PR, Kaplan MC, Alderman MH. Clinical sion 1996;28:98894.
diagnosis of congestive heart failure in patients [33] Omland T, Aakvaag A, Bonarjee V, et al. Plasma
with acute dyspnea. Chest 1990;97:77681. brain natriuretic peptide as an indicator of left ven-
[19] Chakko S, Woska D, Marinez H, et al. Clinical, tricular systolic function and long-term survival
radiographic, and hemodynamic correlations in after acute myocardial infarction. Comparison
chronic congestive heart failure: conicting results with plasma atrial natriuretic peptide and N-termi-
may lead to inappropriate care. Am J Med 1991; nal proatrial natriuretic peptide. Circulation 1996;
90:3539. 93:19639.
[20] Kono T, Suwa M, Hanada H, et al. Clinical signif- [34] Kazanegra R, Cheng V, Garcia A, et al. A rapid test
icance of normal cardiac silhouette in dilated for B-type natriuretic peptide correlates with falling
cardiomyopathy; evaluation based upon echocar- wedge pressures in patients treated for decompen-
diography and magnetic resonance imaging. Jpn sated heart failure: a pilot study. J Card Fail
Circ J 1992;56:35965. 2001;7:219.
[21] Ruskin JA, Gurney JW, Thorsen MK, et al. Detec- [35] Jensen KT, Carstens J, Ivarsen P, et al. A new,
tion of pleural eusions on supine chest radio- fast and reliable radioimmunoassay of brain na-
graphs. AJR Am J Roentgenol 1987;148:6813. triuretic peptide in human plasma. Reference val-
[22] Chait A, Cohen HE, Meltzer LE, et al. The bedside ues in healthy subjects and in patients with
chest radiograph in the evaluation of incipient dierent diseases. Scand J Clin Lab Invest 1997;
heart failure. Radiology 1972;105:5636. 57:52940.
[23] Butman SM, Ewy GA, Standen JR, et al. Bedside [36] Buckley MG, Markandu ND, Miller MA, et al.
cardiovascular examination in patients with severe Plasma concentrations and comparisons of brain
chronic heart failure: importance of resting or in- and atrial natriuretic peptide in normal subjects
ducible jugular venous distention. J Am Coll Car- and in patients with essential hypertension.
diol 1993;22:96874. J Hum Hypertens 1993;3:34550.
586 PEACOCK

[37] Wallen T, Landahl T, Herner T, et al. Brain natri- hormonal activation. Eur J Clin Invest 1998;28:
uretic peptide in an elderly population. J Intern 63642.
Med 1997;242:30711. [52] Tsutamoto T, Wada A, Maeda K, et al. Plasma
[38] Sayama H, Nakamura Y, Saito N, et al. Relation- brain natriuretic peptide level as a biochemical
ship between left ventricular geometry and brain marker of morbidity and mortality in patients
natriuretic peptide levels in elderly subjects. Geron- with asymptomatic or minimally symptomatic left
tology 2000;46:717. ventricular dysfunction. Eur Heart J 1999;20:
[39] Lernfelt B. Aging and left ventricular function in el- 1799807.
derly healthy people. Am J Cardiol 1991;68:5479. [53] Yu CM, Sanderson JE. Plasma brain natriuretic
[40] Kohse KP. Dierential regulation of brain and peptidedan independent predictor of cardiovascu-
atrial natriuretic peptides in hemodialysis patients. lar mortality in acute heart failure. Eur J Heart Fail
Clin Nephrol 1993;40:8390. 1999;1:5965.
[41] McCullough PA, Duc P, Omland T, et al. B-type [54] Hammerer-Lercher A, Neubauer E, Muller S, et al.
natriuretic peptide and renal function in the diag- Head-to-head comparison of N-terminal pro-brain
nosis of heart failure: an analysis from the Breath- natriuretic peptide, brain natriuretic peptide and N
ing Not Properly Multinational Study. Am J terminal pro-atrial natriuretic peptide in diagnos-
Kidney Dis 2003;41(3):5719. ing left ventricular dysfunction. Clin Chim Acta
[42] Cheng V, Kazanagra R, Garcia A, et al. A rapid 2001;310:1937.
bedside test for B-type peptide predicts treatment [55] Harrison A, Morrison LK, Krishnaswamy P, et al.
outcomes in patients admitted for decompensated B-type natriuretic peptide predicts future cardiac
heart failure: a pilot study. J Am Coll Cardiol events in patients presenting to the emergency de-
2001;37:38691. partment with dyspnea. Ann Emerg Med 2002;39:
[43] Tang WH, Girod JP, Lee MJ, et al. Plasma B-type 1318.
natriuretic peptide levels in ambulatory patients [56] Arroliga AC. Noninvasive positive pressure venti-
with established chronic symptomatic systolic heart lation in acute respiratory failure: does it improve
failure. Circulation 2003;108(24):29646. outcomes? Cleve Clin J Med 2001;68:67780.
[44] Niinuma H, Nakamura M, Hirarnori K. Plasma B- [57] Peacock WF, Emerman CL, Costanzo MR, et al.
type natriuretic peptide measurement in a multi- Early initiation of intravenous therapy improves
phasic health screening program. Cardiology heart failure outcomes: an analysis from the AD-
1998;90:8994. HERE registry database. Ann Emerg Med 2003;
[45] Davis M, Espiner E, Richards G, et al. Plasma 42(4):S26.
brain natriuretic peptide in assessment of acute [58] Peacock WF 4th, Albert NM. Observation unit
dyspnoea. Lancet 1994;343:4404. management of heart failure. Emerg Med Clin
[46] McDonagh TA, Robb SD, Morton JJ, et al. Bio- North Am 2001;19:20932.
chemical detection of left-ventricular systolic dys- [59] Peacock WF, Aponte JH, Craig MT, et al. Inpa-
function. Lancet 1998;351:913. tient versus emergency department observation
[47] Maisel AS, Krishnaswamy P, Nowak RM, et al. unit management of heart failure. Ann Emerg
Breathing Not Properly Multinational Study Med 1998;32:S46.
Investigators. Rapid measurement of B-type natri- [60] Fonarow GC, Adams KF Jr, Abraham WT, et al.
uretic peptide in the emergency diagnosis of heart The ADHERE Scientic Advisory Committee,
failure. N Engl J Med 2002;347:1617. Study Group, and Investigators. Risk stratication
[48] Bettencourt P, Ferreira A, Dias P, et al. Evaluation for in-hospital mortality in acutely decompensated
of brain natriuretic peptide in the diagnosis of heart heart failure: classication and regression tree anal-
failure. Cardiology 2000;93:1925. ysis. JAMA 2005;293:57280.
[49] Fujimora M, Yasumura Y, Ishida Y, et al. Im- [61] Diercks DB, Kirk JD, Peacock WF, Weber JE.
provement in left ventricular function in response Identication of emergency department patients
to carvedilol is accompanied by attenuation of neu- with decompensated heart failure at low risk for ad-
rohumoral activation in patients with dilated car- verse events and prolonged hospitalization. J Car-
diomyopathy. J Card Fail 2000;6:310. diac Failure 2004;10(4):S114.
[50] Nishikimi T, Matsuoka H, Ishikawa K, et al. Anti- [62] Burkhardt J, Peacock WF, Emerman CL. Eleva-
hypertensive therapy reduces increased plasma lev- tion in blood urea nitrogen predicts a lower dis-
els of adrenomedullin and brain natriuretic peptide charge rate from the observation unit. Ann
concomitant with regression of left ventricular hy- Emerg Med 2004;44(4):S99.
pertrophy in a patient with malignant hyperten- [63] Rich MW, Beckham V, Wittenberg C, et al. A mul-
sion. Hypertens Res 1996;19:97101. tidisciplinary intervention to prevent the readmis-
[51] Selvais P, Donckier J, Laloux R, et al. Cardiac na- sion of elderly patients with congestive heart
triuretic peptides for diagnosis and risk stratica- failure. N Engl J Med 1995;333:11905.
tion in heart failure: inuences of left ventricular [64] Rich MW, Vinson JM, Sperry JC, et al. Preven-
dysfunction and coronary artery disease on cardiac tion of readmission in elderly patients with
HEART FAILURE MANAGEMENT AND THE ED OBSERVATION UNIT 587

congestive heart failure. J Gen Intern Med 1993;8: Vasodilation in the Management of Acute Conges-
58590. tive Heart Failure. J Card Fail 2000;6(Suppl 2):182.
[65] Albert NM, Peacock WF. Patient outcome and [78] Colucci WS, Elkayam U, Horton DP, et al. Nesiri-
costs after implementation of an acute heart failure tide Study Group: intravenous nesiritide, a natri-
management program in an emergency department uretic peptide, in the treatment of decompensated
observation unit. J Int Soc Heart Lung Transplant congestive heart failure. Nesiritide Study Group.
1999;18:92. N Engl J Med 2000;343:24653.
[66] Vinson JM, Rich MW, Sperry JC, et al. Early read- [79] Silver MA, Ghali JK, Horton DP, et al. Eect of
mission of elderly patients with congestive heart nesiritide versus dobutamine on short-term out-
failure. J Am Geriatr Soc 1990;38:12905. comes in the treatment of acutely decompensated
[67] Quaglietti SE, Atwood JE, Ackerman L, et al. heart failure. J Card Fail 1998;4(Suppl 1):150A.
Management of the patient with congestive heart [80] Peacock WF, Emerman CE, Young J, on behalf of
failure using outpatient, home, and palliative the PROACTION study group. Safety and ecacy
care. Prog Cardiovasc Dis 2000;43:25974. of nesiritide for the treatment of decompensated
[68] Peacock W, Aponte J, Craig M, et al. Predictors of heart failure in emergency department observation
unsuccessful treatment for congestive heart failure unit patients. JACC 2003;4(6)(Suppl A):336A.
in the emergency department observation unit. [81] Colucci WS, Packer M, Bristow MR, et al. US Car-
Acad Emerg Med 1997;4:494. vedilol Study Group: carvedilol inhibits clinical
[69] Cohn JN, Johnson G, Ziesche S, et al. A compari- progression in patients with mild symptoms of
son of enalapril with hydralazine-isosorbide dini- heart failure. Circulation 1996;94:28006.
trate in the treatment of chronic congestive heart [82] Tsuyuki RT, Yusuf S, Rouleau JL, et al. Combina-
failure. N Engl J Med 1991;325:30310. tion neurohormonal blockade with ACE inhibi-
[70] Cohn JN, Archibald DG, Ziesche S, et al. Eect of tors, angiotensin II, antagonists and beta-blockers
vasodilator therapy on mortality in chronic conges- in patients with congestive heart failure: design of
tive heart failure: results of the Veterans Adminis- the Randomized Evaluation of Strategies for Left
tration Cooperative Study. N Engl J Med 1986; Ventricular Dysfunction (RESOLVED) pilot study
314:154752. (phase II). Eur Heart J 1998;19S:308.
[71] The SOLVD Investigators. Eect of enalapril on [83] The International Steering Committee. Rationale,
survival in patients with reduced left ventricular design, and organization of the Metoprolol CR/
ejection fractions and congestive heart failure. XL randomized intervention trial in heart failure
N Engl J Med 1991;325:293302. (MERIT-HF). Am J Cardiol 1997;80(Suppl 9B):
[72] The CONSENSUS Trial Study Group. Eects of 54J8J.
enalapril on mortality in severe congestive heart [84] CIBIS Investigators and Committees. A random-
failure: Results of the Cooperative North Scandi- ized trial of beta-blockade in heart failure: The
navian Enalapril Survival Study (CONSENSUS). Cardiac Insuciency Bisoprolol Study (CIBIS).
N Engl J Med 1987;315:142935. Circulation 1994;90:176573.
[73] Marcus LS, Hart D, Packer M, et al. Hemody- [85] Bristow MR, Gilbert EM, Abraham WT, et al.
namic and renal excretory eects of human brain MOCHA Investigators: carvedilol produces dose-
natriuretic peptide infusion in patients with conges- related improvements in left ventricular function
tive heart failure. A double-blind, placebo- and survival in subjects with chronic heart failure.
controlled, randomized crossover trial. Circulation Circulation 1996;94:280716.
1996;94:31849. [86] The Rales Investigators. Eectiveness of spirono-
[74] Abraham WT, Lowes BD, Ferguson DA, et al. Sys- lactone added to an angiotensin-converting enzyme
temic hemodynamic, neurohormonal, and renal inhibitor and a loop diuretic for severe chronic con-
eects of a steady-state infusion of human brain na- gestive heart failure (The Randomized Aldactone
triuretic peptide in patients with hemodynamically Evaluation Study [RALES]). Am J Cardiol 1996;
decompensated heart failure. J Card Fail 1998;4: 78:9027.
3744. [87] Cio G, Pozzoli M, Forni G, et al. Systemic
[75] Burger AJ, Horton DP, Elkayam U, et al. Nesiri- thromboembolism in chronic heart failure. A pro-
tide is not associated with the proarrhythmic eects spective study in 406 patients. Eur Heart J 1996;
of dobutamine in the treatment of decompensated 17:13819.
CHF: The PRECEDENT Study. J Card Fail [88] Baker DW, Wright RF. Management of heart fail-
1999;5(Suppl 1):178. ure. IV. Anticoagulation for patients with heart
[76] Elkayam U, Silver MA, Burger AJ, et al. The eect failure due to left ventricular systolic dysfunction.
of short-term therapy with nesiritide (B-type natri- JAMA 1994;272:16148.
uretic peptide) or dobutamine on long-term sur- [89] Turpie AG. Thrombosis prophylaxis in the acutely
vival. J Card Fail 2000;6(Suppl 2):169. ill medical patient: insights from the prophylaxis in
[77] Young JB, Stevenson LW, Abraham WT, et al. MEDical patients with ENOXaparin (MEDE-
Rationale and design of the VMAC Trial: NOX) trail. AJC 2000;86(12B):48M52M.
588 PEACOCK

[90] ACC/AHA. Guidelines for the evaluation and [95] Batsford WP, Mickleborough LL, Elefteriades JA.
management of heart failure: report of the Ameri- Ventricular arrhythmias in heart failure. Cardiol
can College of Cardiology/American Heart Associ- Clin 1995;13:8791.
ation task force on practice guidelines (Committee [96] The Cardiac Arrhythmia Suppression Trail
of Evaluation and Management of Heart Failure). (CAST) Investigators. Preliminary report: eect
Circulation 1995;92:276484. of encainide and ecainide on mortality in a ran-
[91] Elkayam U, Weber L, McKay C, et al. Spectrum of domized trial of arrhythmia suppression after
acute hemodynamic eects of nifedipine in severe myocardial infarction. N Engl J Med 1989;321:
congestive heart failure. Am J Cardiol 1985;56: 40612.
5606. [97] Albert NM. Manipulating survival and life quality
[92] Barjon JN, Rouleau JL, Bichet D, et al. Chronic re- outcomes in heart failure through disease state
nal and neurohumoral eects of the calcium entry management. Crit Care Nurse Clin North Am
blocker nisoldipine in patients with congestive 1999;11:12141.
heart failure. J Am Coll Cardiol 1987;9:62230. [98] Cline CMJ, Israelsson BYA, Willenheimer RB,
[93] Elkayam U, Amin J, Mehra A, et al. A prospective, et al. Cost eective management program for heart
randomized, double-blind, crossover study to com- failure reduces hospitalization. Heart 1998;80:
pare the ecacy and safety of chronic nifedipine 4426.
therapy with that of isosorbide dinitrate and their [99] Fonarow GC, Stevenson LW, Walden JA, et al.
combination in the treatment of chronic congestive Impact of a comprehensive heart failure manage-
heart failure. Circulation 1990;82:195461. ment program on hospital readmission and func-
[94] Goldstein RE, Boccuzzi SJ, Cruess D, et al. Diltia- tional status of patients with advanced heart
zem increases late-onset congestive heart failure in failure. J Am Coll Cardiol 1997;30:72532.
post-infarction patients with early reduction in [100] Gattis W, Hasselblad V, Whellan D, et al. Reduc-
ejection fraction. The Adverse Experience Commit- tion in heart failure events by the addition of a clin-
tee and the Multicenter Diltiazem Post-infarction ical pharmacist to the heart failure management
Research Group. Circulation 1991;83:5260. team. Arch Intern Med 1999;159:193945.
Cardiol Clin 23 (2005) 589599

Cost Eectiveness of Chest Pain Units


Sandra Sieck, RN, MBA
Sieck HealthCare Consulting, 9431 Je Hamilton Road, Mobile, AL 36695, USA

Health care is undergoing unprecedented CPUs have continued to evolve to better meet
change resulting from increasing pressures to the needs of the health care system and the acute
deliver quality care to greater numbers of people cardiac patient. More recently, as a result of
in a resource-constrained environment. Although changing reimbursement patterns, CPUs also
all providers of care are aected, the acute have been able to address some of the nancial
hospital facility bears an inordinate amount of pressures placed on the acute care facility by
this pressure, often caught between decreasing third-party payers.
global reimbursements and the need to care for
sicker and more complex patients. The Chest Pain Financial pressures on acute care facilities
Unit (CPU) represents a venue that eectively The Balanced Budget Act (BBA) of 1997 made
merges the clinical needs of a segment of acute sweeping changes to Medicare, generating savings
cardiac patients with the economic needs of the that were critical to extending the life of the
acute care facility. Medicare Trust Fund [2]. In the health care arena,
most of these changes impacted the acute care fa-
Origin of the chest pain unit cility most directly. Lacking a complete under-
standing of the impact of the BBA; however,
Historically, the outpatient CPU concept was many facilities failed to make the necessary
developed to heighten community awareness on changes to oset the new plan, resulting in overall
the early warning signs of a heart attack and decreased reimbursements in 1997. The xed reim-
create an ease of hospital entry for the chest pain bursements of the diagnosis-related group (DRG)
patient. CPUs experienced exponential growth program began to adversely aect hospitals reve-
from 1980 to 1998 but oered little more than nues and prot margins.
successful marketing strategies to increase hospi- In an eort to maintain economic viability,
tal admissions, a factor that often contributed to hospitals targeted two key areas for cost reduction
overcrowding in many community hospitals [1]. eorts: bed days and stang. Intense eorts were
After, the CPU served as a temporary holding made at reducing average length of stay (LOS) and
area, sorting out patients who were at low risk for reducing the number of hospital employees. Al-
developing a myocardial infarction. This triage though this strategy resulted in initial cost benets,
role allowed for reduced use of acute inpatient savings plateaued, and margins eroded further. By
stays by those patients who did not require a high 1999, DRG reimbursement increased only 0.6%,
level of care. This triaging allowed appropriate and hospitals overall average-operating margins
care to be provided to the low-risk chest pain had decreased to 2.76% [3]. These eroding prot
patient while maintaining open bed availability in margins, and a reduction in the number of acute
the intensive care unit (ICU) for patients requiring care facilities as a result of mergers and acquisi-
a higher level of service. tions, created an unstable environment in which
hospitals were struggling to remain solvent [3].
As revenues were held in check, another factor
added to the burdens on the acute care facility:
E-mail address: ssieck@sieckhealthcare.com health care use continued to increase. During the
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.001 cardiology.theclinics.com
590 SIECK

decade leading to 2003, inpatient admissions in- The ED itself is not the cause of overcrowding;
creased nearly 10% while outpatient visits in- rather, it is simply the unit most vulnerable to
creased approximately 61% [3]. The number of gridlock (Fig. 1) [4]. The lack of inpatient beds
emergency department (ED) visits increased is the most commonly cited reason for crowding
steadily from 1997 to 2002, from 99.6 million to in the ED [4,5]. Although most chest pain patients
114.2 million [3]. This immense growth and de- are unlikely to have sustained a myocardial infarc-
mand on emergency services led to an overwhelm- tion or need treatment in an ICU, the high rate of
ing reduction in the availability of these services acute hospital admission adds to the bed-day bur-
and is considered a major contributor to the pro- den of the acute care facility.
cess and economic changes that are now occurring. Acute coronary syndrome (ACS) is a poten-
The ED is acknowledged as the most common tially lethal condition that aects approximately
point of entry into the health care system. As 2 million people in the United States. Most ACS
facilities begin to factor in the expected 8 million patients seek care through an ED, and this patient
annual ED visits for chest pain predicted in the volume is saturating these departments. Treat-
upcoming years, the emergence of the CPUd ment delays, medical errors, poor outcomes,
often termed ED Observation Unitdand other against medical advice, higher re-admissions, di-
specialized short stay units are now under re- versions, increasing costs, and threatened quality
consideration to help alleviate some of the hospi- all play a role in the health care crisis.
tal system burdens, such as ED overcrowding, ACS patients use a tremendous amount of
reduction of unnecessary admissions, and the resources initially in the ED, only to be admitted
opportunity to optimize revenue. into an inpatient bed later for a 2- to 3-day LOS
in the acute facility. From a DRG reimbursement
perspective, all testing done in the ED arena is
Acute coronary syndrome
now debited from the total DRG payment for
Current bottlenecks in the ED are caused by which the patient was admitted. If the admitting
the large, challenging group of patients who pres- diagnosis DRG is not changed from DRG 143
ent with chest pain of possible cardiac origin, or (chest pain) to a higher acuity DRG 140 (unstable
low-probability patients as categorized by the angina), it will most likely contribute negatively to
America Heart Association and the American the hospital bottom line (opportunity revenue loss
College of Cardiology (ACC/AHA) guidelines. of $1500$2000).

Outpatient Post BBA 1997 Inpatient


Supply

ED Decrease Reimbursement
Post BBA
** Most Vulnerable Unit Decrease Workforce
Demand for Gridlock Y2020: 200,000 physicians
Y2020: 800,000 nurses
Increase Uninsured Decrease Ancillary Services
EMATLA Lab Techs
Increase Underinsured EKG Techs
Insurance Bail-out Radiology Techs
Increase Acuity Decrease # Staffed Beds
Aging Population >100 occupancy, not capacity
PCP Access Time
Decrease in Coverage

Fig. 1. Comparison between supply and demand aecting the health care systems access to care, hospital overcrowding
and eciency demands. Hospital overcrowding was considered a unit phenomenon tied to the emergency department.
Issues surrounding the causes and solutions, overcrowding is increasingly perceived as a system problem. (Adapted from
Joint Commission Resources. Managing patient ow: strategies and solutions for addressing hospital overcrowding.
2004. p. 11.)
COST EFFECTIVENESS OF CHEST PAIN UNITS 591

Emergence of the new chest pain unit diagnose and treat, and essentially all are ad-
mitted to the acute care facility. They are often
The CPU is no longer merely a holding area
dicult to distinguish from non-ACS patients
for low-risk cardiac patients or a conduit to the
who present with chest pain.
inpatient service for the higher-risk group. A new
The ACC/AHA guidelines highlight the role of
CPU concept is emerging as a venue for address-
CPUs and short-stay ED coronary care units to
ing some of the previously mentioned clinical and
facilitate a more denitive evaluation while avoid-
economic pressures that face the acute care
ing the unnecessary hospital admission of patients
facility.
who have not developed ACS. Such evaluation
Use of a CPU can show resource advantages
also helps to avoid the inappropriate discharge of
over the typical acute hospital admission. The
patients who develop active myocardial ischemia
typical chest pain patient spends an average of 2.1
without ST-segment elevation. Personnel in these
days in the hospital; when seen and treated only in
units use critical pathways or protocols designed
the CPU, this same patient can be discharged
to arrive at a decision about the presence or
safely in 6 to 12 hours. Each outpatient bed frees
absence of myocardial ischemia and, if present, to
2.2 to 3.5 inpatient beds that can be allocated to
characterize it further as UA or NSTEMI and to
high-acuity patients who require admission [6].
dene the optimal next step in the care of the
Additionally, quality measures can be improved
patient (eg, admission, acute intervention). The
also. For instance, the Impact on the Care of
goal is to arrive at such a decision after a nite
the Emergency Department Chest Pain Patient
amount of time, which usually falls between 6 and
from the Chest Pain Evaluation Registry
12 hours but may extend up to 24 hours, depend-
(CHEPER) study reported a signicant reduction
ing on the policies in individual hospitals. The
in missed myocardial infarctions from 4.5% to
time spent shifts from days of care to hours of
0.4%, a change that enhances quality and directly
care.
impacts the high-cost medical/legal payouts and
The dierence between NSTEMI and UA also
re-admissions from erroneous diagnoses [7].
brings to light an important economic factor. If
Observation units for heart failure recently
a physician diagnoses a patient who presents with
have shown a positive impact on reducing after-
elevated biomarkers with ST depression but fails
health-care use in patients who present with acute
to document the elevation of these cardiac
decompensated heart failure. Table 1 summarizes
markers, the facility is reimbursed as a UA pay-
data from the Cleveland Clinic experience [8].
ment. The dierence in reimbursement for the
NSTEMI and UA patients is approximately
Future directions in acute coronary syndrome $4000 to $6000. Physician documentation can
NonST-segment elevation myocardial infarc- thus play a vital role in how the facilities and oce
tion (NSTEMI) and unstable angina (UA) pa- practice coders submit for reimbursement.
tients represent the next challenging focus for CPUs are able to advance the STEMI and
optimizing cost-ecient acute cardiac care. These NSTEMI/UA patients through the system quickly
patients present to facilities earlier in the throm- and consistently while promptly facilitating safe
bosis progression model of the disease than discharge of the low-probability chest pain pa-
STEMI patients, making them more dicult to tients. The ultimate goal of the CPU is to
transition many of the services now performed
in the ICU setting to the CPU for earlier diagnosis
Table 1 and managementdthe true short-stay coronary
Observation unit results for patients with heart failure care unit.
Utilization Change
parameter after OU Percent change
Reimbursement issues
Revisits 90%51% 44% decrease
ED observation 9% Centers for Medicare and Medicaid Services
discharges (CMS) is the third-party payer for most of the
HF rehospitalizations 77%50% 36% decrease acute cardiac patients who are largely of Medicare
Observation 23% decreased age. CMS is now trying to shift unnecessary
hospitalizations LOS inpatient volume back to the point of entry
Abbreviations: HF, heart failure; OU, observation for more ecient risk stratication, assessment,
unit. and more intense treatment. This approach
592 SIECK

emphasizes the integration of observation unit 0.001) and lower nal rate of hospital admission
services as a viable alternative to inpatient health (47% versus 57%, P ! 0.001) [7]. These clinical
care delivery and is being accomplished by shift- benets were accompanied by an average saving
ing resourcesdor reimbursementdto the outpa- of more than $120 per patient (almost $2.9 million
tient setting, most commonly the ED in a CPU or reduction in true costs for 23,407 patients seen in
observation unit. the ED for chest pain) [7]. In a prospective, ran-
The facility payment for a CPU service is paid domized trial of patients seen at a CPU, no cases
under a new observation code, the Ambulatory existed of missed primary cardiac events and no
Patient Classication (APC) 0339. APC 0339 is erroneous discharges, and use of resources was
a separate observation carve-out payment that more ecient when ACS patients were seen at
CMS has agreed to pay as a third door opportu- the CPU than when they were admitted by way
nity for acute care facilities. One intention of APC of the ED to a cardiac care unit [10]. Practice
0339 is to oer an alternative pathway to di- standards specic to the institution represent an-
agnose, treat, and admit or discharge patients other method of improving quality of care while
rapidly who are diagnosed with chest pain, con- controlling cost. Practice standards can promote
gestive heart failure, and asthma. CMS selected adherence to regulatory directives, cost-contain-
these three medical conditions to focus on four ment measures, and evidence-based guidelines
mains areas of concern: (1) morbidity, (2) mor- for diagnosis and treatment.
tality, (3) appropriateness of admissions, and (4) Hospitals have also reduced costs of treatment
reduction in inappropriate discharges. The current for inpatient stays for low-risk chest pain. Fig. 2
shift of this large class of patients from their shows the eect of a quality-improvement initia-
traditional inpatient treatment protocol to a more tive in scal year (FY) 2000 related to ACS pa-
ecient plan of care initiated in the CPU or ED tients. As part of the initiative, the acute care
observation unit has resulted in a signicant facility modied treatment protocols to adhere
economic impact on the acute care facility [9]. to the AHA/ACC guidelines for the management
The CPU is shown to impact economics and of ACS. From FY 2000 to 2002, inhospital mor-
clinical outcomes in several studies. One multi- tality of ACS patients decreased from 4.8% to
center study demonstrated that observation in 1.9%, and LOS decreased from 5.9 days to 4.6
a CPU resulted in a lower incidence of missed days. The average cost per case dropped from
myocardial infarction (0.4% versus 4.5%, P ! $11,777 to $10,623, saving $1,154 per ACS

ACC/AHA Guidelines in Practice


Improved Patient Care and Reduced Cost:
Results of Baptist Health Systems
Acute Coronary Syndromes Project

Quality Improvement Initiative: Triage to Dschg


Mortality FY 2000 ----- 2002

4.8 ----- 1.9


Length of Stay(LOS) 5.9 ----- 4.6
Cost per Case $11,777 ----- $10,623
Patient Satisfaction Increased ----- 6

**CRUSADE contributed to Journal of Cardiovascular Management, 2004


concurrent Data Mining

Fig. 2. Healthcare facilities increasing must provide high quality medicine at reasonable cost. (Adapted from Jackson S,
Sistrunk H, Staman K, et al. Improved patient care and reduced costs: results of Baptist Heal Systems acute coronary
syndromes project. J Cardiovasc Manag 2003;14:4.)
COST EFFECTIVENESS OF CHEST PAIN UNITS 593

patient. These results demonstrate that institu- risk-stratication protocols to appropriate inpa-
tions can examine care patterns, change them, tient admissions while reducing missed myocar-
and maintain quality clinical outcomes [11]. dial infarctions and providing safe discharges with
Despite these cost savings, the DRG reimburse- fewer returns. Careful up-front risk stratication
ment system does not always result in full coverage not only results in increased eciency of care, but
of costs of care for an inpatient stay (Fig. 3). The also in cost advantages through a more favorable
following example demonstrates the full impact of reimbursement situation.
reimbursement dierences between DRG and Services, such as CPUs, can be structured to
APC coding for chest pain evaluation and treat- provide the resources necessary at the point-of-
ment in a hospital. entry to diagnose and treat ACS patients eciently,
The hospital analysis is based on 4000 chest directing them into the proper service lines at
pain patients, 289 of which qualied for APC maximum eciency. High-acuity and moderate-
0339 outpatient observation versus an inpatient risk patients are now admitted, and low-probabil-
admission using DRG 143. In a 12-month retro- ity patients, who in the past would have been
spective analysis, the patients who were routed admitted, are diagnosed and released safely after
through the CPU generated a prot of $32,000 only hours versus the past pattern of days. The
for the hospital, whereas the patients admitted study performed by Ross on the impact of an ED
under DRG 143 resulted in a loss, costing the observation bed on inpatient bed availability [12]
system $188,000. By merely transferring to indicates that 1 patient treated in an observation
observation and at the time meeting minimal unit bed equals 2.2 to 3.5 inpatient beds. This equa-
testing to meet the APC 0339 payment require- tion is one of the variables that contributes to the
ments, whereas in the outpatient CPU versus economic feasibility of CPUs. This 3:1 ratio in-
admitting to the inpatient system, a $220,000 creases the physicians productivity while driving
variance was created. Fixed cost and variable cost down xed and variable costs allocated to the out-
was factored into these calculations to give a patient services.
more accurate representation of the results.
Although this scenario does not reect every
Federal regulations and the eect on chest pain
hospitals situation, it underscores the potential
units in 2005
adverse nancial implications of providing higher-
intensity care than is medically necessary. An The federal regulations continue to change to
evidence-based, protocol-driven CPU has the accommodate the shift in acuity to the outpatient
ability to route patients rapidly through rapid setting. These regulations do not specify bricks

APC 0339 DRG 143


Chest Pain Chest Pain
Level III/IV < 8 hours..$177 Average Reimbursement $1997
APC 0339 > 8 hours$350 ALOS __2.1 days_

Billable (UB92) Reimbursable


Billable (UB92) Reimbursable
EKG yes yes EKG yes bundled in DRG
Enzymes yes yes yes
Enzymes bundled in DRG
Nuclear Studies yes yes Nuclear Studies yes bundled in DRG
Stress Test yes yes Stress Test yes bundled in DRG

Total $900/$1000
Fixed Cost (FC)($570) Fixed Cost (FC)..($1498)
FC varies per institution FC varies per institution, usually 3 times OP
Mean Variable Cost (VC)...($1168)
Mean Variable Cost (VC).($252) All diagnostics are bundled under corresponding
OP diagnostic testing enhancing bottom line DRG

**PROFITABLE $32,000 **LOSER ($-188,000)

Fig. 3. Comparison of inpatient DRG 143 (chest pain) coverage versus outpatient APC 0339 (chest pain). DRG, diagnosis-
related group; APC, ambulatory patient classication; FC, xed cost; VC, viable cost; OP, outpatient.
594 SIECK

and mortar as part of the reimbursement criteria reducing reimbursement for those who do not
for APC 0339. A CPU or observation unit can be meet the dened indicators; participating in the
located within or near an ED or be a free-standing CMS/Premier Quality Demonstration Project ini-
unit. It is not the location of a CPU that dis- tiative; and experimenting with a sliding scale
tinguishes its services from an ED but the type of concept that pays based on the number of quality
care it provides. indicators facilities meet (Fig. 4).
These regulations are particularly inuential The pay-for-performance approach acknowl-
from a reimbursement standpoint. Financial in- edges the reality that nancial rewards are among
centives have been put into play to ensure ecient the most powerful tools for bringing about behav-
care is provided in the CPU. These changes are ior change [1]. Also, the alignment of payment pro-
more economically favorable to the hospitals that gram incentives to support the provision of safe,
employ CPUs. high-quality care is a complex undertaking because
Specic excerpts from the Federal Register it must simultaneously achieve fair reimbursement
on the recent changes in the APC reimbursement for necessary services, promote desired behavior
as of January 2005 are summarized in Table 2 [13]. change, and avoid unintended consequences. In
The main impacts of the regulatory changes the end, new payment policies and programs
are must work to the advantage of the patient and sup-
port the provision of patient-centered care.
 The duration of the CPU stay must be 8 to 48
In the near future, hospital reimbursement
hours.
likely will be directly proportional to quality
 If an admission occurs after the CPU stay, the
performances. CMS also has several eorts in
hospital is not nancially disadvantaged.
progress to provide hospital quality information
 Previously required laboratory tests have
to consumers and others to improve the care
been eliminated as part of the APC require-
provided by the nations hospitals. These initia-
ment (requires medical necessity).
tives build on previous CMS and quality initiative
 Any tests performed (medically necessary) as
strategies to identify illnesses or clinical conditions
part of a CPU stay can be billed for in addi-
that aect Medicare beneciaries to promote the
tion to the APC 0339 charge.
best medical practice associated with the targeted
clinical disorders; prevent or reduce further in-
stances of these selected clinical disorders; and
Quality implications of chest pain units
prevent related complications [2].
In an eort to support the dramatic restructur-
ing of health care delivery, CMS has also re-
Trans-theoretic Y model application
vamped its image from an institution of strictly
nancial reimbursement to include a healthy plan The separation of quality and nance is no
of quality mixed into its evolving plan of health longer a viable plan of operation and must be
care delivery. CMS is crafting new aggressive redesigned to create an ecient plan focused on
approaches to facilitate implementation of a pro- process to optimize clinical and nancial out-
active strategy to provide resources for care aimed comes. The business world has long recognized
at earlier detection and treatment of acute cardiac this tenet and has created models by which
conditions. In recent years, however, leaders and business units are aligned eciently and strategi-
policy-makers have directed increased attention cally toward achieving a well-dened end product
to strategies for achieving system-wide improve- or service.
ments in health care quality and patient safety A new approach to understanding the entire
that will lead to larger-scale, more rapid changes concept of health care delivery from point-of-
in professional and provider behavior than have entry to discharge is the trans-theoretic Y-model
been experienced to date. (Fig. 5). This model emphasizes a multidisciplinary
Protocols designed around the ACC/AHA team approach to align the care units that aect
guidelines will allow facilities to meet or exceed a cardiac patients progress through the current
certain quality indicators developed for these ACS system (Fig. 6). By understanding how each care
patients, also creating greater patient satisfaction. units operational strategies aect each subse-
CMS is developing pay-for-performance pro- quent care unit from point of entry to discharge,
grams and a bonus reimbursement for meet- a seamless transfer of patient care in outpa-
ing quality standards in certain clinical areas; tient and inpatient settings can optimize quality
COST EFFECTIVENESS OF CHEST PAIN UNITS 595

Table 2
Summary of federal regulation changes in 2005 [12]
Excerpts Analysis/Comments
1. Beginning in early 2001, the APC Panel began APC 0339, outpatient observation, was developed in
discussing the topic of separate payment for 2001 and implemented in April 2002 as a clinical and
observation services. In its deliberations, the APC nancial opportunity for facilities. This new
Panel asserted that observation services following outpatient observation diers from the traditional
clinical and emergency room visits should be paid inpatient observation. Criteria, such as the inclusion
separately, and that observation following surgery of an E&M level code, must be billed with one of
should be packaged into the payment for the three conditions (chest pain, CHF, and asthma).
surgical procedure. For CY 2002, separate payment Once the patient has been in observation for
for observation services (APC 0339) was minimum of 8 hours up to a maximum of 48,
implemented under the OPPS for three medical beginning with the nurses admission note and
conditions: chest pain, CHF, and asthma. A number ending at the time on the physicians discharge
of accompanying requirements were established: the orders. Physician must document progress notes
billing of an evaluation and management visit in during beneciarys stay in observation. Diagnostic
conjunction with the presence of certain specied testing requirements have changed from 2002 to
diagnosis codes on the claim; hourly billing of 2005 and are discussed later in Analysis # 2.
observation care for a minimum of 8 hours up to
a maximum of 48 hours; timing of observation
beginning with the clock time on the nurses
admission note and ending at the clock time on the
physicians discharge orders; a medical record
documenting that the beneciary was under the care
of a physician who specically assessed patient risk
to determine that the beneciary would benet from
observation care; and provision of specic
diagnostic tests to beneciaries based on their
diagnoses.
2. CMS received comments from the community and Specifying which diagnostic tests must be performed as
the APC Panel asserting that the requirements for a prerequisite for payment of APC 0339 may be
diagnostic testing are overly prescriptive and imposing an unreasonable reporting burden on
administratively burdensome, and that hospitals hospitals and may, in some cases, result in
may perform tests to comply with the CMS unnecessary tests being performed.
requirements, rather than based on clinical need.
Therefore, beginning in CY 2005, the authors are
proposing removal of the current requirements for
specic diagnostic testing, and reliance on clinical
judgment in combination with internal and external
quality review processes to ensure that appropriate
diagnostic testing (which the authors expect would
include some of the currently required diagnostic
tests) is provided for patients receiving high-quality,
medically necessary observation care.
Accordingly, beginning in CY 2005, the following tests
should no longer be required to receive payment for
APC 0339 (observation):
For congestive heart failure, a chest radiograph
(71010, 71020, 71030), and ECG (93005), and pulse
oximetry (94760, 94761, 94762).
For asthma, a breathing capacity test (94010) or pulse
oximetry (94760, 94761, 94762).
For chest pain, two sets of cardiac enzyme tests, two
CPK (82550, 82552, 82553) or two troponins (84484,
84512), and two sequential ECGs (93005).
(continued on next page)
596 SIECK

Table 2 (continued )
Excerpts Analysis/Comments
3. Hospitals and the APC panel further suggested that In an eort to reduce hospitals administrative burden
the method for accounting for the beneciarys time related to accurate billing, counting time in
in observation care be modied. observation care should end at the time the
outpatient is actually discharged from the hospital
or admitted as an inpatient. Twentyfour hours is
adequate for the clinical sta to determine what
further care the patient needs. In CY 2005, separate
payment should continue to be made for
observation care based on claims meeting the
requirement for payment of HCPCS code G0244
(observation care provided by a facility to a patient
with CHF, chest pain, or asthma, minimum 8 hours,
maximum 48 hours). However, do not include
claims reporting more than 48 hours of observation
care in calculating the nal payment rate for APC
0339. In CY 2005, OPPS payments for observation
care should increase over CY 2004 levels for two
reasons. First, the proposal to eliminate the
requirement that specic diagnostic tests be
performed to receive separate payment for
observation care will result in more observation
stays being paid for under APC 0339. Several CY
2003 claims with packaged observation services
reported for CHF, asthma, and chest pains would
qualify for separate payment. Adopting the
proposed changes, the volume of claims for payment
under APC 0339 should increase in CY 2005. This
volume increase, combined with the slightly higher
median cost calculated for APC 0339 based on CY
2003 claims, would likely result in higher aggregate
Medicare payments to hospitals for observation care
in CY 2005 than in previous years. The payment is
approximately $403.93, which translates into a 12%
increase over 2004.
4. The beneciary must be in the care of a physician The medical record must include documentation that
during the period of observation, as documented in the physician explicitly assessed patient risk to
the medical record by admission, discharge, and determine that the beneciary would benet from
other appropriate progress notes that are timed, observation care.
written, and signed by the physician.
Abbreviations: CHF, congestive heart failure; CY, calendar year; OPPS, outpatient prospective payment system;
E&M, evolution and management; CPK, creatine phosphokinase; HCPCS, healthcare common procedure coding
system.

improvement and positive economic value. With- plan can be implemented on any disease state,
out each care unit providing vital information to particularly chest pain or ACS patients. Beginning
others in this holistic approach, moving patients in the ED, this concept emphasizes an ecient,
eciently through the system is challenged. rapid assessment and action centered on a seam-
The trans-theoretic Y-model places an empha- less integration of ancillary services, such as the
sis on process improvement while targeting two laboratory, diagnostic imaging, and skilled nurs-
end points: quality and contribution margin ing while understanding the economic impacts
(Fig. 7). This concept begins at the point of entry on decisions made as the patient is directed
and ends at discharge and marries a clinical and through the system. The failure of any one of
nancial strategy that meets quality indicators these services could bring the ED to a halt, thus
while producing desirable prot margins. This supporting the notion that the ED is not
COST EFFECTIVENESS OF CHEST PAIN UNITS 597

Years 1 and 2 Year 3


100
+2 Bonus +2 Bonus
90
$ +1 Bonus $ +1 Bonus
80
70 Quality: Five Clinical Areas
60 1. Coronary artery bypass surgery
2. Heart attack
50
3. Heart failure
$
40
4. Hip and Knee replacements
30 5. Pneumonia
20
-1 Reduction
10 $ -2 Reduction
0
S. Sieck, Sieck HealthCare Consulting, Oct., 2003

Fig. 4. Principles for the construct of pay-for-performance.

necessarily the cause of the gridlock; rather, it is could become negative. The objective is to keep the
the unit that becomes most vulnerable to it [4]. contribution margin at its maximum without
Compare the hospital setting to an industrial compromising quality. If the quality team does
setting. Without uncertainty industrial facilities not communicate its concerns to the protability
can detail the route from raw material to nished team, then major issues of conict arise and the
product with pinpoint accuracy. Checks and bal- system does not operate at optimal eciency (see
ances and alternative plans for the unexpected exist Fig. 6). Without question, the top industry leaders
throughout the process. At every opportunity, can pick a point randomly within their process and
dollars and time are being shaveddcost-avoid- report on the variables within.
ance. The end product is priced to the market based Until health care delivery systems take an
on the operating costs within the process. If the internal look at their plan of operation and the
process varies and these costs or resource use climb, economic plan that coincides, reimbursement will
actions are taken to control the variances. If not, continue its current path. The Joint Commission
the contribution margin is eroded and eventually on Accreditation of Health care Organizations

Quality Cost

P Discharge
R
O
C
E
S
S
Point-of-Entry Sieck HealthCare Consulting, LLC

Fig. 5. Trans-theoretical Y Model: a new approach to understanding the entire concept of health care delivery from
point-of-entry to discharge.
598 SIECK

Chemical Company Hospital System

Quality Cost Quality Cost

S
BU Nursing
I
Radiology BU
X
BU BU Laboratory
S
I BU
BU
G
M Pharmacy
EKG
A RM
ED

Fig. 6. Comparison of hospital system, Zig-Zag model to industry leaders, Vertical model.

(JCAHO) is looking for eciency from point-of- current system is insucient. The principles used
entry to discharge, and CPUs can assist in this in the redesign of the CPU and the overall shift
process because the ED is the most common point of care to the outpatient setting represent initial
of entry into the system. JCAHO has developed stages in that direction.
a new standard, LD.3.15, to monitor the man-
agement of patient ow, attempting to alleviate
overcrowding, eective January 1, 2005. Its main
Summary
focus is identifying and mitigating impediments to
ecient patient ow throughout the hospital [4]. Acute care facilities are struggling to maintain
Health care is a service industry and is learning economic viability in an environment of cost-
to operate like an ecient and quality-driven constraints, declining reimbursements, and in-
business. The CPU represents an early step in creasingly expensive technology. Two-thirds of
this learning that can be applied throughout the facilities are operating at 1% or less, and one
point-of-care continuum. The Institute of Medi- third of all hospitals have reported operating
cine in its recent report Crossing the Quality margins in the negative [15]. To survive, hospitals
Chasm [14], concluded that in order for the health are being challenged to reinvent the way acute
system to meet the future demands of the popula- care is delivered. The experience and evolution
tion of the United States and provide quality care, of CPUs for the treatment of certain segments
drastic reinvention is required; trying to x the of the acute cardiac patient oer insight into the

Quality Cost
SIX NSTEMI/UA BONUS
SIGMA REIMB
Discharge
P Step Down
R Unit

O Cath Lab
C
E ACC/AHA Guidelines
Early ICU 24-48hrs
S Medical Therapy
Diagnosis
EKG
w/heart sounds
S LAB
Minimal Standards
POCT
Sieck HealthCare Consulting, LLC POE: ED

Fig. 7. Applying disease management to Trans-Theoretical Y model.


COST EFFECTIVENESS OF CHEST PAIN UNITS 599

development of new models of acute care that can [4] Joint Commission Resources. Managing patient
blend cost-eciency with quality. ow: strategies and solutions for addressing hospital
CPUs have taken on a new role with clinical overcrowding. 2004, p. 11.
outcomes and quality improvement at their core. [5] National Center for Health Statistics/Center for Dis-
ease Control and Prevention. National Hospital Dis-
The new federal regulations and APC coding
charge Survey. 1998, Series 12, No. 14. September,
provide re-allocation of resources and dollars to 2000.
outpatient services to identify and treat patients [6] Ross M, Wilson AG, McPherson M, et al. The im-
rapidly, thus enhancing the role of the CPU as pact of an ED observation unit on inpatient bed ca-
a cost-ecient entry for acute cardiac patients. pacity. Ann Emerg Med 2001;8:576.
The CPU oers an alternative clinical and [7] Gra LG, Dallara J, Ross MA, et al. Impact on the
nancial strategy to hospitals and physicians in care of the emergency department chest pain patient
the emerging health care reform model. Third- from the Chest Pain Evaluation Registry (CHEPER)
party payers will continue to assess the impact on Study. Am J Cardiol 1997;80(5):5638.
morbidity and mortality as the transition to out- [8] Peacock WF, Emerman CL. Management of acute
decompensated heart failure in the emergency
patient services progresses. In order for hospitals
department. J Am Coll Cardiol 2003;4(Suppl A):
to take advantage of outcomes-based nancial 336A.
incentives oered by the government, realignment [9] Sieck S. Best practice ACS. Journal of Critical Path-
of the business units must occur to maximize ways in Cardiology 2002;1(4):22330.
eciency by including all departments, such as [10] Farouh ME, Smars PA, Reeder GS, et al. A clinical
ancillary services. trial of a chest-pain observation unit for patients
Strategies that emphasize focused care-process with unstable angina. Chest Pain Evaluation in the
improvement initiatives while targeting quality Emergency Room (CHEER) Investigators. N Engl
and economic margins are the wave of the future. J Med 1998;339(26):18828.
The CPU represents the rst wave of the new [11] Jackson S, Sistrunk H, Staman K. Improved patient
care and reduced costs: results of Baptist Health Sys-
model for patients who experience chest pain and
tems acute coronary syndromes project. J Cardio-
heart failure. As this model is rened and shown to vasc Manag 2003;14:4.
impact cost and quality favorably, this experience [12] Ross M. Maximizing use of the emergency depart-
is likely to extend to other cardiac conditions. ment observation unit: a novel hybrid design. Ann
Emerg Med 2001;37:26774.
[13] Centers for Medicare and Medicaid Services. Medi-
References care: Hospital outpatient prospective payment sys-
tem and 2005 CY payment rates, 5044750973
[1] Bayley MD, Schwartz JS, Shofer FS, et al. The nan- [0418427]. Available at: http://www.access.gpo.
cial burden of ED congestion and hospital over- gov/su_docs/fedreg/a040816c.html. Accessed August
crowding for chest pain patients awaiting 16, 2004.
admission. Acad Emerg Med 2002;9:3678. [14] Berwick SM. A users manual for the IOMs Quality
[2] Centers for Medicare and Medicaid Services. Chasm Report. Health A 2002;21:8090.
Available at: http://www.cms.hhs.gov/media/press/ [15] Advancing Health in America. Promises Under
release.asp?Counter351. Accessed February 22, Pressure: Pressures Threaten Prescriptions or
1999. Change; The Care for Hospital Payment Improve-
[3] Modern Healthcare, By the Numbers: Health of our ment. Available at: www.mhanet.org/i4a/pages/
nations hospitals, 2004. p. 12. index/fcm?pageid459. Accessed May 21, 2003.
Cardiol Clin 23 (2005) 601614

Medical Error Prevention in ED Triage for ACS:


Use of Cardiac Care Decision Support and Quality
Improvement Feedback
Denise H. Daudelin, RN, MPHa,b, Harry P. Selker, MD, MSPHa,b,*
a
Tufts University School of Medicine, Boston, MA, USA
b
Institute for Clinical Research and Health Policy Studies, Tufts-New England Medical Center,
750 Washington Street, Box 63, Boston, MA 02111, USA

Acute coronary syndromes (ACS) is the most In ED triage in the United States each year, about
common serious condition requiring emergency 12,000 patients who present with AMI and 14,000
and acute care. Among the 8 million patients who present with unstable angina (UA) are mis-
who present to emergency departments (EDs) in takenly sent home from the ED [6], which nearly
the United States each year with symptoms con- doubles their expected mortality rates [6]. Reect-
sistent with a cardiac problem, about 25% prove ing this problem is that ED cases of missed ACS
to have actual ACS. About one third (8%) of perennially represent one of the largest cost cate-
the presenting group prove to have acute myocar- gories of adult malpractice claims in the United
dial infarction (AMI), of whom approximately States. In ED treatment of patients who present
40%, or 3% of the overall group, have early-stage with AMI, the lifesaving impact of coronary re-
AMI deserving reperfusion treatment. The prob- perfusion therapy is directly related to the timeli-
lem for the ED physician, therefore, is to identify, ness of implementation [7,8], yet many are not
triage, and treat promptly and accurately the treated promptly and about 90,000 per year are
small proportion of patients who require immedi- not treated at all. These errors in triage and treat-
ate emergency care while eciently dealing with ment for ACS are critical to the patient and occur
the great majority who do not have ACS. This is- on a scale that makes them a public health issue.
sue has been a focus of the time-insensitive predic- Thus, they present important opportunities to re-
tive instrument (TIPI) approach to improving ED duce medical errors.
triage and treatment decision-making [15] and Reducing these medical errors requires inno-
the TIPI Information System (TIPI-IS) feedback vative changes to patient care processes, existing
reporting system. equipment, and performance improvement activ-
In the care of such patients, errors are made; ities that can be achieved through the use of
thus, important opportunities exist for improve- information technology (IT).
ment in these ED triage and treatment decisions. Computerized physician order entry systems
and electronic medical error reporting software
are patient safety systems that require changes to,
This article was supported by Grant Numbers RO1 or additional layers in, the patient care workow.
HS/HL07360, RO1 HS 08212, and U18 HS11200 from These systems prove challenging to introduce
the Agency for Healthcare Research and Quality.
eectively and without disruption to patient
* Corresponding author. Institute for Clinical Re-
search and Health Policy Studies Tufts-New England
care. More easily introduced is IT that is seam-
Medical Center 750 Washington Street, Box 63 Boston, lessly built into workow and that enhances
MA 02111. existing patient care and performance improve-
E-mail address: hselker@tufts-nemc.org ment activities. Representing this approach, the
(H.P. Selker). conventional computerized electrocardiograph is
0733-8651/05/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ccl.2005.08.004 cardiology.theclinics.com
602 DAUDELIN & SELKER

an IT tool that can be adapted to improve patient applicable to all six hospitals. Using a pro-
safety by providing clinical decision support in grammed calculator, a patients 0% to 100%
real-time, without changing the patient care pro- probability of having ACI could be computed in
cess. The device also provides workow tools to 20 seconds. In a 2320-patient 11-month prospec-
prevent medical error and identies patients for tive controlled trial of the predictive instruments
inclusion in performance and outcome measure- impact on ED care, its use markedly improved
ment systems. physicians diagnostic specicity (P .002), with
no signicant change in sensitivity. The false pos-
itive diagnosis rate improved signicantly (P
Development of real-time decision support,
.004), whereas the false negative diagnosis rate
concurrent event alerts, and retrospective
did not change. Accordingly, ED triage disposi-
feedback reporting
tions for patients who proved to have ACI were
A series of decision support and risk reduction not changed from the appropriately high CCU ad-
tools for real-time use in the ED have been mission rates, but for patients without ACI, CCU
developed in addition to an information system admission rates dropped from 24% during the
to provide concurrent missed AMI event alerting control periods to 17%, and their ED discharge
and retrospective feedback reports for improve- rates to home increased from 44% during the con-
ment of the management of ACS patients. The trol periods to 51% (P .003), a 30% reduction in
core of this approach is the ability of a time- CCU admissions for patients without ACI. The
insensitive predictive instrument (TIPI) to com- predictive instruments need for a calculator, how-
pute, for every ED patient on presentation, a 0% ever, was cumbersome. Thus, for better ease of
to 100% probability that a patient truly has acute use and attractiveness to clinicians, the authors
cardiac ischemia (ACI), and if having an AMI, the designed the ACI-TIPI, its revised version, to be
likely outcome benets of thrombolytic therapy. A programmed into conventional computerized
key feature of this prediction is the time-insensi- electrocardiographs.
tive aspect of TIPIs, which refers to the fact
that they compute, for a given ED patient, the
Development of the acute coronary ischemia
same probability of value when used in the real-
time-insensitive predictive instrument for real-time
time ED setting or when used retrospectively for
and retrospective decision support
review of medical care. As now available in con-
ventional electrocardiographs, besides printed on Although improvement of CCU use might be
the ECG header as a decision aid for the physician done by real-time and retrospective interventions,
the computed TIPI probabilities are stored in the the latter has had little formal evaluation. Use of
electrocardiographs computer, can be saved in real-time test results and decision aids are more
a database, and then used to trigger concurrent familiar to clinicians than feedback systems. At-
ags about ongoing care and to generate retro- tractive and clinically relevant reports on triage
spective reports and feedback for clinicians. and treatment decisions, however, should facilitate
continuous quality improvement by emergency
medical services (EMS), physicians and institu-
Real-time decision support for emergency
tions self-assessment. (Also, payors and organiza-
department triage: the original acute coronary
tions responsible for hospital and CCU use for
ischemia predictive instrument
coronary disease need a fair and accurate retro-
The original ACI predictive instrument [1] was spective tool by which the appropriateness of
developed to be an easy-to-use method to improve CCU use can be measured.) Thus, there is a need
ED physician triage decisions so that fewer pa- for a clinically valid attractive measure for feed-
tients without ACI would be admitted to cardiac back on ACI/AMI care.
care units (CCUs), but without decreasing proper Ideally, a single tool should be able to serve
admission of those with ACI. Reasoning that prospective and retrospective purposes. Gener-
a 0% to 100% probability value might be incor- ally, however, tools designed for real-time use,
porated easily into physicians clinical decision- such as our original ACI predictive instrument,
making, based on prospectively collected data on have not been validated for retrospective medical
2801 patients seen in six divergent hospitals record review, and tools for retrospective review
EDs for 1 year, a logistic regression equation of care have not been applicable to real-time
was developed that used seven variables and was use, limiting clinicians interest and condence in
ACS PREVENTION IN EMERGENCY DEPARTMENTS 603

their use. Thus, the authors developed a time- exist by all United Stated EMS electrocardiograph
insensitive predictive instrument for ACI (ACI- manufacturers.
TIPI), valid for prospective real-time clinical use Besides enabling presentation of the ACI-TIPI
and retrospective medical record review, that ac- on the ECG, its incorporation into a computerized
curately computes a patients likelihood of having electrocardiograph allows transfer of patients
ACI at the time of presentation [2]. Not only the- ACI probabilities into a database, which is central
oretically desirable as a tool usable by clinicians, to the development of a performance measure-
administrators, and payors, the ACI-TIPI might ment and feedback system, the TIPI information
have the added benet of promoting cooperation system (TIPI-IS). Thus, the probability, ECG,
between these groups. This ACI-TIPI was tested and data entered into the electrocardiograph in
on the 2320 patients seen at the six study hospi- clinical use (age, sex, presenting symptoms, phy-
tals EDs during the second year, along with the sician, location) can be combined with other data
original instrument, and both were compared and used for reports and analyses. This informa-
with ED physician diagnostic performance [2]. tion allows clinicians and operations personnel of
Calculation of the receiver-operating characteris- EMS and hospital organizations to use the same
tic (ROC) curve areas, which simultaneously eval- clinically valid risk-adjusted outcome predic-
uate sensitivity and specicity of a continuous tions/measures (without medical record review).
scale test, yielded values of 0.880.89 for the
ACI-TIPI and the original instrument, demon-
The multicenter acute coronary ischemia
strating excellent and similar diagnostic perfor-
time-insensitive predictive instrument clinical trial
mance by both. Besides the ACI-TIPI, all
subsequent predictive instruments, for mortality In our 10-hospital, 10,689-patient clinical trial
from ACI/AMI [3], for mortality from congestive of the ACI-TIPI electrocardiographs impact on
failure [4], and the TPI [5] (see later discussion) ED triage of patients with symptoms suggestive of
meet TIPI criteria. ACI conducted in public, private, community,
and tertiary hospitals in urban, suburban, and
semi-rural areas, the ACI-TIPI improved ED
triage [11]. In doing so, it demonstrated dieren-
Incorporating acute coronary ischemia
tial impact depending on whether patients had
time-insensitive predictive instrument into
ACI or not, whether the hospital had high or
electrocardiographs to support real-time
low cardiac telemetry unit capacity, and whether
emergency department/emergency medical
the triaging ED physician was an unsupervised
services care
resident or not. Even as the ACI-TIPI reduced un-
Convinced that the predictive instruments necessary admissions, it did not reduce appropri-
probability would be used more were it automat- ate hospital and cardiac unit admission for
ically printed in the text header on ECGs (Fig. 1), patients with true ACI, either UA or AMI. This
the ACI-TIPI was programmed into a computer- result further conrmed the ACI-TIPIs safety
ized electrocardiograph [9]. To do this, an ACI- and eectiveness in ED use. Given that about
TIPI was written using Hewlett-Packard (now 6% of ED patients presenting with symptoms sug-
Philips) ElectroCardiographic Language [9,10]. gesting ACI prove to have stable angina, such re-
This activity was done in repeating iterative cycles ductions would correspond yearly in the United
of programming ECG waveform-based measure- States to approximately 30,000 fewer hospitaliza-
ment criteria and then comparing the programs tions and 20,000 fewer CCU admissions, and
performance on sample ECGs to readings by the about $728 million saved [11].
ECG reader who designed the original ACI-TIPI
ECG variables (HPS).
Studies of failure to hospitalize emergency
Because it was felt that the ACI-TIPI and the
department patients with acute coronary ischemia
approach of incorporating a predictive instrument
into a computerized electrocardiograph would not Among our studies of factors contributing to
benet a maximal number of patients if restricted, ED triage and treatment errors [1224], the au-
we published all details of the ACI-TIPIs formula thors have sought the causes for failing to hospi-
and the incorporation of a TIPI into an electrocar- talize ED patients who have ACI [6,25,26]. In
diograph was put into the public domain. Marketed the authors rst such study, based on the care
or prototype ACI-TIPI/TPI electrocardiographs at the hospitals that participated in the original
604 DAUDELIN & SELKER

Fig. 1. ACI-TIPI ECG header.

ACI predictive instrument trial, approximately under age 55 (6.7 times more likely) had a primary
2% of ED patients with AMI were sent home mis- symptom of shortness of breath rather than chest
takenly [25], most commonly related to problems pain (2.7 times more likely), or a normal or non-
in physician use of the ECG [6,25]. Among ED diagnostic ECG (3.3 times more likely, and 7.7
patients sent home with AMI, 35% had ECG ab- times more likely if having an AMI). These failures
normalities consistent with ACI noted by the phy- to hospitalize showed a statistical trend for greater
sician but not given sucient weight in the triage mortality: nonhospitalized patients with AMI
decision [25]. Additionally, 25% had ECG abnor- were 1.9 times more likely to die than similar pa-
malities suggesting AMI (ST elevation) that were tients who were hospitalized, and nonhospitalized
missed by the ED physician [25]. Analyses show- patients with UAP were 1.7 times more likely to
ing errors in physician ECG reading of ST seg- die than similar patients who were hospitalized.
ments and T waves contributing to suboptimal These ndings possibly reected physicians
ED triage supported these ndings [26]. overdependence on generalities. Among ED pa-
In the authors study of failure to hospitalize tients, it is true that ACI is less likely in younger
ED patients with ACI (AMI or UAP) in the ACI- women, African-Americans, those without chest
TIPI Trial hospitals, the results were consistent pain, and those with normal ECGs; but some such
with earlier ndings [6]. Among the 10,689 ED pa- patients do have ACI. Physicians must be careful
tients studied, of those with AMI (n 889), 2.1% not to over-generalize about certain groups, and
were sent home, and of those with UAP (n 966), for these patients, the ACI-TIPI and TPI may
2.3% were not hospitalized. Yearly in the United help (see later discussion of TPI Trial on use of
States, this corresponds to 26,000 ED patients coronary reperfusion therapy for women). The
with ACI mistakenly not hospitalized: 12,000 wide range in hospitals rates of failure to hospi-
with AMI and 14,000 with UAP. Of note, failure talize (0%11%) suggests that hospitals should
to hospitalize for ACI was more likely if the pa- monitor, and continually improve, their ED
tient was nonwhite (2.2 times more likely, and performance. Of note, chest pain centers did
4.5 times more likely if having an AMI), a woman not have better performance; it seems that what is
ACS PREVENTION IN EMERGENCY DEPARTMENTS 605

important is not the ED label, but that its To assess the potential impact of the risk
physicians fully evaluate the entire patient, and management form, the form was retrospectively
understand the proper use of diagnostic technol- applied to 20 cases that came to malpractice
ogies and receive feedback about diagnostic per- litigation. This information was also compared
formance [27,28]. with what the impact would be were the form
not automatically generated by the electrocardio-
graph, to see whether this approach was war-
Acute coronary ischemia time-insensitive
ranted. With the electrocardiograph-generated
predictive instrument risk management tool: an
version, 61% of cases were seen as much less likely
ECG-based form for avoiding medical errors
or certain to not come to malpractice litigation,
Based on these studies, cases of patients who versus 37% for the nonautomatic version (P
presented with AMI and mistakenly sent home .001). Moreover, for those that would have
were reviewed to devise a way to reduce such come to litigation, had the automatic electrocar-
errors. A risk management form was developed diograph-generated version been used, 80%
and integrated into ACI-TIPI software so the would have had a signicantly better outcome, in-
form is generated and partially lled-out auto- cluding 59% who were judged to be likely to have
matically by the electrocardiograph at the time of a better outcome (P .001). The provision of the
the patients initial ECG (Fig. 2) [29]. The intent ACI-TIPI probability should help the clinician
was that the form prompt consideration and doc- appreciate the importance of such ECG abnor-
umentation of the key clinical factors for such malities as contributors to the patients likelihood
cases, be immediately and conveniently available of having ACI. The clinical information needed to
for real-time use, and include the electrocardio- ll out the rest of the form should provide the
graphs computerized interpretation in its text, kind of documentation that better reveals appro-
along with the ACI-TIPI probability of ACI, to priate care, and the process of lling-in the items
lessen the likelihood that ECG abnormalities was created in the hopes that it deter suboptimal
and high-risk patients would be missed. care in the process of such documentation. Based

Fig. 2. ECG risk management form.


606 DAUDELIN & SELKER

on this, the authors worked with Hewlett-Packard are automatically computed and printed on the
(now Philips) to incorporate the automatic form ECG header. These predictions include probabili-
into their electrocardiograph. ties for acute (30-day) mortality if and if not treated
Financial implications of the impact of the with thrombolytic therapy; 1-year mortality rates if
form exist. For the 12 case records for which and if not treated with thrombolytic therapy; car-
complete nancial data were available, the legal diac arrest if and if not treated with thrombolytic
and processing expenses averaged $45,000 and therapy; thrombolytic therapy-related stroke and
settlements averaged $337,000 [29]. With the form major bleeding requiring transfusion.
automatically generated by the electrocardio- The TPI database, on which the TPIs ve
graph, the mean projected savings per case was component models were developed and tested,
$382,000, corresponding to $1.2 billion yearly sav- include the original data on patients who pre-
ings in the United States. These savings should sented with AMI from 13 clinical trials and
serve to make the ACI-TIPI approach attractive registries, including 107 hospitals of all types
to hospitals, especially those with their own cap- throughout the United States, totaling 4911 pa-
tive malpractice companies. tients [1,4557]. Separate logistic regression pre-
dictive instruments were developed for each
outcome. For each, clinically important and statis-
Development of the thrombolytic predictive
tically signicant variables were selected for
instrument to assist recognition and treatment of
preliminary models, alternative forms and com-
ST-elevation myocardial infarction
binations of these variables were investigated,
Emergent coronary reperfusion therapy, used and models reformulated to optimize performance
promptly, can be lifesaving for patients with ST- while preserving parsimony. This included creat-
elevation myocardial infarction (STEMI) [7,30 ing two special ECG variables to reect two deter-
32]. In emergency settings, however, this can be minants of the impact of thrombolytic therapy:
dicult, especially for less obvious candidates, a measure of AMI size and an indicator of earli-
and when key physician decision-makers are not ness in the AMIs course [56]. In addition, to sat-
on-site. Intensive eorts by physician leaders, the isfy the need for a regression method that could
National Heart Attack Alert Program, organiza- accommodate the rapidly changing inuence of
tions interested in quality of care, and others ECG-based infarct size, QT interval, and coronary
[3241], have helped increase use and promptness reperfusion therapy use in the rst hours of ACI/
of coronary reperfusion therapy [42]. Further im- AMI, a new method of regression was devised [58].
provement is needed [42,43], especially for other
than anterior STEMI, the category of AMI for
Thrombolytic predictive instrument clinical
which thrombolytic therapy was rst recognized
eectiveness trial: impact on use and promptness
as eective [7,30] and for women, who have re-
of coronary reperfusion therapy
ceived less coronary reperfusion therapy than
men [43,44]. Also, a need remains for ways to sup- To test whether the electrocardiograph-based
port prompt and accurate coronary reperfusion TPI improves ED selection of patients for coro-
therapy decisions in hospitals and prehospital nary reperfusion therapy and promptness of
EMS settings where consultation with o-site treatment, a 22-month randomized controlled
physicians is required. clinical eectiveness trial was run on the use and
To assist treatment decisions, the TPI was impact of thrombolytic therapy and overall cor-
developed, a collection of ve component pre- onary reperfusion therapy. Given our interest and
dictive instruments designed to accurately assess the demonstrated need for improvement in the use
the likely patient-specic benets and risks from of coronary reperfusion therapy [42,43], especially
the use of thrombolytic therapy for STEMI. This for other than anterior AMI [7,30] and for women
tool helps clinicians identify patients for coronary [43,44], our study hypotheses focused on these
reperfusion therapy based on their probabilities of groups, and also on hospitals where consultation
benets and complications and facilitates earliest with o-site physicians was required. Study end-
possible use of coronary reperfusion therapy points were percentages of patients receiving (a)
[5,45]. With manufacturers, programs were devel- thrombolytic therapy; (b) thrombolytic therapy
oped for conventional computerized electrocardio- within 1 hour of ED presentation; and (c) all cor-
graphs, so that when signicant ST-segment onary reperfusion therapy, either by thrombolytic
elevation of STEMI is detected, TPI predictions therapy or percutaneous transluminal coronary
ACS PREVENTION IN EMERGENCY DEPARTMENTS 607

angioplasty (PTCA). The trial ran in EDs at 28 group, the ECG had only the header text custom-
urban, suburban, and rural hospitals across the arily used in that ED.
United States, from major cardiac centers to small Collected data included: sociodemographic
community hospitals. data; initial and follow-up clinical features;
Included in the trial were all consenting pa- ECGs and cardiac biomarker test results; triaging
tients at least 35 years old presenting to any study physician training level, specialty, and whether
hospital with STEMI. At participating hospitals, ED-based; whether on-site or o-site (telephone)
software-generating TPI predictions was installed consultation was used for the coronary reperfu-
on conventional computerized electrocardio- sion therapy decision; whether the patient received
graphs. When a signicant ST-elevation charac- thrombolytic therapy or PTCA; and how long
teristic of STEMI was automatically detected, the after chest pain onset was the therapy received.
electrocardiograph randomly assigned the patient TPI software automatically acquired clinical var-
to the control or intervention group. If assigned to iables required for TPI calculations and ECG
the intervention group, the electrocardiograph Q-wave, ST-segment, and T-wave measurements.
automatically prompted the user to enter infor- Site physicians, blinded to study group, assigned
mation needed to compute the TPI predictions: conrmed diagnoses based on presentation, clin-
age, sex, history of hypertension or of diabetes, ical course, initial and follow-up ECGs, and
blood pressure, and time since ischemic symptom biomarker tests, using the World Health Organi-
onset. The remaining variables, based on ECG zation criteria [59]. Patients ED care was classi-
waveform measurements, were automatically ac- ed by whether consultation with an o-site
quired by the electrocardiograph. Then the ECG physician was used in making the treatment deci-
was printed with TPI predictions on its header sion. Hospital size, type, whether having on-site
(Fig. 3). If variables were missing, TPI predictions ED sta, and physician type were used as poten-
were not calculated, but an alert listing missing tial explanatory variables.
variables was printed, allowing entry of missing Of 2875 patients who presented with AMI
data, if available. For patients in the control at the participating hospitals, 1243 (43%) had

Fig. 3. TPI ECG header.


608 DAUDELIN & SELKER

ST-segment elevation. Of these, 1197 were ran- therapy and are more easily recognized in the pre-
domly assigned to study groups; 732 (61%) had cordial leads of the conventional ECG, and thus
inferior STEMI, and 465 (39%) had anterior physicians already do a good job of detecting
STEMI. Of patients with inferior STEMI in the them as appropriate candidates for coronary
control group, compared with the TPI group, reperfusion therapy, and treat them. Similarly,
61% compared with 68% (P .03) received AMI presents more classically in men and is
thrombolytic therapy; 53% compared with 59% already better recognized (and treated) in men
(P .08) received thrombolytic therapy within 1 than women [6,60,61]. The high rates of coronary
hour; and 68% compared with 75% (P .03) re- reperfusion therapy for anterior STEMI and men
ceived coronary reperfusion therapy overall, ei- when seen in experienced centers may require little
ther as thrombolytic therapy or primary PTCA. improvement. Less than 5% of the TPI Trials pa-
Of patients with anterior STEMI in the control tients, however, were seen in hospitals without on-
group compared with the TPI group, 60% com- site ED physicians, and no patients were included
pared with 54% (P O .2) received thrombolytic while in transit by way of EMS (when all physi-
therapy; 51% compared with 45% (P O .2) re- cian input would be remote). Thus, it is important
ceived thrombolytic therapy within 1 hour; and that the TPI did improve the speed and use of
68% compared with 64% of TPI patients (P O coronary reperfusion therapy among all patients
.2) received overall coronary reperfusion therapy. who presented with AMI (including anterior and
Among women (n 398) in the control group men) when the ED physician or needed physician
compared with the TPI group, 48% compared consultants are o-site. Because many hospitals
with 58% (P .03) received thrombolytic ther- EDs still are not staed 24 hours daily, and, espe-
apy; 41% compared with 48% (P .10) received cially in rural settings, long transport times may
thrombolytic therapy within 1 hour; and 56% mandate EMS use of thrombolytic therapy or
compared with 66% (P .04) received overall a decision for direct transport to one of the
coronary reperfusion therapy. Of patients who 20% of United States hospitals that are primary
required physician consultation by telephone PTCA-capable [62], it seems likely that the
(n 271) in the control group compared with TPI may improve coronary reperfusion therapy
the TPI group, 47% compared with 63% use for patients with anterior STEMI in these
(P .01) received thrombolytic therapy; 41% settings.
compared with 54% (P .04) received thrombo- The TPI is particularly attractive for EMS use.
lytic therapy within 1 hour; and 51% compared Prehospital thrombolytic therapy is indicated for
with 66% (P .01) received overall coronary re- situations in which transport times are long [63
perfusion therapy. Thus, the trial showed that 66], and the recent availability of single bolus
the TPI increased use of thrombolytic therapy, thrombolytic therapy [6769] and the increasing
use of thrombolytic therapy within 1 hour, and use of 12-lead ECGs in ambulances [7074]
use of overall coronary reperfusion therapy by provide a foundation for this approach [7577].
11% to 12% for patients with inferior STEMI, The provision of TPI predictions for patients
18% to 22% for women, and 30% to 34% for pa- with STEMI in the eld can support EMS use
tients with an o-site physician. The TPIs eect of thrombolytic therapy. Also, for patients who
was minimal on patients who exhibited high base- present with contraindications to thrombolytic
line coronary reperfusion therapyrates, such as for therapy, prehospital identication as a reperfusion
men who presented with anterior STEMIs. For candidate by the TPI can help avoid transport to
the targeted groups (those more often missed, a facility lacking primary PTCA capability, avoid-
women and those with less obvious STEMIs, ing the need for re-transport to a PTCA-capable
and situations in which physicians were o-site), center. Clarifying reperfusion needs and options
however, the TPI increased recognition of STEMI in the eld, especially in rural areas, can save
and use and timeliness of coronary reperfusion time, and improve patient outcomes [78]. That
therapy. the TPI increased overall coronary reperfusion
In the TPI Trial, it is of interest that the TPIs therapy use suggests it should facilitate decision-
impact was not seen among patients with anterior making for both types of reperfusion in prehospi-
STEMI, and when combining STEMI locations, it tal EMS use. Also, even with hospital-based
was not seen among men. This is thought to be coronary reperfusion therapy, advance notice by
because anterior STEMIs have received longer- EMS, as in the MITI Trial [63], prompted by
standing emphasis for coronary reperfusion the TPI, should reduce delays after arrival.
ACS PREVENTION IN EMERGENCY DEPARTMENTS 609

Time-insensitive predictive instrument information the ECG Management System data, ADT, Lab
system cardiac error reduction system based on and ICD9 diagnostic coding data into a data
acute coronary ischemia time-insensitive warehouse database; and (2) create a TIPI-IS web
predictive instrument demonstration project server to distribute TIPI-IS reports and allow
physicians (using a standard web browser) to
Having shown that ACI-TIPIs predictions on
explore the underlying patient data (Fig. 4). The
ECGs provide eective real-time decision support,
web-based user interface was fully redesigned af-
the TIPI-IS demonstration project was undertaken.
ter repeated usability testing and enhanced with
An ACI-TIPI-IS was created to measure the impact
additional outcome information, including con-
of real-time, concurrent, and retrospective ACI-TIPI
rmed diagnoses of ACI after inpatient evaluation
interventions on medical errors in ED triage. Also,
and whether patients had cardiac catheterization,
through a series of user tests, the systems attractive-
PTCA or coronary bypass surgery. The system
ness and use for various users was evaluated, and its
evaluation process included an iterative review
reports were modied to improve their usefulness in
with ED physicians leading to the redesign of
identifying and addressing medical error.
screens and reports. Reports for comparison and
In the ACI-TIPI-IS demonstration project [79],
benchmarking across groups of hospitals and for
the TIPI-IS provided concurrent and retrospective
drill-down within a hospital were created and
reporting and feedback for ED physicians and their
were additionally reviewed and revised by health
hospitals. Implemented and used at ve hospitals,
plans and malpractice insurers to ensure that the
the project provided a range of concurrent alerts
reports met the requirements for risk reduction
and feedback reports to over 70 physicians at these
and medical error prevention programs (Fig. 5).
facilities. This project has been successful in dem-
Implementation of the TIPI-IS had ve major
onstrating the use of the ACI-TIPI probabilities
components: (1) preparation of the ECG equip-
within an information system providing concurrent
ment and ECG computerized management system;
alerting and retrospective feedback reports.
(2) installation of the TIPI-IS server within each
A prototype TIPI-IS database and web server-
hospitals intranet; (3) linking and integrating each
based user interface was created to (1) aggregate

Fig. 4. TIPI-IS diagram. (Courtesy of Clinical Care Systems, Inc., Bedford, MA.)
610 DAUDELIN & SELKER

Fig. 5. TIPI-IS feedback report. (Courtesy of Clinical Care Systems, Inc., Bedford, MA.)

hospitals key computerized databases through which was successfully demonstrated, is the ability
electronic data interfaces to the TIPI-IS server; to compile large amounts of information for feed-
(4) testing of interface data; and (5) user training. back reporting and error reduction through an au-
Testing interfaces from feeder systems to the tomatic electronically compiled database. This
TIPI-IS at each site was conducted to ensure process leveraged the clinical information systems
TIPI-IS captured data completely and mapped already in place and minimized the resources re-
key data elements accurately for subsequent re- quired to collect and analyze data.
porting. Electronic data collection eased the re- During the intervention period, ED QI nurses
sources required to compile a large database, but and physicians were given a general overview of
also posed challenges in collecting accurate data the project and training in the use of the TIPI-
because it depended on the diligence of sta IS, the content and signicance of the concurrent
entering the data for operational purposes rather alerts and feedback reports. These meetings were
than research purposes. conducted with the ED physician leader who dis-
Access to data and reports through the web- tributed physician-specic feedback reports with
based interface improved availability of data, and information from baseline data collection and a se-
despite user testing and redesign of the original cure password to access their reports online.
system, the web-based system continued to be Concurrent alerts were designed within the
available for use. Frequent users of the system database system to signal patient cases that met
were project sta and ED department Quality criteria for immediate follow-up to prevent possi-
Improvement (QI) sta who had operational ble missed diagnosis of ACI. These thresholds
needs to access the data. When new reports or in- included patients sent home from the ED with an
formation were posted on the website, each physi- ACI-TIPI value in the high-risk category or with
cian received an e-mail with the link to the web a positive cardiac biomarker. The alert triggered
page. A key advantage of the TIPI-IS approach, an e-mail or pager message to the ED and ED
ACS PREVENTION IN EMERGENCY DEPARTMENTS 611

physician indicating a patient required follow-up/ who presents with ACS or to hospitalize a patient
review (Fig. 6). The results of follow-up on these who does not prove to be experiencing ACS to the
alerts were entered into the alert follow-up screen. CCU, and errors in treatment, such as the failure
Patients were contacted to return to the EDs if to promptly use reperfusion therapy for patients
needed and hospitalized for further evaluation with ST-elevation AMI. ECG-based ACI-TIPI
when appropriate. and TPI predictive instruments, with a linked
Retrospective review of a sample of patients sent TIPI-IS, provide real-time, concurrent, and retro-
home from the ED with moderate- to high-risk spective decision support tools and feedback for
ACI-TIPI probabilities was conducted to identify the prevention of medical errors in the care of
cases of missed diagnosis of ACI. This review, patients who present with ACS. In real-time, ACI-
conducted at each site by the ED director, identied TIPI probabilities printed on the ECG header for
a series of issues ranging from incomplete docu- the ED physician, provide an additional piece of
mentation to variations in practice, sometimes information for triage decision making, and the
within, though sometimes not within, standard ACI-TIPI risk management form reduces liability
expectations of that ED. Results were addressed risk by prompting consideration and documenta-
with sta during monthly morbidity and mortality tion of key clinical factors in the diagnosis of ACI.
review, in individual feedback discussions, and in Also in real-time, the TPI increases overall coro-
some cases, by changes in the EDs practices. nary reperfusion therapy use. Concurrent agging
by TIPI-IS uses electronically acquired ECG and
Summary hospital data to provide concurrent alerts about
potential misdiagnosis or mis-triage of patients
Medical errors in the care of patients who
with ACS. Retrospectively TIPI-ISbased feed-
present with ACS include errors in ED triage, such
back reports allow performance improvement.
as the decision to send home a patient from the ED
These examples of information technology tools,
integrated into ECG equipment already used in
hospitals to deliver patient care, demonstrate the
potential to adapt other existing equipment or
other patient care activities to enhance patient
safety and error reduction.

References
[1] Pozen MW, DAgostino RB, Selker HP, et al. A pre-
dictive instrument to improve coronary unit admis-
sion practices in acute ischemic heart disease.
N Engl J Med 1984;310:12738.
[2] Selker HP, Grith JL, DAgostino RB. A tool for
judging coronary care unit admission that is appro-
priate for both real-time and retrospective use:
a time-insensitive predictive instrument (TIPI) for
acute cardiac ischemia: a multicenter study. Med
Care 1991;29:61027.
[3] Selker HP, Grith JL, DAgostino RB. A time-
insensitive predictive instrument for acute myocar-
dial infarction mortality: a multicenter study. Med
Care 1991;29:1196211.
[4] Selker HP, Grith JL, DAgostino RB. A time-
insensitive predictive instrument for acute mortality
due to congestive heart failure: development, testing,
and use for comparing hospitals: a multicenter
study. Med Care 1994;32:104052.
[5] Selker HP, Grith JG, Beshansky JR, et al. Patient-
specic predictions of outcomes in myocardial in-
farction for real-time emergency use: a thrombolytic
Fig. 6. TIPI-IS alert message. (Courtesy of Clinical Care predictive instrument. Ann Intern Med 1997;127:
Systems, Inc., Bedford, MA.) 53848.
612 DAUDELIN & SELKER

[6] Pope JH, Aufderheide TA, Ruthazer R, et al. A mul- a prospective multicenter study. Circulation 1996;
ticenter prospective study of missed diagnoses of 94:938.
acute myocardial infarction and unstable angina [19] Maynard C, Beshansky JR, Grith JL, et al. Causes
pectoris in the emergency department. N Engl J of chest pain and symptoms suggestive of acute car-
Med 2000;342:116370. diac ischemia in African American patients present-
[7] Gruppo Italiano per lo Studio Della Streptochinasi ing to the emergency department: a multicenter
NellInfarcto Miocardico (GISSI). Eectiveness of study. J Natl Med Assoc 1997;89:66571.
intravenous thrombolytic therapy in acute myocar- [20] Picken HA, Zucker DR, Grith JL, et al. Insurance
dial infarction. Lancet 1986;1:397401. type and the transportation to emergency depart-
[8] Weaver WB, Eisenberg MS, Martin JS, et al. Myo- ments of patients with acute cardiac ischemia: the
cardial infarction triage and intervention project ACI-TIPI insurance study. Am J Man Care 1998;
phase I: patient characteristics and feasibility of pre- 4:8217.
hospital initiation of thrombolytic therapy. J Am [21] Feldman JA, Fish SS, Beshansky JR, et al. Acute
Coll Cardiol 1990;15:92531. cardiac ischemia in patients with cocaine-associated
[9] Cairns CB, Niemann JT, Selker HP, et al. A comput- complaints: results of a multicenter trial. Ann Emerg
erized version of the time-insensitive predictive in- Med 2000;36:46976.
strument: use of the Q wave, ST segment, T wave [22] Udelson JE, Beshansky JR, Ballin DS, et al. Emer-
and patient history in the diagnosis of acute myocar- gency department assessment of Tc-99m sestamibi
dial infarction by the computerized ECG. J Electro- imaging for evaluation and triage of patients with
cardiol 1991;24:S469. suspected acute cardiac ischemia: a randomized,
[10] Lobodzinski SM, Laks MM. Present and future con- multicenter, controlled clinical trial. JAMA 2002;
cepts in computerized electrocardiography. Intern 288:2693700.
Med Specialist 1987;8:15292. [23] Pope JH, Ruthazer R, Kontos MC, et al. The impact
[11] Selker HP, Beshansky JR, Grith JL, et al. Use of of electrocardiographic left ventricular hypertrophy
the acute cardiac ischemia time-insensitive predictive and bundle branch block on the triage and outcome
instrument (ACI-TIPI) to assist emergency depart- of emergency department patients with a suspected
ment triage of patients with chest pain or other acute coronary syndrome: a multicenter study. Am
symptoms suggestive of acute cardiac ischemia: J Emerg Med 2004;22(3):15663.
a multicenter controlled clinical trial. Ann Intern [24] Coronado BE, Pope JH, Grith JL, et al. Clinical
Med 1998;129:84555. features, triage and outcome of patients presenting
[12] Selker HP, Grith JL, Dorey FJ, et al. How do to the emergency department with a suspected acute
physicians adapt when the coronary care unit is coronary syndrome but without pain: a multicenter
full? A prospective multicenter clinical study. study. Am J Emerg Med 2004;22(7):56874.
JAMA 1987;257:11815. [25] McCarthy BD, Beshansky JR, DAgostino RB, et al.
[13] McCarthy BD, Wong JB, Selker HP. Detecting Missed diagnoses of acute myocardial infarction in
acute cardiac ischemia in the emergency department: the emergency department: results from a multicen-
a review of the literature. J Gen Intern Med 1990;5: ter study. Ann Emerg Med 1992;22:57982.
36573. [26] Jayes RL, Larsen GC, Beshansky JR, et al. Physi-
[14] Fleming C, DAgostino RB, Selker HP. Is coronary cian electrocardiogram reading in the emergency de-
care unit admission restricted for elderly patients? A partmentdaccuracy and eect on triage decisions:
multicenter study. Am J Public Health 1991;81: ndings from a multicenter study. J Gen Intern
11216. Med 1992;7:38792.
[15] Jayes RL, Beshansky JR, DAgostino RB, et al. Do [27] Zalenski RJ, Rydman RJ, Ting S, et al. A national
patients coronary risk factors predict acute cardiac survey of emergency department chest pain centers
ischemia in the emergency room? J Clin Epidemiol in the United States. Am J Cardiol 1998;81:13059.
1992;45:6216. [28] Zalenski RJ, Selker HP, Cannon CP, et al. National
[16] Georgeson S, Linzer M, Grith JL, et al. Acute car- Heart Attack Alert Program position paper: chest
diac ischemia in patients with syncope: importance pain centers and programs for the evaluation of
of the initial electrocardiogram. J Gen Intern Med acute cardiac ischemia. Ann Emerg Med 2000;35:
1992;7:37986. 46271.
[17] Larsen GC, Grith JL, Beshansky JR, et al. Electro- [29] Selker HP, Beshansky JR, Pozen JD, et al. Electro-
cardiographic left ventricular hypertrophy in cardiograph-based emergency department risk man-
patients with suspected acute cardiac ischemiadits agement tool based on the ACI-TIPI; potential
inuence on diagnosis, triage and short-term prog- impact on care and malpractice claims. J Healthc
nosis: a multicenter study. J Gen Intern Med 1994; Risk Manag 2002;118.
9:66673. [30] Fibrinolytic Therapy Trialists (FTT) Collaborative
[18] Maynard C, Beshansky JR, Grith JL, et al. The in- Group. Indications for brinolytic therapy in sus-
uence of sex on the use of cardiac procedures in pected acute myocardial infarction: collaborative
patients presenting to the emergency department: overview of early mortality and major morbidity
ACS PREVENTION IN EMERGENCY DEPARTMENTS 613

results from all randomized trials of more than 1000 American College of Cardiology/American Heart
patients. Lancet 1994;343:31122. Association Task Force on Practice Guidelines
[31] ISIS-2 (Second International Study of Infarct Sur- (Committee on Management of Acute Myocardial
vival) Collaborative Group. Randomized trial of in- Infarction). Circulation 1999;100:101630.
travenous streptokinase, oral aspirin, both, or [42] Rogers WJ, Canto JG, Lambrew CT, et al, for the
neither among 17,187 cases of suspected acute myo- Investigators in the National Registry of Myocardial
cardial infarction: ISIS-2. Lancet 1988;2:34960. Infarction 1, 2, and 3. Temporal trends in the treat-
[32] Cannon CP, Antman EM, Walls R, et al. Time as an ment of over 1.5 million patients with myocardial in-
adjunctive agent to thrombolytic therapy. J Thromb farction in the US from 1990 through 1999. J Am
Thrombolysis 1994;1:2734. Coll Cardiol 2000;36:205663.
[33] National Heart Attack Alert Program Coordinating [43] Berger AK, Radford MJ, Krumholtz HM. Factors
Committee 60 Minutes to Treatment Working associated with delay in reperfusion therapy in el-
Group. Emergency department: rapid identication derly patients with acute myocardial infarction:
and treatment of patients with acute myocardial in- analysis of the cooperative cardiovascular project.
farction. Ann Emerg Med 1994;23:31129. Am Heart J 2000;139:98592.
[34] National Heart Attack Alert Program Coordinating [44] Barron HV, Bowlby LJ, Breen T, et al, for the Na-
Committee Access to Care Subcommittee. Emer- tional Registry of Myocardial Infarction 2 Inves-
gency medical dispatching: rapid identication and tigators. Use of reperfusion therapy for acute
treatment of acute myocardial infarction. Am J myocardial infarction in the United States. Data
Emerg Med 1995;13:6773. from the National Registry of Myocardial Infarc-
[35] National Heart Attack Alert Program Coordinating tion 2. Circulation 1998;97:11506.
Committee Access to Care Subcommittee. 91-1: [45] Selker HP, Grith JL, Beshansky JR, et al. The
rapid identication and treatment of acute myocar- thrombolytic predictive instrument project: combin-
dial infarction. Am J Emerg Med 1995;13:18895. ing clinical study databases to take medical eec-
[36] Rogers WJ, Bowlby LJ, Chandra NC, et al. Treat- tiveness research to the streets. Washington DC:
ment of myocardial infarction in the United States AHCPR, DHHS; 1992.
(1990 to 1993) observations from the national regis- [46] Kennedy JW, Martin GV, Davis KB, et al. The
try of myocardial infarction. Circulation 1994;90: Western Washington intravenous streptokinase in
210314. acute myocardial infarction randomized trial. Circu-
[37] National Heart Attack Alert Program Coordinating lation 1988;77:34552.
Committee Access to Care Subcommittee. Access to [47] Proceedings of symposium on rapid identication
timely and optimal care of patients with acute coro- and treatment of acute myocardial infarction. Na-
nary syndromescommunity planning considera- tional Heart, Lung and Blood Institute, DHHS,
tions: a report by the National Heart Attack Alert USA, 1992.
Program. J Thromb Thrombolysis 1998;6:1936. [48] Weaver WB, Eisenberg MS, Martin JS, et al. Myo-
[38] National Heart Attack Alert Program Coordinating cardial infarction triage and intervention project
Committee Working Group on Educational Strate- phase I: Patient characteristics and feasibility of pre-
gies to Prevent Prehospital Delay in Patients at hospital initiation of thrombolytic therapy. J Am
High Risk for Acute Myocardial Infarction. Educa- Coll Cardiol 1990;15:92531.
tional strategies to prevent prehospital delay in [49] Pozen MW, DAgostino RB, Mitchell JB, et al. The
patients at high risk for acute myocardial infarction: usefulness of a predictive instrument to reduce inap-
a report by the National Heart Attack Alert Pro- propriate admissions to the coronary care unit. Ann
gram. J Thromb Thrombolysis 1998;6:4761. Intern Med 1980;92:23842.
[39] Sims RJ, Topol EJ, Holmes DR, et al. Link between [50] Rosati RA, McNeer JF, Starmer CF, et al. A new in-
the angiographic substudy and mortality outcomes formation system for medical practice. Arch Intern
in a large randomized trial of myocardial reperfu- Med 1975;135:101724.
sion: importance of early and complete infarct artery [51] Pryor DP, Cali RM, Harrell FE, et al. Clinical
reperfusion. Circulation 1995;91:19238. databases: Accomplishments and unrealized poten-
[40] Ryan TJ, Anderson JL, Antman EM, et al. ACC/ tial. Med Care 1985;23:62347.
AHA guidelines for the management of patients [52] Topol EJ, Cali RM, George BS, et al. A random-
with acute myocardial infarction: a report for the ized trial of immediate versus delayed elective angio-
American College of Cardiology/American Heart plasty after intravenous tissue plasminogen
Association Task Force on Practice Guidelines activator in acute myocardial infarction. N Engl J
(Committee on Management of Acute Myocardial Med 1987;317:5818.
Infarction). J Am Coll Cardiol 1996;28:1328428. [53] Topol EJ, Cali RM, George BS, et al. Coronary
[41] Ryan TJ, Antman EM, Brooks NH, et al. 1999 up- arterial thrombolysis with combined infusion of
date: ACC/AHA guidelines for the management of recombinant tissue-type plasminogen activator and
patients with acute myocardial infarction: executive urokinase in patients with acute myocardial infarc-
summary and recommendations: a report of the tion. Circulation 1988;77:11007.
614 DAUDELIN & SELKER

[54] Topol EJ, George BS, Kereiakes DJ, et al. A ran- activator compared with front-loaded alteplase in
domized controlled trial of intravenous tissue plas- acute myocardial infarction. Results of the TIMI
minogen activator and early intravenous heparin in 10B trial. Circulation 1998;98:280514.
acute myocardial infarction. Circulation 1989;79: [68] Van de Werf F, Cannon CP, Luyten A, et al. Safety
2816. assessment of single bolus administration of TNK-
[55] Wall TC, Strickland J, Masoud J, et al. Thrombolytic tPA in acute myocardial infarction. The ASSENT-
therapy on the homefront: intravenous urokinase in 1 trial. Am Heart J 1999;137:78691.
community hospitals. N C Med J 1989;50:3636. [69] Fong WC, Fox NL, Cannon CP, et al. Determina-
[56] Cali RM, Topol EJ, Stack RS, et al. Evaluation of tion of a weight-adjusted dose of TNK-tissue plas-
combination thrombolytic therapy and timing of minogen activator. Am Heart J 2001;141:3340.
cardiac catheterization in acute myocardial infarc- [70] Aufderheide TP, Hendley GE, Thakur RK, et al.
tion: results of the thrombolysis and angioplasty in The diagnostic impact of prehospital 12-lead electro-
myocardial infarction-phase 5 randomized trial. Cir- cardiography. Ann Emerg Med 1990;19:12807.
culation 1991;83:154356. [71] Karagounis L, Ipsen SK, Jessop MR, et al. Impact
[57] Althouse R, Maynard C, Cerqueira MD, et al. The of eld-transmitted electrocardiography on time to
Western Washington myocardial infarction registry in-hospital thrombolytic therapy in acute myocar-
and emergency department tissue plasminogen acti- dial infarction. Am J Cardiol 1990;66:78691.
vator treatment trial. Am J Cardiol 1990;66: [72] Canto JG, Rogers WJ, Bowlby LJ, et al. The preho-
1298303. spital electrocardiogram in acute myocardial infarc-
[58] Schmid CH, DAgostino RB, Grith JL, et al. A lo- tion: is its full potential being realized? J Am Coll
gistic regression model when some events precede Cardiol 1997;29:498505.
treatment: the eect of thrombolytic therapy for [73] Kereiakes DJ, Gibler WB, Martin LH, the Cincin-
acute myocardial infarction on the risk of cardiac ar- nati Heart Project Study Group, et al. Relative im-
rest. J Clin Epidemiol 1997;50:121929. portance of emergency medical system transport
[59] Gillum RF, Fortmann SP, Prineas RJ, et al. Interna- and the prehospital electrocardiogram on reducing
tional diagnostic criteria for acute myocardial infarc- hospital time delay to therapy for acute myocardial
tion and acute stroke. Am Heart J 1984;108:1508. infarction. A preliminary report from the Cincinnati
[60] Zucker DR, Grith JL, Beshansky JR, et al. Presen- Heart Project. Am Heart J 1992;123:83540.
tations of acute myocardial infarctions in men and [74] Kudenchuk PJ, Maynard C, Cobb LA, et al. Utility
women: results from a prospective, multicenter of the prehospital electrocardiogram in diagnosing
study. J Gen Intern Med 1997;12:7987. acute coronary syndromes: the Myocardial Infarc-
[61] Coronado BE, Grith JL, Beshansky JR, et al. Hos- tion Triage and Intervention (MITI) project. J Am
pital mortality in women and men with acute cardiac Coll Cardiol 1998;32:1727.
ischemia: a prospective multicenter study. J Am Coll [75] Cannon CP, Sayah AJ, Walls RM. Prehospital
Cardiol 1997;29:14906. thrombolysis: an idea whose time has come. Clin
[62] Lange RA, Hillis LD. Should thrombolysis or pri- Cardiol 1999;22:109.
mary angioplasty be the treatment of choice for [76] Amit G, Weiss T, Zahger D. Coronary angioplasty
acute myocardial infarction? N Engl J Med 1996; or intravenous thrombolysis: the dilemma of opti-
335:13117. mal reperfusion in acute myocardial infarction:
[63] Weaver WD, Cerqueira M, Hailstrom AP, et al. Pre- a critical review of the literature. J Thromb Throm-
hospital-initiated vs. hospital initiated thrombolytic bolysis 1999;8:11323.
therapy the myocardial infarction triage and inter- [77] Stern R, Arntz HR. Prehospital thrombolysis in
vention trial. JAMA 1993;270:12116. acute myocardial infarction. Eur J Emerg Med
[64] Selker HP, Zalenski RJ, Antman EM, et al. An eval- 1998;5:4719.
uation of technologies for identifying acute cardiac [78] Widimsky P, Groch L, Zelizko M, et al, on behalf of
ischemia in the emergency department: a report the PRAGUE Study Group Investigators. Multi-
from a national heart attack alert working group. centre randomized trial comparing transport to pri-
Ann Emerg Med 1997;29:187. mary angioplasty vs. immediate thrombolysis vs.
[65] Rawles J. Halving of mortality at 1 year by domicil- combined strategy for patients with acute myocar-
iary thrombolysis in the Grampian Region Early dial infarction presenting to a community hospital
Anistreplase Trial (GREAT). J Am Coll Cardiol without a catheterization laboratory. The PRAGUE
1994;23:15. study. Eur Heart J 2000;21:82331.
[66] The European Myocardial Infarction Project [79] Daudelin D, Selker H, Kwong M, et al. Information
Group. Prehospital thrombolytic therapy in patients technology to reduce errors in emergency cardiac
with suspected acute myocardial infarction. N Engl J care. In: Henriksen K, Battles JB, Marks E, et al, edi-
Med 1993;329:3839. tors. Advances in patient safety: from research to
[67] Cannon CP, Gibson CM, McCabe CH, et al, for implementation. Vol. 3. Implementation issues.
the Thrombolysis in Myocardial Information Rockville, MD: Agency for Healthcare Research
(TIMI) 10B Investigators. TNK-tissue plasminogen and Quality; 2005 in press.

You might also like