Relationship of Atrial Fibrillation and Stroke After
Coronary Artery Bypass Graft Surgery: When is
Anticoagulation Indicated?
Andras Kollar, MD, PhD, Scott D. Lick, MD, Kathleen N. Vasquez, PA, and
Vincent R. Conti, MD
Division of Cardiothoracic Surgery, Department of Surgery, University of Texas Medical Branch, Galveston, Texas
Background. Atrial fibrillation (AF) is considered as a
risk factor for stroke after coronary artery bypass grafting
operations.
Methods. A retrospective search in our hospitals med-
ical record database was done to identify patients with
postoperative strokes who underwent coronary artery
bypass grafting operations from January 1, 1993, until
December 31, 2004. All cases were individually reviewed,
and the temporal relationship between neurologic event
and postoperative episodes of AF was determined. Dur-
ing the study period it was our consistent policy to use
only Coumadin anticoagulation limited to patients who
had persistent AF or were to be discharged in AF.
Results. Of the 2,964 coronary artery bypass grafting
operations, 576 patients (19.4%) had AF and 32 patients
(1.1%) suffered stroke. Seventeen stroke patients main-
tained normal sinus rhythm during their hospital stay.
Of the remaining 15 patients, 9 presented with neuro-
logic deficit before the first episode of AF, with 5 having
P ostoperative atrial fibrillation (AF) is the single most
common cardiac complication, occurring in 16% to
35% [110] of patients after coronary artery bypass graft-
ing (CABG) surgery, and is generally believed to repre-
sent an increased hazard for stroke [1 4, 7, 9]. In medical
patients with chronic or recurrent AF, the cause and
effect relationship between the arrhythmia and the cere-
brovascular event has been unquestionably proven [11
14], and long-term anticoagulation with sodium warfarin
(Coumadin) has emerged as the preferred method to
prevent thrombus formation and embolic strokes in
higher risk patients [14]. Postoperative AF, however, has
a self-limiting nature [8, 14], and a similar clear-cut
relationship between the above two entities has not been
confirmed.
Although many studies have confirmed that chemical
prophylaxis successfully reduces the incidence of post-
operative AF [10], there is no report whether this reduc-
tion had any effect on the stroke incidence. The American
College of Cardiology/American Heart Association
Accepted for publication March 14, 2006.
Address correspondence to Dr Kollar, Department of Surgery, University
of Texas Medical Branch, 301 University Blvd, Galveston, TX 77555;
e-mail: ankollar@utmb.edu.
2006 by The Society of Thoracic Surgeons
Published by Elsevier Inc
CARDIOVASCULAR
intraoperative and 4 having postoperative stroke. Of the
6 patients with AF before neurologic event, three strokes
occurred within 1 week after spontaneous conversion to
normal sinus rhythm. One patient with preoperative and
also with intraoperative AF who underwent emergency
coronary artery bypass grafting woke up with stroke. In
the remaining two cases, the AF or atrial flutter episodes
lasted less than 6 hours each before the neurologic event.
More aggressive anticoagulation as suggested in the
published guidelines could not have prevented strokes
in any of these 6 patients.
Conclusions. This retrospective analysis does not sup-
port the use of aggressive anticoagulation, particularly
full intravenous heparinization as a bridging therapy to
decrease the already low incidence of postoperative
strokes after routine coronary artery bypass grafting
surgery.
(Ann Thorac Surg 2006;82:51523)
2006 by The Society of Thoracic Surgeons
(ACC/AHA) guidelines [9, 14] strongly recommend early
aggressive anticoagulation to prevent thromboembolic
strokes, particularly when cardioversion is attempted.
According to one of the largest prospective studies on
postoperative AF, involving 70 institutions in 17 coun-
tries, 56.2% of patients in AF received intravenous hep-
arin while only 17.6% of patients were started on Cou-
madin [8]; however, there are no data to suggest that such
a practice could prevent postoperative strokes in any
significant number.
In our institution, we do not use aggressive antico-
agulation for postoperative AF, and we rarely use
electrical cardioversion except for intubated patients
on the intensive care unit and for those with significant
hemodynamic compromise. Full heparinization is re-
served for patients with pulmonary embolism or doc-
umented deep venous thrombosis, and Coumadin
therapy is introduced gradually when the patients
have had AF preoperatively or remain in new, contin-
uous AF for more than 48 hours and they are expected
to be discharged in AF. The aim of this study was to
examine the temporal relationship between postoper-
ative AF and cerebrovascular accidents after CABG
operations and to determine whether any of the
0003-4975/06/$32.00
doi:10.1016/j.athoracsur.2006.03.037
 516 KOLLAR ET AL
ATRIAL FIBRILLATION AND STROKE AFTER CABG
strokes could have been prevented by a more aggres-
sive anticoagulation protocol.
Material and Methods
CARDIOVASCULAR
After appropriate institutional review board approval,
which included waiver of the requirement to obtain prior
consent of the individuals affected (September 24, 2004),
a retrospective search in our university hospitals medi-
cal record database was done on all CABG patients
operated on from January 1993 until December 2004.
Patients with an additional discharge ICD code of 433.00-
01, 433.20-21, 434.00-01, 434.90-91, 433.10-11, 433.30-31,
434.10-11, and 436.0, indicating neurologic events, were
identified and their hospital chart was reviewed. Simi-
larly, all hospital readmissions within 30 days of dis-
charge after CABG surgery were screened for delayed
strokes. In addition, our departments computerized So-
ciety of Thoracic Surgeons database was also searched
for perioperative strokes or transient ischemic attacks
(TIAs), and the list was compared with the hospital
records.
Preoperative Studies and Stroke Prevention in
Coronary Artery Bypass Graft Patients
In our program, during the study period carotid duplex
screening was done on patients with carotid bruits and
when there were symptoms or history of cerebrovascular
accidents or TIA. Symptomatic patients were also evalu-
ated by head computed tomography and vascular sur-
gery if more than 50% internal carotid artery stenosis was
noted on carotid duplex. In case of concomitant disease
and high-grade (80%) internal carotid artery stenosis,
we preferred staged procedures: carotid endarterectomy
first, followed by CABG. After CABG surgery, all patients
were started on 325 mg of aspirin 8 hours after arrival to
the surgical intensive care unit unless excessive postop-
erative bleeding was recorded by means of the chest
tubes.
Surgical Technique
During the study period, five surgeons performed all
CABG operations using essentially the same surgical
technique. In our institution, intraoperative transesoph-
ageal echocardiography (TEE) has become a routine
procedure during the second half of the study period,
and epiaortic echocardiography is used only when as-
cending aortic disease was suspected by palpation and
TEE was not in use. However, we adopted the routine use
of single aortic cross-clamping to minimize manipula-
tions on the ascending aorta. The operations were done
with cardiopulmonary bypass using moderate hypother-
mia and intermittent cold potassium blood cardioplegia
except for a few patients in the second half of the study
period who had operations without cardiopulmonary
bypass. The right atrium was cannulated with a two-
stage venous cannula, and during aortic cross-clamping
the ascending aorta was vented by gravity. Left atrial or
ventricular venting was not routinely used. Proximal
anastomoses were constructed first, followed by the dis-
Ann Thorac Surg
2006;82:51523
tal (side-to-side or end-to-side) vein and internal thoracic
artery anastomoses. All patients received two right ven-
tricular and two right atrial temporary pacing wires, and
atrial or atrioventricular sequential pacing was instituted
as deemed appropriate.
Postoperative Care and Arrhythmia Detection
Most patients were transferred from the intensive care
unit to a surgical telemetry unit on the first postoperative
day where surface electrocardiogram and atrial electro-
grams are monitored simultaneously until discharge.
Any arrhythmia episodes are automatically printed and
recorded in the chart by the nursing staff. Our routine
protocol consists of oral -blocker prophylaxis unless
contraindicated. Episodes of AF are treated for rate
control with combined digoxin and -blocker or calcium-
channel blocker if -blocker is not tolerated. Amioda-
rone, quinidine, or procainamide are only used occasion-
ally. Electrical cardioversion is routinely used in patients
intubated on the intensive care unit and in those with
significant hemodynamic compromise. Atrial flutter is
typically converted back to sinus rhythm or to AF with
rapid atrial pacing. The pacing wires are removed 2 or
more hours before discharge.
Our practice is to institute oral anticoagulation in only
those patients who have persistent AF more than 48
hours and were to be discharged in AF. Full hepariniza-
tion was not used because of AF, although some received
venous thromboprophylaxis levels of heparin (5,000 U
subcutaneously every 8 to 12 hours). Therapeutic pro-
thrombin and international normalized ratio levels (in-
ternational normalized ratio, 2.0 to 3.0) was frequently
achieved only after discharge. These policies were part of
a written protocol for perioperative care that was agreed
on and used by each of the surgeons with cases during
the study period.
Results
During the 12-year study period, 2,964 isolated CABG
operations were performed in our program. Thirty-two
patients (1.1%) suffered stroke or TIA, and 576 patients
(19.4%) developed AF or atrial flutter during their hospi-
tal stay. Overall mortality for elective primary CABG
cases was 1.4%, and 3 patients with stroke died (9.3%
mortality). Our current (2001 to 2004) median postoper-
ative length of stay for CABG patients without AF or
atrial flutter (n 485) is 5.0 days (range, 3 to 141 days)
and for patients with postoperative AF or atrial flutter (n
109) it is 6.0 days (range, 4 to 67 days).
The above 32 patients with stroke or TIA represent our
current database. There were 19 men and 13 women with
a mean age of 61 years (range, 45 to 85 years). History and
clinical data are presented in Table 1, and perioperative
surgical data are presented in Table 2. Head computed
tomography verification was done in all but one case
(TIA only). Twenty-five patients had radiographic evi-
dence of fresh ischemic infarct. Twelve patients woke up
with neurologic deficit and were defined as having suf-
fered an intraoperative stroke. The other 20 patients were
Ann Thorac Surg KOLLAR ET AL 517
2006;82:51523 ATRIAL FIBRILLATION AND STROKE AFTER CABG

Table 1. Preoperative History and Clinical Data occurred 12 hours to 3 days after the stroke. Of the
remaining 5 patients, one (no. 7) developed intraopera-
Variable n
tive AF, which was electrically terminated, then atrially
Medical history paced for 24 hours. On postoperative day 3 he had
Hypertension 24 another episode of AF that converted to sinus rhythm

CARDIOVASCULAR
Diabetes mellitus 23 within 6 hours, but he suffered a stroke the same evening.
Myocardial infarction 16 Transesophageal echocardiography showed a small atrial
Recent MI 2 septal defect with bidirectional shunt suggesting para-
Congestive heart failure 4 doxical embolism. One patient (no. 22) developed atrial
Chronic atrial fibrillation 1 flutter on day 2 (successfully rapid paced into sinus
Acute atrial fibrillation (post MI) 1 rhythm) and had a first recurrence of flutter with similar
COPD 3
conversion again on postoperative day 8 and a stroke the
same day. One patient (no. 24) had four short episodes
Smoking 15
(30 minutes to 6 hours) of AF and was discharged in sinus
History of stroke/TIA 13
rhythm on postoperative day 7. He was readmitted in
Carotid bruit 5
normal sinus rhythm with symptoms of stroke and short-
Preoperative medications
ness of breath and was confirmed to have a pulmonary
Aspirin 22
embolism and a patent foramen ovale with a right-to-left
Coumadin 0 shunt. One patient (no. 25) had three short episodes (30
-blocker 26 minutes to 4 hours) of AF and was discharged home in
Ca antagonist 9 sinus rhythm on postoperative day 6. This patient had
ACE inhibitor 14 documented normal sinus rhythm on return clinic visits
Digoxin 5 (postoperative day 17 and postoperative day 31) and also
Diuretic 10 at readmission to the hospital with a massive stroke and
right internal carotid artery occlusion (postoperative day
ACE angiotensin-converting enzyme; COPD chronic obstructive
pulmonary disease; MI myocardial infarction; TIA transient 35 after CABG). The last patient (no. 27) again had three
ischemic attack. short episodes (30 minutes to 4 hours) of AF until
postoperative day 4. This patients postoperative recov-
ery was slower than usual, but he had no neurologic
neurologically intact after extubation and presented with deficit until he had a sudden respiratory arrest on post-
sudden hemiparesis, focal deficit, or signs of encephalop- operative day 13. His cardiac recovery was complete (no
athy between postoperative days 2 and 35. Four patients more AF episodes), but he suffered bilateral anoxic brain
with focal neurologic deficit and no evidence of infarct damage and subsequently died.
recovered completely (TIA), and all other surviving pa- In our series no patient with neurologic event after
tients were discharged in improved neurologic condition
after a median postoperative length of stay of 14 days
(range, 6 to 90 days). Table 2. Intraoperative and Perioperative Data
From the available clinical data and study reports, each
Variable n
case was retrospectively reconstructed, and the likely
source or cause of stroke was determined (Tables 3, 4). Of Elective CABG 28
the 12 patients with intraoperative strokes, 6 remained in Emergency CABG 4
normal sinus rhythm and 5 patients had subsequent AF Number of grafts (n/patient) 3.4 (26)
without additional neurologic sequel. One patient required Cardiopulmonary bypass time (min) 128.6 (70216)
emergency CABG for acute myocardial infarction and had Aortic cross-clamp time (min) 91.6 (46166)
preoperative AF after temporizing percutaneous translumi- Systemic hypothermia
nal coronary angioplasty for which he was fully anticoagu- Moderate (2832C) 29
lated with heparin. He suffered an intraoperative stroke, Deep (1820C) 3
and a TEE on postoperative day 3 confirmed left atrial Intraoperative arrhythmia (AF) 2
appendage thrombus (intraoperative TEE was not available Intraoperative antiarrhythmic drug 0
at that time). Post-bypass pacemaker
Of the 20 patients with postoperative strokes, 11 had no
Atrial pacing 16
episodes of AF or atrial flutter (Table 3). Suspected causes
AV sequential pacing 5
were vertebral artery disease (5 patients), carotid artery
Normal sinus rhythm 11
disease (2 patients), aortic arch atheromas (2 patients),
Inotropic support 24 hours 4
and unexplained (2 patients).
Ventilator support 24 hours 5
The remaining 9 patients, then, are of greatest interest
for this analysis: they had postoperative strokes or TIA Reoperation for bleeding 1
(days 2 to 8) and AF or atrial flutter (Table 4). In 4 patients Perioperative MI 1
(2 had strokes on postoperative day 2 and 2 on postop- AF atrial fibrillation; AV atrioventricular; CABG coronary
erative day 3) the first episodes of atrial fibrillation artery bypass grafting; MI myocardial infarction
 CARDIOVASCULAR

518
ATRIAL FIBRILLATION AND STROKE AFTER CABG
KOLLAR ET AL
Table 3. Detailed Clinical Data on Stroke Patients Without Postoperative Atrial Fibrillation
Case Carotid HX of CVA/ Postoperative Clinical Postoperative
No. Cardiac Status Bruit TIA Carotid Duplex Stroke Picture Likely Source AF Occurence Comment

2 Stable No Recent TIA Bilateral 40% Intraoperative Mental change Unexplained No ...
4 Unstable IABP No No Intraoperative Hemiparesis Aortic atherosclerosis No ...
5 Stable No No POD 4 Focal TIA Vertebral artery No ...
athersclerosis
9 Stable No No POD 3 Hemiparesis L carotid disease No ...
13 Stable Recent MI No No Intraoperative Hemiparesis LV thrombus No POD 4 autopsy
15 Stable Yes Recent TIAs Bilateral 15% POD 3 Mental change Unexplained No Patient refused
TEE study
16 Stable No No POD 2 Hemiparesis Aortic atherosclerosis No ...
18 Stable Yes No Bilateral 50%70% Intraoperative Mental change Carotid disease No ...
19 Stable Yes Recent TIA L, 65% R, 15% POD 3 Mental change Carotid disease No ...
20 Stable No No ... Intraoperative Mental change Unexplained No ...
21 Stable No No ... POD 4 Focal TIA Unexplained No ...
23 Stable No Remote ... Intraoperative Hemiparesis Unexplained No ...
26 Stable No No ... POD 12 Hemiparesis Vertebral artery No Readmission with
occlusion stroke
28 Stable No Remote Bilateral 16%49% POD 4 Focal TIA Aortic atherosclerosis No ...
29 Stable No No ... POD 18 Hemiparesis Unexplained No Readmission with
stroke
30 Stable No No ... POD 31 Focal TIA Vertebral artery No Readmission with
occlusion stroke
31 Stable No No ... POD 8 Focal TIA Unexplained No Readmission with
TIA and SOB
(confirmed PE)

AF atrial fibrillation; CVA cerebrovascular accident; HX history; IABP intra aortic balloon pump; L left; LV left ventricular; MI myocardial infarction; PE
pulmonary embolism; POD postoperative day; R right; SOB shortness of breath; TEE transesophageal echocardiography; TIA transient ischemic attack.

Ann Thorac Surg

2006;82:51523
 2006;82:51523
Ann Thorac Surg
Table 4. Detailed Clinical Data on Stroke Patients With Postoperative Atrial Fibrillation
Case Cardiac Carotid HX of CVA/ Postoperative Postoperative AF Before
No. Status Bruit TIA Carotid Duplex Stroke Clinical Picture Likely Source AF Occurence Stroke Comment on AF

1 Stable No No ... POD 2 Hemiparesis Unexplained POD 2 No 6 h after stroke

3 Unstable No Recent TIA L, 100% R, normal POD 2 Hemiparesis Carotid disease POD 4 No 2 days after stroke
6 Acute MI No No ... Intraoperative Hemiparesis LAA thrombus POD 0 Yes 4 h postoperative
Preop. AF
7 Stable No Remote TIA Bilateral 15% POD 3 Hemiparesis Atrial septal Intraoperative, Yes Paradoxical
aneurysm/shunt then POD 3 embolism?
8 Stable No No ... Intraoperative Mental change Unexplained POD 2 No 2 days after stroke
10 Stable Yes No R vertebral, 80% POD 3 Mental change Vertebral artery POD 4 No 1 day after stroke
90% disease
11 Stable No No Intraoperative Mental change Aortic atherosclerosis POD 4 No 4 days after stroke
12 Stable No Remote CVA Bilateral 15% POD 3 Mental change Unexplained POD 6 No 3 days after stroke
14 Stable Yes Recent CVA L, 70% R, 100% Intraoperative Mental change Carotid disease POD 1 No ...
17 Stable No Recent CVA L, 15% R, 50% Intraoperative Hemiparesis Unexplained POD 3 No 3 days after stroke
22 Stable No Recent TIA ... POD 8 Hemiparesis Unexplained POD 2 Yes Atrial flutter
episodes

ATRIAL FIBRILLATION AND STROKE AFTER CABG

24 Stable No No ... POD 20 Hemiparesis Atrial R-L shunt POD 2 Yes Readmission with
Paradoxical stroke/NSR
embolism? (confirmed PE)
25 Stable No No ... POD 35 Hemiparesis R. carotid artery POD 2 Yes Readmission with
occlusion stroke/NSR
27 Stable No Remote Bilateral 16%49% POD 13 Bilateral diffuse Asphyxia, then POD 2 Yes Patient never
brain injury cardiac arrest woke up after
resuscitation
32 Unstable No No ... Intraoperative Bilateral diffuse Perioperative POD 2 No Patient never
brain injury prolonged LCO woke up after
and tamponade surgery

AF atrial fibrillation; CVA cerebrovascular accident; HX history; L left; LAA left atrial appendage; LCO low cardiac output; NSR normal sinus rhythm; PE
POD postoperative day; R right; TIA transient ischemic attack.

KOLLAR ET AL
pulmonary embolism;

519
CARDIOVASCULAR
 520 KOLLAR ET AL
ATRIAL FIBRILLATION AND STROKE AFTER CABG
postoperative AF had a long enough AF episode to
warrant aggressive anticoagulation suggested by the
ACC/AHA guidelines.
CARDIOVASCULAR
Comment
New-onset postoperative AF is generally believed to be
self-limited, with 95% of cases developing within 5 to 6
days after the operation [8]. The peak occurrence of AF is
the second postoperative day, after which most patients
spontaneously convert and remain in normal sinus
rhythm [8]. Approximately 40% of patients have more
than one episode typically recurring within 24 hours, but
only a small percentage are discharged with persistent
AF [8].
The pathophysiology of postoperative AF is not com-
pletely understood. Apart from obvious comorbid condi-
tions such as valvular heart disease, atrial enlargement,
congestive heart failure, and history of preoperative atrial
arrhythmias [13, 5], several other risk factors predispose
cardiac surgical patients for postoperative AF. Advanced
age is the strongest, followed by systemic hypertension,
left ventricular hypertrophy, peripheral vascular disease,
and chronic lung disease [1 6, 8]. Longer cardiopulmo-
nary bypass time and aortic cross-clamp time have been
shown to be associated with increased incidence of
postoperative AF [1, 15, 16]. Similarly, various surgical
manipulations on the atria (eg, left atrial venting [2, 3]),
the need for prolonged inotropic or mechanical support,
and the patients own adrenergic status [5, 7, 8] appear to
have an impact, although it is not clear whether these
factors are true independent risk factors or they are just
representative of sicker patients. Postoperative pericar-
dial fluid collection and pericarditis have also been
associated with atrial arrhythmias [17, 18], whereas tech-
nical modifications, such as the preservation of anterior
epicardial fat pad, have been found to decrease the
overall incidence of AF [19].
In nonsurgical patients, the relationship between
chronic AF and thromboembolic stroke has been exten-
sively studied and well established [14]. Chronic AF leads
to decreased blood flow velocities, particularly in the left
atrial appendage, with consequent thrombus formation
and systemic embolization [20, 21]. Because the economic
and health care implications of stroke are enormous,
there have been many efforts to reduce stroke rates in
these patient groups. It is now universally accepted that
long-term anticoagulation with Coumadin significantly
decreases the risk of strokes in higher risk medical
patients [14]. However, studies that risk-stratified these
patients have led to accepted guidelines that have de-
fined lower risk patients with chronic AF in whom the
relative risks do not favor Coumadin anticoagulation
[1113, 2225]. On the basis of these studies, for younger
patients (65 years of age) with no or minimal other risk
factors who have an annual stroke risk of less than 1.5%
per year, aspirin alone is considered sufficient, and the
risk of Coumadin-related bleeding may actually out-
weigh the benefit of therapy [22].
It is not entirely evident whether cardioversion and
Ann Thorac Surg
2006;82:51523
rhythm control is a successful strategy to prevent strokes.
In the AFFIRM study [26] performed in the United States
and in another study [27] from the Netherlands, medical
patients with chronic AF were randomized into rate
control plus anticoagulation versus rhythm control plus
anticoagulation groups. After 2 to 3 years of follow up, a
similar number of strokes occurred in both groups.
Altogether the rhythm control strategy offered no sur-
vival advantage, and the rate control group had advan-
tages including less frequent need for hospitalization and
fewer adverse reactions to antiarrhythmic drugs.
How can these data be applied to our CABG patient
population? The combined ACC/AHA/European Society
of Cardiology practice guidelines (2001) for AF [14] rec-
ommend anticoagulation with heparin or oral anticoag-
ulation in general and continuous heparin anticoagula-
tion when cardioversion is attempted if postoperative AF
persists beyond 48 hours. These guidelines also make the
following statement regarding nonsurgical patients who
are orally anticoagulated for chronic AF: It is the con-
sensus of the writing group that anticoagulation may be
interrupted for a period of up to 1 week for surgical or
diagnostic procedures that carry high risk of bleeding
without substituting heparin [14]. The recently pub-
lished ACC/AHA 2004 guideline update for CABG sur-
gery [9] recommends warfarin anticoagulation if the
arrhythmia persists beyond 24 hours; however, in the
subsequent text there is further speculation that an
aggressive anticoagulation and cardioversion philosophy
may reduce the neurological complications associated
with this arrhythmia, and it may be advisable to use
intravenous heparin.
Besides the obvious discordance between those guide-
lines, it is not quite clear why a new-onset AF occurring
in the early postoperative period with its self-limited and
intermittent nature represents such a high stroke risk
that an aggressive anticoagulation protocol is advised
within 24 hours, whereas medical patients with chronic
AF undergoing noncardiac operations can be left without
anticoagulation for up to 1 week without significant risk
of stroke.
The cardiac surgical literature has been relatively con-
sistent on this question, considering postoperative AF as
an increased stroke risk [1 4, 7, 9]. However, in a recent
single institutional study involving more than 9,000 pa-
tients, AF dropped out as a risk factor for stroke on
multivariate analysis [6], lending doubt regarding the
strength of this relationship. It is surprising, though, that
there is only one paper in the entire English literature on
this topic discussing temporal relationship between these
two entities. Lahtinen and colleagues [28] from Finland
analyzed data of 52 stroke patients after CABG operation
and found that in 19 patients (36%) the first AF episode
preceded the development of stroke by a mean of 21.3
hours (average, 2.5 AF episodes before stroke). The stroke
was attributed to calcification in the ascending aorta in 13
patients (25%), and 16 patients (31%) had greater than
70% internal carotid artery stenosis. The study does not
state the overall incidence of AF in their CABG popula-
tion or their routine arrhythmia prophylaxis and man-
Ann Thorac Surg
2006;82:51523
agement. Additional variables were preoperative stroke
or TIA (21 cases), significant valvular disease (18 cases),
and history of preoperative AF (5 cases), and 8 of the
above AF 19 cases suffered a stroke in spite of anticoag-
ulation. Similarly no information was given on other
potential intraoperative risk factors, such as left atrial
vent placement, that might be relevant in surgical
patients.
A recent analysis from the Texas Heart Institute Car-
diovascular Research Database [7] compared 994 CABG
patients with postoperative AF and 5,481 patients without
arrhythmia. Atrial fibrillation was found to be an inde-
pendent predictor of long-term mortality at 5 years
(adjusted odds ratio, 1.5). It was also associated with
greater in-hospital mortality (odds ratio, 1.7), with more
perioperative strokes (odds ratio, 2.02), and with pro-
longed hospital stay (14 versus 10 days). However, pa-
tients who developed AF were older, more often hyper-
tensive, had chronic obstructive pulmonary disease,
noncoronary vascular disease, congestive heart failure,
and more severe underlying coronary artery disease, all
of which are strong predictors for stroke without surgery.
They were also likely to have had an intraaortic balloon
pump placed and longer cardiopulmonary bypass time.
Although the authors performed a case-matched suba-
nalysis to strengthen their argument, the adequacy of the
adjustment was limited and therefore the reported asso-
ciations may not be truly independent, as pointed out in
the editorial comment [29]. Also no temporal relationship
between AF and stroke was reported in this series.
The Multicenter Study of Perioperative Ischemia Re-
search Group and investigators of the Ischemia Research
and Education Foundation have recently published their
prospective study performed in 70 hospitals on 4 conti-
nents [8]. This study included more than 5,000 patients
undergoing CABG operations with or without valve
surgery on cardiopulmonary bypass. Patients with post-
operative AF were significantly older (67.8 years versus
61.8 years), and a significantly larger number had history
of AF (14.6% versus 6.0%), valvular disease (27.8% versus
14.9%), congestive heart failure (40.1% versus 32.0%),
chronic obstructive pulmonary disease (14.0% versus
8.6%), and prior neurologic event (13.0% versus 9.3%).
The overall incidence of postoperative AF was 32.3%,
with 43% of patients having more than one episode and
22% having more than two episodes. The overall stroke
rate for patients without AF was 1.2%, whereas it was
0.93% in patients with one episode of AF and 1.4% in
patients with more than one episode of AF. The incidence
of composite neurologic outcomes (stroke, encephalopa-
thy, and stroke score changes postoperatively) was sig-
nificantly higher for patients with more than one episode
of AF as compared with patients with normal sinus
rhythm, but not for patients with only one episode of AF.
Patients with more than one episode of AF had an overall
higher mortality (4.7% versus 2.1%) than patients without
AF, and composite complication outcome was also sig-
nificantly higher (22.6% versus 15.4%, respectively). In
this study, 56.2% of all patients with AF were started on
intravenous heparin, and 17.6% received Coumadin, al-
KOLLAR ET AL 521
ATRIAL FIBRILLATION AND STROKE AFTER CABG
though the effectiveness of anticoagulation and its effect
on neurologic outcomes was not analyzed. The authors
conclude that sicker patients develop AF more often and
they also have a higher incidence of other postoperative
complications, which, however, did not seem to apply to
CARDIOVASCULAR
patients with only one episode of postoperative AF.
The Northern New England Cardiovascular Disease
Group analyzed more than 11,000 CABG patients data
from 1996 to 2001 [30] and found AF was a positive
predictor for stroke (odds ratio, 1.82). Interestingly, how-
ever, prolonged cardiopulmonary bypass time (longer
than 114 minutes; odds ratio, 2.36) and prolonged inotro-
pic agent use (odds ratio, 2.59), which had previously
been established as predisposing factors for AF, were
stronger predictors for stroke than AF by itself. In spite of
their large database, a temporal relationship could only
be determined in less than one third of the stroke
patients, but even those cases were unconfirmed. This
same group also analyzed the etiologic mechanism of
stroke in 388 patients operated on from 1992 to 2000 [31,
and reported that almost two thirds of all cerebrovascular
events occurred within 2 days of the operation and 38%
were classified as nonembolic. In these patients even
early anticoagulation with intravenous heparin could not
have prevented stroke.
Our 12-year surgical experience is limited compared
with large institutional or pooled data and includes
slightly more than 2,900 CABG cases with 19.4% AF
incidence, and a stroke incidence of 1.1% (32 patients).
We have not performed a detailed retrospective risk
factor analysis, but according to the Society of Thoracic
Surgeons database yearly reports, our overall risk profile
is comparable to the national database population. Our
telemetry monitoring system, however, includes simul-
taneous continuous monitoring of atrial electrograms
and surface electrocardiograms, enabling us to identify
the exact nature and onset of postoperative arrhythmias
more precisely, particularly supraventricular arrhyth-
mias, and to reliably detect all episodes of AF. Therefore,
we believe that our recorded AF incidence is accurate,
with atrial arrhythmias rarely if ever missed during
hospitalization. Moreover, the prompt identification of
the nature of the arrhythmia allowed us to establish a
temporal relationship between these episodes and the
neurologic complication in every case.
According to our initial analysis the overall incidence
of stroke in patients with AF was higher (2.6%), than in
patients who maintained normal sinus rhythm (0.7%)
during their hospital stay. However, of the 15 patients
who had both stroke and AF, 9 patients had their neuro-
logic event before the first episode of arrhythmia, with 6
having intraoperative stroke. Of the 6 patients who
experienced atrial arrhythmia before stroke (1.0% true
incidence of stroke in AF patients), 1 was an emergency
case with preoperative myocardial infarction and AF
episodes who developed left atrial appendage thrombus
in spite of heparin anticoagulation and suffered an intra-
operative stroke (case 6). Three (nos. 24, 25, and 27)
strokes occurred more than 1 week after the patients
were successfully converted and stayed in normal sinus
 522 KOLLAR ET AL
ATRIAL FIBRILLATION AND STROKE AFTER CABG
rhythm (no indication for anticoagulation). The last 2
patients (nos. 7 and 22) had preoperative history of TIA
without significant carotid disease; both of these patients
had two to three short episodes (6 hours duration) of
AF or atrial flutter with successful conversion each time
CARDIOVASCULAR
and again no indication (arrhythmia lasting more than 24
to 48 hours) for intravenous heparinization. Our policy of
routine -blocker plus aspirin plus rate control for AF
episodes, and delayed oral anticoagulation (started only
after 48 hours of persistent or recurrent AF) yielded no
strokes that could have been prevented with more ag-
gressive anticoagulation as suggested by the ACC/AHA
practice guidelines.
From our clinical experience and the literature review
presented above, we suggest that there is insufficient
evidence indicating a high enough incidence of strokes as
a direct result of AF after CABG to justify early aggressive
anticoagulation, particularly full heparinization as bridg-
ing therapy in the recently operated patients. The only
data not consistent with this view are the previously
described study from Finland. In that study the average
time between the first AF episode and the neurologic
event was 21.3 hours, less than the ACC/AHA recom-
mended time interval to start anticoagulation for postop-
erative AF. In other words, these neurologic episodes
may not have been preventable by anticoagulation.
We conclude that the development of AF, particularly
recurrent AF in spite of routine -blocker prophylaxis,
may simply identify sicker patients with a higher stroke
risk profile, and the apparent association between these
two entities may be in large part a parallel phenomenon
similar to medical patients. Furthermore, there is no clear
evidence to suggest that aggressive chemical or electrical
cardioversion and prompt anticoagulation with intrave-
nous heparin could prevent strokes in any significant
number. Many patients with postoperative AF have an
overall low risk profile for thromboembolic stroke, and
considering the self-limiting nature of the postoperative
arrhythmia it seems appropriate to treat these patients
similar to medical patients, with aspirin alone. There is
no convincing evidence to suggest that early aggressive
anticoagulation, particularly intravenous heparinization
as bridging therapy, will decrease the already low inci-
dence of postoperative strokes. On the contrary, there is
evidence that early postoperative heparin administration
increases the risk of bleeding complications [32].
It is our opinion that if continuous AF persists, lasting
more than 48 hours, and the patient is expected to remain
in AF, it seems appropriate to start Coumadin, and
therapeutic oral anticoagulation (international normal-
ized ratio. 2.0 to 3.0) could be achieved gradually in an
outpatient setting without an increased risk of cardiac
tamponade. Such a policy should not and in our series
has not inordinately delayed hospital discharge.
References
1. Creswell LL, Schuessler RB, Rosenbloom M, Cox JL. Hazards
of postoperative atrial arrhythmias. Ann Thorac Surg 1993;
56:539 49.
Ann Thorac Surg
2006;82:51523
2. Mathew JP, Parks R, Savino JS, et al. Atrial fibrillation
following coronary artery bypass graft surgery: predictors,
outcomes, and resource utilization. MultiCenter Study of
Perioperative Ischemia Research Group. JAMA 1996;276:
300 6.
3. Almassi GH, Schowalter T, Nicolosi AC, et al. Atrial fibril-
lation after cardiac surgery: a major morbid event? Ann Surg
1997;226:50113.
4. Hravnak M, Hoffman LA, Saul MI, et al. Atrial fibrillation:
prevalence after minimally invasive direct and standard
coronary artery bypass. Ann Thorac Surg 2001;71:14915.
5. Amar D, Shi W, Hogue CW Jr, et al. Clinical prediction rule
for atrial fibrillation after coronary artery bypass grafting.
J Am Coll Cardiol 2004;44:1248 53.
6. DAncona G, Saez de Ibarra JI, Baillot R, et al. Determinants
of stroke after coronary artery bypass grafting. Eur J Cardio-
thorac Surg 2003;24:552 6.
7. Villareal RP, Hariharan R, Liu BC, et al. Postoperative atrial
fibrillation and mortality after coronary artery bypass sur-
gery. J Am Coll Cardiol 2004;43:742 8.
8. Mathew JP, Fontes ML, Tudor IC, et al. A multicenter risk
index for atrial fibrillation after cardiac surgery. JAMA
2004;291:1720 9.
9. Eagle KA, Guyton RA, Davidoff R, et al. ACC/AHA 2004
guideline update for coronary artery bypass graft surgery:
summary article. A report of the American College of Car-
diology/American Heart Association Task Force on Practice
Guidelines (Committee to Update the 1999 Guidelines for
Coronary Artery Bypass Graft Surgery). J Am Coll Cardiol
2004;44:1146 54, e213310.
10. Crystal E, Connolly SJ, Sleik K, Ginger TJ, Yusuf S. Inter-
ventions on prevention of postoperative atrial fibrillation in
patients undergoing heart surgery: a meta-analysis. Circu-
lation 2002;106:75 80.
11. Risk factors for stroke and efficacy of antithrombotic therapy
in atrial fibrillation. Analysis of pooled data from five ran-
domized controlled trials. Arch Intern Med 1994;154:1449
57.
12. The SPAF III Writing Committee for the Stroke Prevention
in Atrial Fibrillation Investigators. Patients with nonvalvular
atrial fibrillation at low risk of stroke during treatment with
aspirin: Stroke Prevention in Atrial Fibrillation III Study.
JAMA 1998;279:12737.
13. Hart RG, Halperin JL, Pearce LA, et al. Lessons from the
Stroke Prevention in Atrial Fibrillation trials. Ann Intern
Med 2003;138:831 8.
14. Fuster V, Ryden LE, Asinger RW, et al. ACC/AHA/ESC
guidelines for the management of patients with atrial fibril-
lation: executive summary. A report of the American College
of Cardiology/American Heart Association Task Force on
Practice Guidelines and the European Society of Cardiology
Committee for Practice Guidelines and Policy Conferences
(Committee to Develop Guidelines for the Management of
Patients With Atrial Fibrillation): developed in collaboration
with the North American Society of Pacing and Electrophys-
iology. J Am Coll Cardiol 2001;38:1231 66.
15. Frye RL, Kronmal R, Schaff HV, Myers WO, Gersh BJ. Stroke
in coronary artery bypass graft surgery: an analysis of the
CASS experience. The participants in the Coronary Artery
Surgery Study. Int J Cardiol 1992;36:21321.
16. Gardner TJ, Horneffer PJ, Manolio TA, et al. Stroke following
coronary artery bypass grafting: a ten-year study. Ann
Thorac Surg 1985;40:574 81.
17. Angelini GD, Penny WJ, el-Ghamary F, et al. The incidence
and significance of early pericardial effusion after open heart
surgery. Eur J Cardiothorac Surg 1987;1:165 8.
18. Kuralay E, Ozal E, Demirkili U, Tatar H. Effect of posterior
pericardiotomy on postoperative supraventricular arrhyth-
mias and late pericardial effusion (posterior pericardiotomy).
J Thorac Cardiovasc Surg 1999;118:4925.
19. Cummings JE, Gill I, Akhrass R, Dery M, Biblo LA, Quan KJ.
Preservation of the anterior fat pad paradoxically decreases
Ann Thorac Surg
2006;82:51523
the incidence of postoperative atrial fibrillation in humans.
J Am Coll Cardiol 2004;43:994 1000.
20. Sparks PB, Jayaprakash S, Mond HG, Vohra JK, Grigg LE,
Kalman JM. Left atrial mechanical function after brief dura-
tion atrial fibrillation. J Am Coll Cardiol 1999;33:3429.
21. Stoddard MF, Dawkins PR, Prince CR, Ammash NM. Left
atrial appendage thrombus is not uncommon in patients
with acute atrial fibrillation and a recent embolic event: a
transesophageal echocardiographic study. J Am Coll Cardiol
1995;25:4529.
22. Beyth RJ, Milligan PE, Gage BF. Risk factors for bleeding in
patients taking coumarins. Curr Hematol Rep 2002;1:419.
23. Gage BF, van Walraven C, Pearce L, et al. Selecting patients
with atrial fibrillation for anticoagulation: stroke risk strati-
fication in patients taking aspirin. Circulation 2004;110:2287
92.
24. Rockson SG, Albers GW. Comparing the guidelines: antico-
agulation therapy to optimize stroke prevention in patients
with atrial fibrillation. J Am Coll Cardiol 2004;43:929 35.
25. Wang TJ, Massaro JM, Levy D, et al. A risk score for
predicting stroke or death in individuals with new-onset
atrial fibrillation in the community: the Framingham Heart
Study. JAMA 2003;290:1049 56.
The Society of Thoracic Surgeons Policy Action Center
The Society of Thoracic Surgeons (STS) is pleased to
announce a new member benefitthe STS Policy Action
Center, a website that allows STS members to participate
in change in Washington, DC. This easy, interactive,
hassle-free site allows members to:
Personally contact legislators with ones input on key
issues relevant to cardiothoracic surgery
Write and send an editorial opinion to ones local media
2006 by The Society of Thoracic Surgeons
Published by Elsevier Inc
KOLLAR ET AL 523
ATRIAL FIBRILLATION AND STROKE AFTER CABG
26. Wyse DG, Waldo AL, DiMarco JP, et al. A comparison of rate
control and rhythm control in patients with atrial fibrillation.
N Engl J Med 2002;347:182533.
27. Van Gelder IC, Hagens VE, Bosker HA, et al. A comparison
of rate control and rhythm control in patients with recurrent
persistent atrial fibrillation. N Engl J Med 2002;347:1834 40.
CARDIOVASCULAR
28. Lahtinen J, Biancari F, Salmela E, et al. Postoperative atrial
fibrillation is a major cause of stroke after on-pump coronary
artery bypass surgery. Ann Thorac Surg 2004;77:1241 4.
29. Levy D, Kannel WB. Postoperative atrial fibrillation and
mortality: do the risks merit changes in clinical practice?
J Am Coll Cardiol 2004;43:749 51.
30. Likosky DS, Leavitt BJ, Marrin CA, et al. Intra- and postop-
erative predictors of stroke after coronary artery bypass
grafting. Ann Thorac Surg 2003;76:428 35.
31. Likosky DS, Marrin CA, Caplan LR, et al. Determination of
etiologic mechanisms of strokes secondary to coronary ar-
tery bypass graft surgery. Stroke 2003;34:2830 4.
32. Fanikos J, Tsilimingras K, Kucher N, Rosen AB, Hieblinger
MD, Goldhaber SZ. Comparison of efficacy, safety, and cost
of low-molecular-weight heparin with continuous-infusion
unfractionated heparin for initiation of anticoagulation after
mechanical prosthetic valve implantation. Am J Cardiol
2004;93:24750.
E-mail senators and representatives about upcoming
medical liability reform legislation
Track congressional campaigns in ones districtand
become involved
Research the proposed policies that help or hurt
ones practice
Take action on behalf of cardiothoracic surgery
This website is now available at www.sts.org/takeaction.
Ann Thorac Surg 2006;82:523 0003-4975/06/$32.00