Erythrocyte sedimentation rate, coronary atherosclerosis, and

cardiac mortality
Andrea Natali , Antonio L'Abbate , Ele Ferrannini
DOI: http://dx.doi.org/10.1016/S0195-668X(02)00741-8 639-648 First published online: 1 April
2003
Article
Figures & data

Information

Explore

PDF

Abstract
Aims To test whether the erythrocyte sedimentation rate (ESR) is related to the extension of
coronary atherosclerosis (ATS) and predicts cardiac mortality.
Methods and results Hospital-based, retrospective observational (median follow up: 92 months)
cohort study. In 1726 consecutive patients undergoing angiography, coronary ATS and
subsequent mortality were related to ESR and to classical risk factors. Patients
undergoing angiography for reasons different from ischemic heart disease (IHD), served as
control. ESR was progressively higher in the presence of 1, 2, or 3-vessel disease. Age-and sexadjusted ESR was positively related to ATS both in univariate analysis (
,
) and in
a multivariate model including principal risk factors (partial
,
). Similar
associations were observed in the control group. Over the follow-up period, 170 patients died of
a cardiac cause. When male and female patients in their upper ESR quartile (>18 and >23mm/h,
respectively) were compared with the remainder of the cohort, their age-and gender-adjusted
odds ratio for cardiac mortality was 1.72 (
,
). This result held true, in
men, also when using a full set of risk factors in a Cox model.
Conclusions ESR is an independent correlate of coronary ATS and, in male patients with
probable IHD, a predictor of cardiac death.

Erythrocyte sedimentation rate

Inflammation

Coronary atherosclerosis

Cardiac mortality

Risk factors

Ischemic heart disease

1 Introduction
Histopathologic studies have consistently shown that inflammatory cells not only participate in
the initiating events leading to plaque formation(adesion of monocytes to activated endothelial
cells, migration into subendothelial layers, and transformation into macrophages), but also affect
the subsequent evolution of the lesion. Activated T lymphocytes are present in the lipid core of
atherosclerotic lesions and, through the release of cytokines, stimulate macrophage migration
andactivation, inhibit collagen synthesis, and induce neovascularization.1 Thus, various indices
of systemic inflammation (white blood cell count, serum interleukin-6, fibrinogen, C-reactive
protein, and albumin concentrations) have been reported to predict ischemic heart disease (IHD)
development both in healthy subjects and in at-risk patients.2 Whether inflammation is a risk
factor for IHDbecause it favours the occurrence of clinical events (e.g. through plaque
destabilization and/or blood clotting) or because it heralds the presence of a more extensive
coronary atherosclerosis (ATS), is unknown. Similarly, the notion that rheologic characteristics
of blood contribute to the pathogenesis of IHD is gaining growing support. The major
components of blood viscosity are the blood cell mass (i.e. the hematocrit), the intrinsic
resistance of the plasma to flow (commonly measured by capillary viscometry), and red blood
cell aggregability, which can be estimated through the erythrocyte sedimentation rate (ESR). A
recent meta-analysis of population-based studies3 calculated that the IHD risk ratios of subjects
in the top vs bottom tertile of the distribution of hematocrit, plasma viscosity, and ESR were
1.60, 1.57, and 1.33, respectively. Even higher risk ratios (1.81 for hematocrit and 2.60 for
plasma viscosity) were found in patients with preexisting cardiovascular disease, whereas no
data are available on the prognostic value of ESR in these patients. In general, in comparison
with the other indices of blood viscosity (19 studies of haematocrit and six of plasma viscosity,3)
ESR has been somewhat neglected, with two large4,5 and three very small68 prospective studies
published in the last three decades. One reason might be that the ESR can only be measured on
fresh bloodsamples, and is confounded by gender, age,hematocrit, as well as the presence of
many acute or chronic illnesses. On the other hand, the ESR measurement by the Westergren
method is standardized, accurate, universally available, and cheap. Furthermore, when red blood
cell physical characteristics are taken into account by adjusting for hematocrit, ESR largely
reflects the plasma concentration of acute phase response proteins resulting in a compound index
of both viscosityand inflammation. Despite the rationale, relatively scarce information is
available on theassociation of ESR with the anatomical status ofthe coronary circulation. This
prompted us to carry out the present retrospective analysis of patients undergoing coronary
angiography for the clinical workup of ischemia, in whom we tested whether ESR was related to
the severity of coronary ATS and whether it predicted subsequent cardiac death.

2 Methods
2.1 Patients
We analyzed data from 2263 consecutive patients undergoing coronary angiography over the
decade 19831992. Most of the patients
underwent coronary angiography as part of the
clinical workup for IHD; of these, 228 (12%) had previously suffered a myocardial infarction
(PMI), 45 (2%) were admitted with an acute myocardial infarction (AMI), and 1679 (86%)

presented with either resting or effort angina. ESR data were available for 1726 of these patients
(88%); the clinical characteristics of this group (Table 1) were superimposable on those of the
whole population (
for all variables by MannWhitney U test, data not shown). In the
remaining 311 subjects, coronary angiography was performed as part of the clinical evaluation
for valvular disease
, dilated or hypertrophic cardiomyopathy
, arrhythmia
, or other
. ESR was available for 86%
of these patients.

Enlarge table

Table 1
Clinical characteristics of the patients
Mean
SD Range
n
1726
M/F
1380/346
Age (years)
55
10
1784
BMI (kg m2)
26.4
3.2 14.939.5
Familial IHD (score)
1.1
2.1 022
Alcohol consumption (g/week)
166
182 01893
Diabetes (%)
12.2
Smoking habits (n/ex/y; in %)
24/53/23
a
Cigarette consumption (pack-years) 32
(32) 0.1225
Systolic BP (mmHg)
128
14
90188
Diastolic BP (mmHg)
81
8
59110
Heart rate (bpm)
64
8
4298
Hematocrit (%)
40.1
3.9 26.056.1
Serum total cholesterol (mmol/l)
5.57
1.32 2.5921.76
b
Serum triglycerides (mmol/l)
1.66
(1.02) 0.4025.4
a Geometric mean and (interquartile range) for the 1278 subjects who were past- or
current-smokers.
b Geometric mean and (interquartile range).

IHD=ischemic heart disease; BMI=body mass index; BP=arterial blood pressure;

n/ex/y=never smokers/past-smokers/ current-smokers.

2.2 Database
Since 1983, a standard record was completedfor each patient undergoing angiography by the
cardiologist in charge. Before permanent storage, each record was checked by a computer expert
and senior cardiologist, for logical or clinical inconsistencies as well as uniformity of entry
criteria.;comptd;;center;stack;;;;;6;;;;;width> ;comptd;;center;stack;;;;;6;;;;;width> Each record
consisted of the following three blocks of information: (a) Demographics: Gender, age, height,
weight; (b) Clinical data: PMI or AMI, resting, effort, and unstable angina all

diagnosedaccording to clinical criteria; (c) Angiographic data: Any stenosis (graded as 50, 75,
90, and 100%) along each coronary artery (main stem (MS), left anterior descending artery
(LAD), circumflex artery (CX), and right coronary artery (RCA)) as well as secondary branches
was described. A coronaryvessel was considered to have a clinicallysignificant obstruction if it
was stenosed by at least 50% with respect to the prestenotic segment; when multiple stenoses
were present on the same main vessel, the grade of the narrowest one was considered. These
descriptions constituted the basis for the diagnosis of 1-, 2-, or 3-vessel disease.From the
description of the angiograms, a score of cumulative coronary ATS (expressed in arbitrary units9)
was calculated as the sum of all detectable percent narrowings in the entire coronarytree
(including multiple stenoses along thesame vessel; irregularities were considered asa single
40% stenosis). Formal validation ofthe visual grading system against quantitative angiography
computer-assisted edge evaluation (Kontron M14, Kontron Image Analysis Division,Munich,
Germany)was carried out in a subset
of stenoses randomly selected from the present
data collection in a blind fashion. When the percent reduction of the stenotic segment relative to
the pre-stenotic segment was calculated as the ratio of the two areas, the values coded as
irregularities, 50, 75, and 90% by visual grading corresponded to 535, 743, 801, 831%
narrowing, respectively, on quantitative angiography; and (d) Follow-up data: Information
concerning death was collected yearly up to 15 years after discharge from the following sources:
outpatient visits, telephone interviews (with the patients, close relatives, or the referring
physician), questionnaires sent to all patients, and death certificates. The data were interpreted by
the cardiologist in charge of the database, who classified the events as: no event, cardiac death
(which included sudden death, acute left ventricular dysfunction, fatal myocardialinfarction, and
death during or immediately after coronary bypass surgery), and noncardiac death (which
included cancer, accident, stroke, acute pulmonary disease, etc.).

Download figure
Open in new tab

Download powerpoint

Fig. 1
ESR according to the number of vessel with clinically significant (50%) coronary stenosis (top)
and to the extent of coronary ATS units (bottom) in 1380 men and 346 women. Each point plots
the geometric mean.

Download figure
Open in new tab

Download powerpoint

Fig. 2
ATS by gender and quartile of ESR in diabetic and nondiabetic patients (top) and in patients with
normal or high (>5.2mmol/l) serum total cholesterol concentrations (bottom). Bars are geometric
means of ATS.

2.3 Clinical chart review

The original database was integrated with additional information collected retrospectively
through detailed review of the individual clinical charts, as follows: (a) Familial IHD was
expressed as a score derived from the number of first-degree relatives affected (1 for a parent,
for each affected of siblings). Second-degree relatives were also included (with a score of
0.1). Theses values were multiplied by 5, if the subject was affected before 60 years of age (for
men, 70 for women); (b) Diabetes mellitus was defined bythe presence of antidiabetic treatment
or if two measurements of fasting plasma glucose were greater than 7.0mmol/l.10 Hypertension
wasdefined by current antihypertensive treatment, or blood pressure values (as the average
morning reading) 140/90mmHg11 or a documented history of high blood pressure; (c) Patients
were classified as current-smokers, past-smokers, or never-smokers. Cigarette consumption was

also estimated, and expressed in pack-years (when not specified, age 17 was assumed as the
beginning). Current alcohol consumption was expressed in grams per week based on the average
weekly consumption of wine, beer, and spirits. An average ethanol content of 11g for each wine
serving (120ml), 13g for each beer serving (250ml), and 15g for each spirit serving (40ml)
wasassumed; (d) From the routine chemistry done on fasting blood on admission, the following
data were also taken: plasma creatinine and uric acid, serum total cholesterol, and triglycerides,
leukocyte count, ESR, and hematocrit.

2.4 Methods
Cardiac catheterization was performed by the Seldinger technique. During ventricular and
coronary angiography, at least three different views were recorded for each patient on a 35-mm
film. Blood samples were analysed by the samelaboratory. Plasma glucose, cholesterol, and
triglyceride concentrations were measured by routine enzymatic methods; hematocrit was
measuredby microultracentrifugation. ESR was measured by the modified Westergren method
(automated ona VES-Matic); in our laboratory, upper limits of normal are 12 and 17mm/h, in
adult men and women, respectively and the coeficient of variation of duplicates is 9%.

2.5 Statistical analysis

Dataare presented as meanSD. ESR and serum triglyceride values were log-transformed for
statistical analysis and mean data given as geometric mean and interquartile range. Forcigarette
consumption and the ATS score, the
transformation was used in statistical analysis.
Between-group mean differences were tested by the MannWhitney U test (KruskallWallis test
for three or more groups) or the 2test, for continuous variables and proportions,respectively.
When comparing mean group values stratified by another ordinal variable, two-way analysis of
variance (ANOVA) was used. Simple and multiple regression analyses were carried outby
standard methods. Survival was analyzed by generating KaplanMeier cumulative hazard plots
for groups, which were then compared by the logrank (MantelCox) test. Odds ratios (OR)and
their 95% confidence intervals (CI) werecalculated using Cox regression models.

Enlarge table

Table 2
Multiple regression analysis
ATS
Intercept
Male sex
Age
Diabetes
Serum cholesterol
Smoking

LAD
CX
RCA
Total
3.00 (4.03) 3.21 (2.71) 3.51 (3.53)2.94 (5.13)
0.32 (0.30) 0.25 (0.26) 0.30 (0.26)0.39
(0.33)
0.23 (0.19) 0.19 (0.21) 0.19 (0.18)0.23
(0.24)
0.11 (0.17) 0.16 (0.21) 0.13 (0.14)0.14
(0.17)
0.13 (ns)
0.13 (ns)
0.16 (0.10)
0.16 (0.12)
0.05 (ns)
0.12 (ns)
0.12 (ns)
0.13 (ns)

ATS
LAD
CX
RCA
Total
ESR
0.09 (0.30) 0.07 (0.16) 0.10 (0.20)
0.12 (0.24)
Alcohol consumption ns (0.10)
0.08 (0.12) 0.07 (0.12)0.06 (0.12)
Hypertension
0.05 (ns)
ns (ns)
ns (ns)
0.05 (ns)
Explained variance (%) 20 (29)
17 (21)
20 (25)25
30 (29)
Entries are standardized coefficients, at a significance level less than 0.05, for the patients
undergoing angiography for the clinical work-up of IHD and in the control group (in
bracket); ns=not significant. LAD=left anterior descending coronary artery;
CX=circumflex coronary artery; RCA=right coronary artery. ESR was entered as
,
smoking (pack-years) and coronary stenosis as
.

3 Results
ESR was higher in women than in men (14 [15] vs 8 [10]mm/h,
); in either sex, it
increased with age (by 2.4mm/h per decade,
,
in men; 4.7mm/h per decade,
,
in women), and decreasedwith increasing hematocrit (by 1.3mm/h per unit
of hematocrit,
,
in men; 0.8mm/h per unit of hematocrit,
,
in women). In nonsmokers, ESR was significantly higher than in past- or currentsmokers (16 vs 13 vs 12mm/h,
), whereas the hematocrit showed the reverse trend (38.4
vs 40.5 vs 40.8%,
); past-and current-smokers were more frequently men (93 and 88%,
respectively), whereas women were more prevalent (57%) among never smokers. After adjusting
for sex and hematocrit, ESR was not affected by smokinghabits. Of the 213 diabetic patients
(155 men, 58 women), five had type 1 diabetes and 208 had type 2 diabetes. In the whole
diabetic group, ESR was higher than in nondiabetic subjects, in men (10 [13] vs 8 [10]mm/h,
) as well as women (20 [17] vs 13 [15]mm/h,
). In the 689 patients with
hypertension (535 men, 154 women), ESR was higher than in normotensive subjects, in men (9
[11] vs 8 [9]mm/h,
) and women (16 [16] vs 13 [15]mm/h,
). In univariate
analysis, ESR was found to be positively related to body mass index, heart rate, serum total
cholesterol, and smoking, and inversely related to alcohol consumption. In multivariate analysis,
hematocrit, age,serum total cholesterol, heart rate, body massindex, diabetes, gender, and alcohol
consumption (in that order) were independent predictors of ESR (
and
for the
model).
ESR was progressively higher in the presence of angiographically documented major narrowing
(>50%) of 0, 1, 2, or 3 major coronary vessels(
by two-way ANOVA), and was
positively related to the ATS score in univariate analysis in both men (
,
) and
women (
,
) (Fig. 1). The ATS score was higher in men than women (140 [280] vs
19 [271] units,
), in diabetic than nondiabetic patients (240 [280] vs. 82 [267] units,
), in hypertensive than normotensive patients (125 [300] vs 76 [285] units,
), and
in past-smokers (139 [285] units) than in current or never-smokers (103 [240] and 36 [344] units,
respectively,
). In univariateassociation, the ATS score was also higher
with;comptd;;center;stack;;;;;6;;;;;width> increasing age, serum total cholesterol and triglyceride
levels, and hematocrit. By stepwiseregression analysis, sex, age, serum cholesterol, presence of
diabetes, cigarette consumption,alcohol consumption (negatively), hypertension (in that order)

were all independently associated withthe ATS score, together accounting for 28% of its
variability.

Download figure
Open in new tab

Download powerpoint

Fig. 3
Cumulative hazard plot for all-cause and cardiac mortality in men (squares) and women (circle)
in the top quartile (filled symbols) or in the lower three quartiles (empty symbols) of ESR
distribution. These patients (1380 men and 346 women) underwent coronary angiography for the
clinical work up of ischemic heart disease. The values refer to the logrank test.
When patients were stratified by sex and presence of diabetes, the independent
associationbetween ESR and ATS was still statistically significant (by two-way ANOVA) in both
men
and women
(Fig. 2). By defining hypercholesterolemia as a serum total
cholesterol >5.2mmol/l, 999 patients (206 women) werehypercholesterolemic. Upon stratifying
the cohort by sex and hypercholesterolemia, ATS was related to ESR in both men and women (
for the independent effect of ESR) (Fig. 2). To further test the independent association
between ESR and ATS, a multivariate model was set up that included covariates common to both
(age, diabetes, and serum cholesterol), as well as all variables that were univariately related to

either ESR or ATS. The results (Table 2) show that coronary stenosisremained independently
associated with ESR (for each of the three main coronary arteries andglobally). Furthermore, the
model yielded similar standardized regression coefficients in the patients undergoing
angiography for the clinical workup of IHD and in the control group. The latter had a better
cardiovascular risk profile (with a higher prevalence of never smokers (39 vs 24%), lower serum
cholesterol (4.981.19 vs 5.571.32mM) and triglycerides (1.58 [0.82] vs 1.67 [1.02]mM)
concentrations,
for all) and a much lower ATS score (5 [50] vs 94 [290] units,
).
Nevertheless, the standardized coefficients indicate a similar, close association between ESR and
coronary ATS compared to the IHD group.
Over a median follow-up period of 92 months (range 0191), 259 patients in the IHD group
died; of these, 170 died of a documented cardiac cause. When the survival of the patients in the
upper quartile of ESR distribution (>18 in men and >23mm/h in women) was compared with that
of the remainder of the cohort, there was a significant increase in both total and cardiac mortality
in men as well as women (
for both) (Fig. 3). The age-and gender-adjusted OR for
cardiac death was 1.72 (
,
).
When the full set of risk factors was used in a Cox model (also including the hematocrit), ESR
values in the upper quartile still behaved as a negative prognostic index, particularly for cardiac
mortality (Table 3). An ESR greater than 18 for men and 23mm/h for women was a predictive
factor also, after adjusting for the presence of PMI and unstable angina (OR: 1.67,
). Fully adjusted cumulative hazard plots for men in the upper ESR quartile (>18mm/h) and for
men with a PMI against the rest of the cohort are compared in Fig.
4;comptd;;center;stack;;;;;6;;;;;width> .

Enlarge table

Table 3
Mortality risk
All causes
p
Cardiac causes p
2.08 (1.323.29) 0.002 2.02 (1.173.50) 0.01
2.01 (1.722.36) <0.001 1.90 (1.552.32) <0.001
1.63 (1.172.27) 0.004 1.86 (1.262.73) 0.002
1.16 (0.901.51) ns
1.11 (0.811.26) ns
0.97 (0.901. 04) ns
0.90 (0.821.00) ns

Sex (male)
Age (10 years)
Diabetes
Hypertension
Alcohol (100g/week)
Smoking
Past
1.50 (1.102.04) 0.01 1.37 (0.932.01) ns
Current
1.73 (1.112.69) 0.02 1.75 (1.013.03) 0.046
High cholesterol (>5.2mM) 1.40 (1.061.84) 0.02 1.53 (1.092.14) 0.014
Hematocrit (5units)
0.99 (0.821.19) ns
1.05 (0.791.32) ns
High ESR (>15mm/h)
1.30 (0.971.74) 0.08 1.71 (1.202.43) 0.003
Entries are OR and (95% CI) generated by the Cox multiple regression model.

Download figure
Open in new tab

Download powerpoint

Fig. 4
Residual (following adjustment for age, diabetes, hypertension, alcohol consumption, smoking
habits, hypercholesterolemia, and hematocrit) cumulative risk of cardiac death in men with an
ESR>18 mm/h (filled symbols) or 18 mm/h (empty symbols). The lower panel shows the
equivalent plot for men with and without previous myocardial infarction (PMI). These patients
(1380 men) underwent coronary angiography for the clinical work up of ischemic heart disease.
The value refer to the log-rank test.

4 Discussion
ESRwhich measures the tendency of red blood cells to aggregateis a time-honored, routine
analysis mainly used to screen for the presence of hidden inflammation, and to help in the
diagnosis and follow-up of chronic diseases. In our cohortof patients largely selected for the
presence of symptoms or signs of IHD, ESR showed an interesting behavior. After adjusting for
sex and hematocrit, ESR was directly associated not only with age, serum cholesterol,
hypertension, heart rate, BMI, and diabetes, and inversely related to alcohol consumption i.e.

with a set of variables that are also well known risk factors for IHD, but also with the extent of
coronary ATS. Since the major determinants of ESR are the concentration of positively charged
inflammatory proteins (fibrinogen, IgM, alpha2-macroglobulins) and the number and the size of
red blood cells, the previously mentioned associations being independent from hematocrit
strongly suggest that subclinical, chronic inflammation is related not only to the presence of each
of the previously mentioned conditions, but also to the ATS that they cause. Acute phase
response indices have been found to be elevated in type 2 diabetics,12 particularly in those
patients with the clinical phenotype of the metabolic syndrome.13 Similarly, a recent analysis of a
large, population-based study has shown that three indices of inflammation, namely C-reactive
protein, white blood cell count, and fibrinogen, tend to cluster with obesity, central distribution
of body fat, higher blood pressure, hyperglycemia, lower HDL-cholesterol, and
hyperinsulinemia,14 which are all risk factors for ATS. To compound the matter, inflammation
indices have been shown to predict the occurrence of IHD events. Thus, C-reactive protein
predicted myocardial infarction in a hospital-based study,15 and ESR-predicted future IHD events
in a population-based study.3 Whether this predictive power results from a direct association of
ESR with the extent of coronary ATS, is unclear. Few studies have tested whether indices of
inflammation or blood viscosity reflect theanatomical status of the coronary vasculature.In a
study of 573 patients undergoing coronary angiography, the leukocyte count was
directlyassociated with the severity of coronary ATS, but the association lost statistical
significance after adjusting for smoking.16 In another study, serum C-protein, amyloid protein A,
and interleukin-6 concentrations were all found to be elevated in 100 patients with coronary
artery disease in comparison to the 100 matched healthy subjects; however, no difference was
observed among patients with 1-, 2-, or 3-vessel disease.17 No study has tested whether indices of
blood viscosity are associated with more severe coronary ATS.
Our data demonstrate that more severe degrees of angiographically documented coronary ATS
expressed either as number of severely (>50%)stenosed vessels or as the ATS scoreare
associated with higher ESR values (Fig. 1). To rule out the possibility that this association was
entirely due to the relation of ESR with the individual IHD risk factors, we first analyzed the data
after stratifying for other risk factors, such as diabetes andhypercholesterolemia (Fig. 2), and
then used amultivariate regression model on the wholedataset (Table 2). By either approach, ESR
was an independent correlate of coronary ATS. In statistical terms, ESR explained only a small
fraction of coronary ATS; this, however, was also true for all the classical IHD risk factors, in
keeping withprevious observations.1820
Although our estimate of ATS does not take into consideration the histological characteristics of
the lesions (i.e. their stability), the follow-up data demonstrate that in patients presenting
withprobable IHD, ESR is an independent predictor of cardiac death (
).
Remarkably in men, an ESR greater than 18mm/h carried the same negative prognosis as a PMI
(Fig. 4). This result was similar whether survival was analyzedby quartile of ESR or with ESR as
a continuous variable.
To establish whether the prognostic power of ESR was solely explained by its association with
coronary ATS, we included ATS in a Cox regression model with cardiac deaths as end-points:
whereas sex, hypercholesterolemia, and smoking became statistically nonsignificant predictors,
ESR maintained its predictive power (adjusted OR: 1.49 [2.091.06],
). Furthermore,

neither thepresence of ischemia nor a PMI modified the risk of cardiac death associated with a
high ESR (adjusted OR: 1.41 [1.961.02],
). These analysessuggest alternative pathways
through which ESR influenced the prognosis of our patients.
Our conclusions only apply to patients at high risk for IHD, and cannot be readily extrapolated to
the general population. Nevertheless, it is of interest that ESR was independently related to ATS
(partial
,
adjusted for sex, age,hematocrit, diabetes, and hypertension) also in the
small group of patients
in whom coronary angiography was performed as part of the
clinical workup for valvular disease, cardiomyopathy, or arrhythmia. In this group, coronary ATS
was less extensive than in the primary cohort, and therisk factor profile was more favorable; yet,
the adjusted ESR value was still positively associated with a higher ATS score. In addition, in a
survey conducted in healthy middle-aged working men followed for 23 years, a very similar
estimate of the relative risk for coronary artery disease mortality (OR: 1.59 [1.971.27]) was
observed in the subjects with ESR values grater than 15mm/h.5
In summary, ESR is an independent correlate of coronary ATS and a predictor of cardiac death in
patients with probable IHD. A mild, persistentelevation of ESR should direct physician's
attention to coronary ATS and its consequences.

Acknowledgments
We wish to thank Sara Burchielli for her assistance.

The European Society of Cardiology

References
1.

MethaJ, Saldeen T, Rand K. Interactive role of infection, inflammation and traditional

risk factors in atherosclerosis and coronary artery disease. J Am Coll Cardiol.
1998;31:12171225.
CrossRefMedlineWeb of Science
2.

DaneshJ, Collins R, Appleby P et al. Association of fibrinogen, C-reactive protein,

albumin, or leukocyte count with coronary artery disease: meta-analysis of prospective
studies. JAMA. 1998;279:14771482.
CrossRefMedlineWeb of Science
3.

DaneshJ, Collins R, Peto R et al. Haematocrit, viscosity, erythrocyte sedimentation rate:

meta-analyses of prospective studies of coronary heart disease. Eur Heart J.
2000;21:515520.
FREE Full Text
4.

GillumR, Mussolino M, Makuc D. Erythrocyte sedimentation rate and coronary heart

disease: the NHANES I epidemiologic follow-up study. J Clin Epidemiol. 1995;48:353
361.
CrossRefMedlineWeb of Science
5.

ErikssenG, Liestol K, Bjornholt JV et al. Erythrocyte sedimentation rate: a possible

marker of atherosclerosis and a strong predictor of coronary heart disease mortality. Eur
Heart J. 2000;21(19):16141620.
Abstract/FREE Full Text
6.

BottigerL, Carlson L. Risk factors for ischemic vascular death for men in the Stockholm
prospective study. Atherosclerosis. 1980;36:389408.
CrossRef
7. TiblinG, Wilhelmsen L, Werko L. Risk factors for myocardial infarction and death due to
ischemic heart disease and other causes. Am J Cardiol. 1975;35:514522.
CrossRefMedlineWeb of Science
8. RaffnsonV, Bengtsson C. Erythrocyte sedimentation rateand cardiovascular disease:
results from a population studyof women in Goteborg, Sweden. Atherosclerosis.
1982;42:97107.
CrossRefMedlineWeb of Science
9.

RoweG, Thomsen J, Sternlund RR et al. A study of hemodynamics and coronary blood

flow in man with coronary artery disease. Circulation. 1969;39:139148.
Abstract/FREE Full Text

10.

Report of the expert committee on the diagnosis andclassification of diabetes mellitus.

Diabetes Care. 1997;20:11831197.
FREE Full Text
11.

National high blood pressure education program working group report on primary
prevention of hypertension. Arch Int Med. 1993;153:186208.
CrossRefMedlineWeb of Science
12.

TanKC, Chow WS, Tam SC et al. Atorvastatin lowers C-reactive protein and improves
endothelium-dependent vasodilation in type 2 diabetes mellitus. J Clin Endocrinol
Metab. 2002;87(2):563568.
CrossRefMedlineWeb of Science
13.

PickupJ, Mattock M, Chusney G et al. NIDDM as a disease of the innate immune system:
association of acute-phase reactants and interleukin-6 with metabolic syndrome X.
Diabetologia. 1997;41:12411248.
14.

FestaA, D'agostino R, Howard G et al. Chronic subclinical inflammation as part of the

insulin resistance syndrome. The insulin resistance atherosclerosis study (IRAS).
Circulation. 2000;102:4247.
Abstract/FREE Full Text
15.

HaverkateF, Thompson S, Pyke S et al. Production of C-reactive protein and risk of

coronary events in stable and unstable angina. European Concerted Action on
Thrombosis and Disabilities Angina Pectoris Study Group. Lancet. 1997;349:462466.
CrossRefMedlineWeb of Science
16.

KostisJ, Turkevich D, Sharp J. Association between leukocyte count and the presence and
extnt of coronary atherosclerosis as determined by coronary arteriography. Am J
Cardiol. 1984;53(8):997999.
CrossRefMedlineWeb of Science
17.

RifaiN, Jourbran R, Yu H et al. Inflammatory markers in man with angiographically

documented coronary heart disease. Clin Chem. 1999;45(11):19671973.
Abstract/FREE Full Text
18.

VlietstraR, Kronmal R, Frye R et al. Factors affecting the extent and severity of coronary
artery disease in patients enrolled in the Coronary Artery Surgery Study. Arteriosclerosis.
1982;2:208215.
Abstract/FREE Full Text
19. FreedmanD, Gruchow H, Bamrah V et al. Diabetes mellitus and arteriographicallydocumented coronary artery disease. J Clin Epidemiol. 1988;41(7):659668.
CrossRefMedlineWeb of Science
20. HochmanJ, Phillips W, Ruggieri D et al. The distribution of the atherosclerotic lesion in
the coronary arterial tree: relation to cardiac risk factors. Am Heart J. 1988;116:1217
1222.
CrossRefMedlineWeb of Science