Download Relationship Between C-Reactive Protein Levels and

Journal of the American College of Cardiology
© 2007 by the American College of Cardiology Foundation
Published by Elsevier Inc.
Vol. 49, No. 5, 2007
ISSN 0735-1097/07/$32.00
doi:10.1016/j.jacc.2006.09.040
Relationship Between C-Reactive
Protein Levels and Regional Left
Ventricular Function in Asymptomatic Individuals
The Multi-Ethnic Study of Atherosclerosis
Boaz D. Rosen, MD,* Mary Cushman, MD,† Khurram Nasir, MD, MPH,*
David A. Bluemke, MD, PHD,‡ Thor Edvardsen, MD, PHD,* Verônica Fernandes, MD, PHD,*
Shenghan Lai, MD, PHD,§ Russell P. Tracy, PHD,† João A. C. Lima, MD, FACC*‡
Baltimore, Maryland; and Burlington, Vermont
Objectives
This study sought to investigate the relationship between C-reactive protein (CRP) and regional left ventricular
(LV) function in asymptomatic individuals without a history of cardiovascular disease.
Background
C-reactive protein is associated with an increased risk for developing cardiovascular disease. However, the relationship between CRP and subclinical LV dysfunction has not been evaluated in asymptomatic individuals.
Methods
Regional myocardial function was analyzed as peak systolic circumferential shortening strain (Ecc) using the
harmonic-phase method by tagged magnetic resonance imaging in 1,164 individuals without symptomatic cardiovascular disease from the MESA (Multi-Ethnic Study of Atherosclerosis) trial (age 66.4 ⫾ 9.6 years old). Regions were
defined by coronary territories: left anterior descending artery (LAD), left circumflex artery (LCX), and right coronary
artery (RCA). The relationship between log-CRP concentration and Ecc was studied by multivariable linear regression
after adjustment for demographic characteristics, risk factors, and therapy (including hormone replacement therapy).
Results
For each region, associations differed by gender with no association of CRP and regional LV function among
women. In men, after adjustment, higher log-CRP was significantly associated with lower (absolute) Ecc in the
LAD and RCA regions (regression coefficient 0.37 per unit higher log-CRP [95% confidence interval [CI] 0.08 to
0.65] and 0.31 [95% CI 0.02 to 0.59], respectively) and peak systolic Ecc overall (regression coefficient 0.32
[95% CI 0.05 to 0.58]). In the LCX region, the association was weaker (p ⫽ 0.06).
Conclusions
Among individuals without evident heart failure or other cardiovascular disorders, higher CRP was associated
with lower systolic myocardial function in all regions in men but not in women. These findings support the role of
inflammation and atherosclerosis in incipient myocardial dysfunction. (Multi-Ethnic Study of Atherosclerosis;
http://clinicaltrials.gov/ct/show/NCT00005487). (J Am Coll Cardiol 2007;49:594–600) © 2007 by the
American College of Cardiology Foundation
Inflammation plays an important role in the development
and progression of atherosclerosis and congestive heart
failure (CHF) (1). The inflammatory process contributes to
the formation of early atherosclerotic plaques in the form of
lipid-laden macrophages and induces plaque weakening and
rupture leading to acute coronary syndromes and sudden
death (2,3). Furthermore, many circulating markers of
inflammation, particularly C-reactive protein (CRP), are
associated with increased morbidity and mortality in asymptomatic individuals and in patients with cardiovascular
disease and CHF (4 –9).
Coronary artery disease (CAD) is the major cause of left
ventricular (LV) dysfunction and CHF (10). In this regard,
traditional risk factors for CAD as well as CRP are
predictors of CHF. Indeed, in the Health ABC (Health,
Aging, and Body Composition) study, among all major
*From the Division of Cardiology, Johns Hopkins Medical Institutions, Baltimore, Maryland; †Departments of Medicine and Pathology, University of
Vermont, Burlington, Vermont; ‡Radiology Department, Johns Hopkins Medical
Institutions, Baltimore, Maryland; and §Department of Epidemiology,
Bloomberg School of Public Health and Hygiene, Johns Hopkins University,
Baltimore, Maryland. This study was supported by the National Heart, Lung, and
Blood Institute grant (RO1-HL66075-01) and the Multi-Ethnic Study of Athero-
sclerosis study contracts (NO1-HC-95162, NO1-HC-95168, and NO1-HC95169). Dr. Lima is also supported by the Johns Hopkins Reynolds Center, and Dr.
Rosen is supported by the Israeli Heart Society and the Organization of American
Physician Fellowship for Medicine in Israel. Peter Libby, MD, acted as Guest Editor
for this article.
Manuscript received May 16, 2006; revised manuscript received September 8, 2006,
accepted September 28, 2006.
JACC Vol. 49, No. 5, 2007
February 6, 2007:594–600
cardiovascular events, CRP was most strongly associated
with future CHF (11).
Because CAD is regional, it is intuitive to expect that the
progressive LV dysfunctional process that underlies CHF
would commence as a regional process that eventually
results in global dysfunction and failure. In this regard,
although most studies relating inflammatory markers to the
development of CAD used clinical events as end points,
none have yet examined the relationship between systemic
markers of inflammation and regional LV dysfunction in
asymptomatic individuals.
The goal of this study was to investigate the association
between CRP and the presence and extent of regional LV
dysfunction in asymptomatic participants of the MESA
(Multi-Ethnic Study of Atherosclerosis) trial. Our hypothesis was that higher levels of CRP would be related to
regional LV dysfunction involving 1 or more coronary
territories independently of traditional risk factors.
Methods
The MESA trial is a prospective study designed to evaluate
mechanisms underlying the development and progression of
subclinical cardiovascular disease in asymptomatic individuals (12). There were 6,814 men and women, 45 to 85 years
of age, from 4 ethnicities (non-Hispanic white, African
American, Hispanic, and Chinese) enrolled at 6 field
centers (Winston Salem, North Carolina; New York, New
York; Baltimore, Maryland; Minneapolis, Minnesota; Chicago, Illinois; and Los Angeles, California). At entry, all
participants underwent extensive evaluation including clinical history, physical examination, and laboratory tests
including fasting glucose, lipid panel, and CRP. The
MESA protocol was approved by the institutional review
boards in all participating centers. Informed consent was
obtained from all study participants.
Tagged MRI studies. In an ancillary study of MESA,
1,184 participants randomly selected underwent tagged
magnetic resonance imaging (MRI) studies at enrollment
(from September 2001 to September 2002) in 6 centers.
Images were acquired by whole-body scanners (1.5 CVi,
General Electric Medical Systems, Waukesha, Wisconsin,
and Sonata/Symphony Siemens Medical Solutions, Erlangen, Germany) using electrocardiograph-triggered segmented k-space fast spoiled gradient-echo pulse sequence
during breath holds. After completing the standard protocol, 3 tagged short-axis slices (base to apex) were obtained.
Parallel striped tags were prescribed in 2 orthogonal
orientations (0° and 90°) using electrocardiograph-triggered
fast gradient echo sequence with spatial modulation of
magnetization.
Parameters for tagged images were: field of view 40 cm;
slice thickness 7 to 8 mm; repetition time 6 ms; echo time
3.0 ms; flip angle 10° to 12°; phase encoding views 128 with
6 phase encoding views per segment; temporal resolution 40
ms; tag spacing 7 mm.
Rosen et al.
Association Between CRP and LV Function
595
The LV mass was determined
Abbreviations
and Acronyms
for each participant using dedicated commercially available
CAD ⴝ coronary artery
software (MASS, version 4.2,
disease
Medis, Leiden, the Netherlands)
CHF ⴝ congestive heart
at end diastole.
failure
Strain analysis. Short-axis
CI ⴝ confidence interval
tagged slices were analyzed using
CRP ⴝ C-reactive protein
HARP (harmonic phase). The
Ecc ⴝ circumferential strain
HARP program (Diagnosoft,
IQR ⴝ interquartile range
Palo Alto, California) enables a
LAD ⴝ left anterior
fast determination of myocardial
descending artery
strain (13–15). In the present
LCX ⴝ left circumflex artery
study, peak systolic midwall cirLV ⴝ left ventricular
cumferential strain (Ecc) was deMESA
ⴝ Multi-Ethnic Study
termined in 12 segments in 3
of
Atherosclerosis
slices. By convention, systolic
RCA ⴝ right coronary artery
Ecc is normally negative because
of circumferential shortening, and
reduced (absolute) Ecc values reflect decreased regional function (for example, Ecc ⫽ ⫺12%
reflects lower regional function when compared with Ecc ⫽
⫺18%). Regional strains were analyzed according to coronary
perfusion areas (left anterior descending artery [LAD], left
circumflex artery [LCX], and right coronary artery [RCA]
territories). Assignment to coronary territories was done according to published standards (16). Peak global systolic strain
was defined as the peak midwall Ecc averaged across all
midwall segments.
Risk factors and CRP concentrations. Hypertension was
defined as diastolic blood pressure ⱖ90 mm Hg, systolic
blood pressure ⱖ140 mm Hg, or receiving treatment for
hypertension. High-normal diastolic blood pressure was
defined as ⬍90 and ⱖ85 mm Hg. Dyslipidemia was defined
as total cholesterol ⱖ240 mg/dl, low-density lipoprotein
cholesterol ⱖ160 mg/dl, triglycerides ⱖ150 mg/dl, highdensity lipoprotein cholesterol ⬍45 mg/dl, or receiving
treatment for hyperlipidemia. Diabetic individuals were defined as either having fasting plasma glucose ⱖ126 mg/dl or
receiving treatment for diabetes. Impaired fasting glucose was
defined as glucose ⱖ100 and ⬍126 mg/dl. Smoking status was
defined as current smoking, former smoking, or never smoked.
C-reactive protein was measured using the BNII nephelometer (N High-Sensitivity CRP; Dade Behring, Deerfield, Illinois). Analytical intra-assay coefficients of variation
(CVs) range from 2.3% to 4.4%, and interassay CVs range
from 2.1% to 5.7%.
Statistical analysis. The distribution of CRP was skewed;
therefore, log-transformation was performed. Data for CRP
are presented as median and interquartile range (IQR). All
analyses were done using STATA-8 software (Stata Corp.,
College Station, Texas). The CRP levels in different subgroups were compared using the Mann-Whitney U and
Kruskal-Wallis tests.
Linear regression was used to study the relationship
between log-CRP and regional Ecc. To test whether the
596
Rosen et al.
Association Between CRP and LV Function
JACC Vol. 49, No. 5, 2007
February 6, 2007:594–600
relationship between regional LV function and CRP differed by gender, age, ethnicity, or central obesity indexed by
waist circumference, interaction terms between CRP and
these factors were included in the initial multivariable
regression models. Only the gender ⫻ log-CRP interaction
was significant in most of the regions (p values for interactions were 0.1, 0.07, 0.03, and 0.04 for Ecc in the LAD,
LCX, RCA, and peak global Ecc, respectively). Therefore,
men and women were analyzed separately.
Multivariable linear regression was used to study the
independence of the association of log-CRP with regional
Ecc. Variables included as potential confounders were age,
ethnicity, waist circumference, hypertension, total cholesterol, treatment for hypertension or high cholesterol, diabetes mellitus, and smoking status. In women, hormone
replacement therapy use was also included (current use vs.
former and never use). Regression was performed in 3 steps.
Model I (demographic parameters) included age, ethnicity,
and waist circumference. Model II (risk factors): history of
hypertension, antihypertensive medication, cholesterol, antihyperlipidemic medication, smoking status, and history of
diabetes mellitus in addition to the variables included in
Model I. In women, adjustment for hormone replacement
therapy status (current users vs. nonusers) was added. Model
III included demographic parameters (I), risk factors (II),
and LV mass or coronary artery calcium scores in the
corresponding arteries (e.g., coronary artery calcium scores
in the LAD for Ecc LAD). Statistical significance was
defined as p ⬍ 0.05.
Results
Risk factor profiles of the study participants are shown in
Table 1. A total of 1,184 individuals underwent tagged
Risk Factor Profile of the Study Participants
Table 1
Risk Factor Profile of the Study Participants
Men
Women
Gender Difference
(p Value*)
Characteristics
Number
635 (54)
549 (46)
191 (30)
173 (30)
Ethnicity
Caucasian
Asian
70 (11)
65 (12)
African American
169 (27)
153 (28)
Hispanic
205 (32)
158 (29)
0.6
66 ⫾ 10
67 ⫾ 9
0.6
27.2 ⫾ 3.9
28.3 ⫾ 5.2
⬍0.001
SBP (mm Hg)
128 ⫾ 19
130 ⫾ 23
0.1
DBP (mm Hg)
75 ⫾ 9
69 ⫾ 11
⬍0.001
Age (yrs)
Body mass index (kg/m2)
Risk factors
Blood pressure
Hypertensive
265 (41.7)
261 (47.5)
0.05
Treatment for HTN
208 (32.8)
190 (34.6)
0.5
Lipids
Total cholesterol (mg/dl)
186 ⫾ 33
202 ⫾ 36
LDL cholesterol (mg/dl)
116 ⫾ 30
120 ⫾ 30
0.02
HDL cholesterol (mg/dl)
46 ⫾ 12
56 ⫾ 15
⬍0.001
Triglycerides (mg/dl)
130 ⫾ 84
132 ⫾ 77
⬍0.001
0.6
Dyslipidemia
248 (39)
197 (36)
0.3
Treatment for hyperlipidemia
118 (19)
116 (21)
0.3
Never smoked
268 (42)
338 (62)
Former smoker
282 (45)
159 (29)
Current smoker
82 (13)
48 (9)
⬍0.001
⬍0.001
Smoking
Glucose
Plasma glucose (mg/dl)
110 ⫾ 31
103 ⫾ 21
Diabetes mellitus
76 (12)
58 (11)
CRP (mg/l)†
1.3 (0.7–3.0)
2.6 (1.2–5.7)
⬍0.001
LV mass (g)
166 ⫾ 39
122 ⫾ 27
⬍0.001
0.4
Percentages are shown in parentheses. Continuous variables are presented as mean ⫾ standard deviation. *The p values were calculated by t test
for continuous variables and chi-square for categorical variables; †because of the skewed distribution of CRP, median values and interquartile
ranges are presented. Mann-Whitney U test was used to compare CRP levels. To convert values for total, HDL, and LDL cholesterol to millimoles per
liter, multiply by 0.0259; to convert values for glucose to millimoles per liter, multiply by 0.0555; to convert values for triglycerides to millimoles per
liter, multiply by 0.0113.
CRP ⫽ C-reactive protein; DBP ⫽ diastolic blood pressure; HDL ⫽ high-density lipoprotein; HTN ⫽ hypertension; LDL ⫽ low-density lipoprotein;
LV mass ⫽ left ventricular mass; SBP ⫽ systolic blood pressure.
Rosen et al.
Association Between CRP and LV Function
JACC Vol. 49, No. 5, 2007
February 6, 2007:594–600
MRI, of whom there were 1,164 (98.3%) participants with
available CRP data. Their mean age was 66.4 ⫾ 10 years.
Compared with the overall MESA trial cohort (total 6,814
participants), the studied subgroup included a higher percentage of men (54% vs. 47%, p ⬍ 0.001), a greater number
of Caucasians, and a smaller number of Hispanic individuals
(p ⬍ 0.001 for both ethnic groups). However, the distributions of risk factors and CRP concentrations did not differ
from the overall MESA trial cohort. Men had a higher
diastolic blood pressure and LV mass compared with
women, and were more likely to be current or former
smokers. Women had a higher body mass index and higher
cholesterol levels (total, low-density lipoprotein, and highdensity lipoprotein cholesterol) than men.
CRP and participant characteristics. Median CRP concentrations in men were 1.3 mg/l (IQR 0.7 to 3.0 mg/l), and
in women 2.6 mg/l (IQR 1.2 to 5.7 mg/l, p ⬍ 0.001).
Among women but not men, CRP was significantly lower
among those receiving lipid-lowering therapy compared
with nonusers (2.1 mg/l [IQR 0.6 to 3.0 mg/l] vs. 2.7 mg/l
[IQR 1.3 to 5.9 mg/l], p ⫽ 0.02). A total of 27% of the
women included in the cohort were current users of hormone replacement therapy. Women who were currently
using hormone replacement therapy had significantly higher
CRP levels compared with former or never-treated women
(3.5 mg/l [IQR 1.8 to 6.7 mg/l] vs. 2.3 mg/l [IQR 1.1 to 4.9
mg/l], p ⫽ 0.0004). In both genders, current cigarette
smokers had higher CRP levels than former smokers or
persons who had never smoked (p ⫽ 0.001).
Finally, there was a positive relationship between LV
mass and log-CRP in women (regression coefficient 4.5 g
LV mass per 1 unit higher log-CRP [mg/l], 95% confidence
interval [CI] 2.4 to 6.5, p ⬍ 0.001) and in men (regression
coefficient 5.0 g LV mass per 1 unit higher log-CRP [mg/l],
95% CI 2.0 to 8.0, p ⫽ 0.001). This relationship remained
significant after multivariate adjustment for demographic
characteristics and risk factors.
Relationships between log-CRP and regional LV function. Because of a significant gender interaction, results are
shown separately for men and women. The unadjusted
associations of CRP and regional LV function for men and
women are shown in Figure 1. In addition, these relationships are shown as regression coefficients in Table 2. A
positive regression coefficient indicates an association between higher log-CRP and lower absolute Ecc values, i.e.,
reduced regional LV function. Higher log-CRP was associated with lower regional LV function in all regions in men
but not in women (Fig. 1, Table 2).
In men, after adjustment for age, ethnicity, and traditional risk factors, the relationship between log-CRP and
lower regional LV function was attenuated slightly for
different regions (Table 3, models I and II). Similar results
were seen when waist circumference was substituted by body
mass index in the regression model. However, after further
adjustment for LV mass, the relationship between log-CRP
and Ecc weakened and remained statistically significant only
Figure 1
597
Peak Regional Systolic Ecc (%) by
CRP Quartiles (mg/l) in Men and Women
(A) Mean ⫹ SE error bars of peak systolic circumferential strain (Ecc) in the
left anterior descending artery (LAD), left circumflex artery (LCX), and right coronary artery (RCA) regions are shown. Open bars, dotted bars, ruled bars, and
solid bars indicate 1st, 2nd, 3rd, and 4th quartiles of C-reactive protein (CRP)
levels, respectively. (B) Simple linear regression was used to determine the
trend. The p values indicate significance levels. By convention, systolic Ecc is
normally negative because of circumferential shortening. Therefore, lower
(absolute) values of Ecc reflect decreased regional function.
in the LAD territory (Table 3, model III). In contrast, the
results were not markedly changed after adjusting for
coronary calcium scores in the corresponding regions (data
not shown).
In women, in addition to the other covariates, hormone
replacement therapy use was added to the regression model.
However, no associations between log-CRP and regional
Ecc values were observed (Table 3). Results remained
unchanged after including LV mass in the regression model.
Discussion
The main finding of this study is that higher CRP concentration is related to lower regional LV function determined
by Ecc in men, but not in women without evident heart
failure or CAD in the MESA trial. This association was
only modestly attenuated with adjustment for demographic
parameters and risk factors. Additional adjustment for LV
mass measured by MRI weakened this relationship.
Rosen et al.
Association Between CRP and LV Function
598
JACC Vol. 49, No. 5, 2007
February 6, 2007:594–600
Relationship Between Regional Ecc and Log-CRP in Men and Women
Table 2
Relationship Between Regional Ecc and Log-CRP in Men and Women
Men
Region
Women
Regression Coefficient
95% CI
p Value
Regression Coefficient
LAD
0.45
0.17–0.72
0.002
0.09
⫺0.27–0.44
95% CI
p Value
0.6
LCX
0.39
0.08–0.71
0.014
⫺0.08
⫺0.5–0.34
0.7
RCA
0.35
0.06–0.63
0.016
⫺0.13
⫺0.47–0.21
0.5
Global Ecc
0.38
0.13–0.63
0.003
⫺0.06
⫺0.40–0.27
0.7
Relationship is expressed as a regression coefficient (unadjusted). Regression coefficients are the differences in Ecc (%) per 1 U higher log-CRP
(mg/l). A positive value indicates an association of higher CRP with decreased LV function. The p values indicate the significance levels of the
regression coefficients for changes in regional strain for each gender.
CI ⫽ confidence interval; CRP ⫽ C-reactive protein; Ecc ⫽ circumferential strain; LAD ⫽ left anterior descending artery; LCX ⫽ left circumflex
artery; LV ⫽ left ventricular; RCA ⫽ right coronary artery.
Inflammatory markers including CRP are associated with
an increased risk for cardiovascular morbidity and mortality
in asymptomatic men and women, and in patients with
acute coronary syndromes (1,4 – 6,8,17,18). Furthermore, in
the Health ABC study, among major cardiovascular events,
CRP was most closely correlated with risk of future CHF (11).
Our findings document the association between CRP and LV
dysfunction among subjects without clinical heart failure or
coronary disease. In contrast to our study, gender interactions
were not observed in previous studies of clinical outcomes.
We hypothesize that the presence of subclinical atherosclerosis, endothelial dysfunction, and LV hypertrophy
might underlie the association between CRP and regional
LV dysfunction. These mechanisms are probably not mutually exclusive, and each may contribute substantially to the
relationship between CRP and regional LV dysfunction.
Inflammation plays an important role in the development
and progression of atherosclerosis. In the early stages of
atherogenesis, local inflammatory stimuli including oxidized
low-density lipoprotein lead to cytokine production
(1,2,19). These cytokines in turn induce endothelial adhesion and migration of monocytes into the lesion area,
producing early atherosclerotic lesions. The CRP seems to
play a role in this process by facilitating the activation of
inflammatory cells and enhancing cytokine production.
Furthermore, the inflammatory process is associated with
plaque weakening, rupture, and thrombus formation (1– 4).
Thus, it is logical to think that the relationship between
CRP and regional LV dysfunction may be related to the
association of CRP with these inflammatory processes
known to promote local atherosclerosis. In this regard, it is
important to mention that the relationship between CRP
and regional LV function was not modulated by coronary
calcium score.
This process may or not be associated with small regions
of myocardial infarction or scar tissue, even in the absence of
epicardial CAD, possibly because of local thrombosis or
embolism. These areas can be detected using delayed
contrast hyperenhancement (20,21). Indeed, scars indicative
of unrecognized myocardial infarctions have been detected
by delayed hyperenhancement in a considerable number
(20%) of elderly individuals participating in a communitybased study in Sweden (22). Unfortunately, this technique
has not been used in the MESA trial, and this plausible
explanation could not be studied.
Additionally, CRP may be related to incipient myocardial
dysfunction through its well-established association with
Relationship Demographic
Parameters;
Between Regional
Parameters
Ecc and
and
Log-CRP
Risk Factors;
in Men and
and Demographic
in Women After
Parameters,
Adjustment
Risk
for Factors,
Demographic
and LV Mass
Relationship Between Regional Ecc and Log-CRP in Men and in Women After Adjustment for Demographic
Table 3
Parameters; Demographic Parameters and Risk Factors; and Demographic Parameters, Risk Factors, and LV Mass
I
Region
Regression Coefficient
II
III
p Value
Regression Coefficient
p Value
Regression Coefficient
p Value
Men
LAD
0.44 (0.15–0.72)
0.002
0.37 (0.08–0.65)
0.01
0.31 (0.01–0.61)
0.04
LCX
0.37 (0.05–0.69)
0.022
0.32 (⫺0.01–0.64)
0.06
0.32 (⫺0.03–0.66)
0.07
RCA
0.34 (0.05–0.62)
0.020
0.31 (0.02–0.59)
0.04
0.18 (⫺0.12–0.49)
0.23
Global Ecc
0.37 (0.12–0.63)
0.004
0.32 (0.05–0.58)
0.02
0.26 (0.01–0.53)
0.06
Women
LAD
0.10 (⫺0.25–0.46)
0.57
0.16 (⫺0.44–0.76)
0.60
0.24 (⫺0.36–0.86)
0.42
LCX
⫺0.05 (⫺0.48–0.37)
0.80
⫺0.20 (⫺0.90–0.49)
0.57
⫺0.20 (⫺0.92–0.52)
0.59
RCA
⫺0.13 (⫺0.48–0.21)
0.45
⫺0.20 (⫺0.77–0.36)
0.49
⫺0.27 (⫺0.85–0.31)
0.36
Global Ecc
⫺0.04 (⫺0.38–0.30)
0.80
⫺0.09 (⫺0.65–0.48)
0.76
⫺0.06 (⫺0.64–0.52)
0.84
Regression coefficients are the differences in Ecc (%) per 1-U higher log-CRP (mg/l). A positive value indicates an association of higher CRP with LV dysfunction. The numbers in parentheses indicate 95%
confidence intervals. The p values indicate significance levels of the regression coefficients for changes in regional strains. Demographic parameters (model I) include age, ethnicity, and waist circumference.
Risk factors (model II) include history of hypertension, antihypertensive medication, cholesterol, antihyperlipidemic medication, smoking status, and history of diabetes mellitus. In women, adjustment for
hormone replacement therapy status (current users vs. nonusers) was also included. Model III includes demographic parameters (I), risk factors (II), and LV mass.
Abbreviations as in Table 2.
JACC Vol. 49, No. 5, 2007
February 6, 2007:594–600
endothelial dysfunction. The CRP may directly alter endothelial function by reducing local production of nitric oxide,
or indirectly by increasing levels of reactive oxygen species,
thus reducing endothelial induced vasodilation (2,23–26).
A third possible mechanism pertains to the association
between CRP and LV mass. We observed a significant
association between CRP and LV mass. This relationship
has been shown previously in patients with hypertension
and renal failure (27,28). It has been previously shown that
concentric LV hypertrophy is related to decreased myocardial function (29 –32). In the current study, after adjustment
for LV mass, the association of higher CRP concentrations
and reduced myocardial strain (Ecc) was attenuated, suggesting that the relationship between CRP and regional LV
function might be linked to the association between CRP
and LV mass.
An unexpected finding of this study was the gender
difference in the association of CRP and regional LV
dysfunction. In men but not women, higher CRP was
significantly associated with regional LV dysfunction. Substantial gender differences in the presence of risk factors,
shown in Table 1, may partly explain this difference, but the
gender difference remained after controlling for those variables. It is theoretically possible, but unlikely, that women
with elevated CRP and reduced regional LV function
developed symptoms earlier than men, and were therefore
excluded from the MESA trial. A third and more plausible
explanation is that gender-specific differences in the association between CRP and regional LV function are modulated through gender-related differences in LV mass. Previous studies have described gender-specific differences in
LV remodeling and the development of LV dysfunction and
CHF (33–38). The association between CRP levels and LV
mass and the gender-specific differences in the CRP/
regional dysfunction relationship fit this finding of gender
differences in the cardiac remodeling process. In contrast,
the Cardiovascular Health Study and the Health ABC
investigators reported a significant association between
higher CRP levels and the incidence of CHF in both
genders (11,39). These discrepancies suggest that in older
individuals, gender-related differences in the CRP LV
dysfunction relationships might disappear as women catch
up with men in the degree of atherosclerosis and inflammation.
However, this possibility should be further explored.
Finally, the finding of regional differences in the association between CRP and Ecc is noteworthy. The relationship between CRP and regional function in the LAD and
RCA seems to be stronger than in the LCX territory. This
may be attributable to statistical power, because the variation in the LCX has been found to be slightly greater than
in other regions, probably because of a decreased signal-tonoise ratio in the posterior wall. In addition, in a different
study, we have shown a greater coronary artery calcium score
and a higher extent of significant CAD in the LAD region
compared with other regions (unpublished data). Further
studies should be performed to clarify these findings.
Rosen et al.
Association Between CRP and LV Function
599
Methodological considerations. Our study includes 1,164
subjects, making this one of the largest MRI studies of
quantitative regional myocardial function that we are aware
of. The HARP software is a robust analytical tool that
allows for efficient and reliable determinations of regional
LV function by MRI tagging (13,15). Aside from its size,
the MESA trial entails a heterogeneous population from
different ethnicities.
On the other hand, the cross-sectional study design
impairs the ability to establish the temporal nature of the
association. Participants were selected to participate in the
MESA trial through a population-based process designed to
minimize biases typically associated with studies of volunteers. However, they may not represent a random sample of
the population. In addition, excluding symptomatic patients
could have affected the results of this study, as discussed
earlier.
Temporal resolution is an important determinant of the
quality of strain measurement. In the present study, the
mean temporal resolution was 40 ms. This temporal resolution is lower than the resolution that can be achieved by
echocardiography (i.e., 7 to 10 ms). However, MRI is not
limited by acoustic windows and scanning angles and can
provide a detailed strain map of the entire myocardium. In
addition, a previous study has shown a high correlation
between strain values measured by echocardiography and
tagged MRI (40).
Finally, except for a small study that measured only for
regional myocardial perfusion, contrast-enhanced MR images were not acquired routinely in the MESA trial.
Therefore, we could not address the possibility that areas
with reduced regional function may indeed reflect silent infracted regions manifesting as a delayed enhancement pattern.
In conclusion, we document an inverse relationship between CRP levels and regional LV function in men. This
association was seen in all regions and did not change after
adjusting for age, traditional risk factors, or medical treatment. In contrast, such a relationship was not evident in
women. The association between CRP and LV function
was attenuated by adjustment for LV mass. Further studies
are warranted to elucidate the interrelationship between
inflammation, LV mass, and gender in the development of
LV dysfunction and heart failure.
Acknowledgments
The authors thank the participants of the MESA trial and
the entire community of MESA investigators and staff for
their support and valuable contributions. A full list of
participating MESA investigators and institutions can be
found at http://www.mesa-nhlbi.org.
Reprint requests and correspondence: Dr. João A. C. Lima,
Cardiology Division, Blalock 524, Johns Hopkins Hospital, 600
North Wolfe Street, Baltimore, Maryland 21287-0409. E-mail:
[email protected].
600
Rosen et al.
Association Between CRP and LV Function
REFERENCES
1. Ross R. Atherosclerosis—an inflammatory disease. N Engl J Med
1999;340:115–26.
2. Willerson JT, Ridker PM. Inflammation as a cardiovascular risk factor.
Circulation 2004;109:II2–10.
3. Yeh ET. CRP as a mediator of disease. Circulation 2004;109:II11– 4.
4. Lind L. Circulating markers of inflammation and atherosclerosis.
Atherosclerosis 2003;169:203–14.
5. Ridker PM, Hennekens CH, Buring JE, Rifai N. C-reactive protein
and other markers of inflammation in the prediction of cardiovascular
disease in women. N Engl J Med 2000;342:836 – 43.
6. Ridker PM, Cushman M, Stampfer MJ, Tracy RP, Hennekens CH.
Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 1997;336:973–9.
7. Ridker PM. Intercellular adhesion molecule (ICAM-1) and the risks
of developing atherosclerotic disease. Eur Heart J 1998;19:1119 –21.
8. Ridker PM, Buring JE, Shih J, Matias M, Hennekens CH. Prospective study of C-reactive protein and the risk of future cardiovascular
events among apparently healthy women. Circulation 1998;98:731–3.
9. Anand IS, Latini R, Florea VG, et al. C-reactive protein in heart
failure: prognostic value and the effect of valsartan. Circulation
2005;112:1428 –34.
10. Packer M. Consensus recommendations for the management of
chronic heart failure. Am J Cardiol 1999;83:1A–38A.
11. Cesari M, Penninx BW, Newman AB, et al. Inflammatory markers
and onset of cardiovascular events: results from the Health ABC study.
Circulation 2003;108:2317–22.
12. Bild DE, Bluemke DA, Burke GL, et al. Multi-ethnic study of
atherosclerosis: objectives and design. Am J Epidemiol 2002;156:
871– 81.
13. Garot J, Bluemke DA, Osman NF, et al. Fast determination of
regional myocardial strain fields from tagged cardiac images using
harmonic phase MRI. Circulation 2000;101:981– 8.
14. O’Dell WG, McCulloch AD. Imaging three-dimensional cardiac
function. Annu Rev Biomed Eng 2000;2:431–56.
15. Osman NF, Prince JL. Visualizing myocardial function using HARP
MRI. Phys Med Biol 2000;45:1665– 82.
16. Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized
myocardial segmentation and nomenclature for tomographic imaging
of the heart: a statement for healthcare professionals from the Cardiac
Imaging Committee of the Council on Clinical Cardiology of the
American Heart Association. Circulation 2002;105:539 – 42.
17. Ridker PM. C-reactive protein and risks of future myocardial infarction and thrombotic stroke. Eur Heart J 1998;19:1–3.
18. Ridker PM. Inflammation, atherosclerosis, and cardiovascular risk:
an epidemiologic view. Blood Coagul Fibrinolysis 1999;10 Suppl 1:
S9 –12.
19. Verma S, Szmitko PE, Yeh ET. C-reactive protein: structure affects
function. Circulation 2004;109:1914 –7.
20. Lima JA, Judd RM, Bazille A, Schulman SP, Atalar E, Zerhouni EA.
Regional heterogeneity of human myocardial infarcts demonstrated by
contrast-enhanced MRI. Potential mechanisms. Circulation 1995;92:
1117–25.
21. Kim RJ, Fieno DS, Parrish TB, et al. Relationship of MRI delayed
contrast enhancement to irreversible injury, infarct age, and contractile
function. Circulation 1999;100:1992–2002.
22. Ebeling BC, Bjerner T, Johansson L, Lind L, Ahlstrom H. Myocardial scars more frequent than expected magnetic resonance imaging
detects potential risk group. J Am Coll Cardiol 2006;48:765–71.
23. Verma S, Kuliszewski MA, Li SH, et al. C-reactive protein attenuates
endothelial progenitor cell survival, differentiation, and function:
further evidence of a mechanistic link between C-reactive protein and
cardiovascular disease. Circulation 2004;109:2058 – 67.
JACC Vol. 49, No. 5, 2007
February 6, 2007:594–600
24. Pasceri V, Willerson JT, Yeh ET. Direct proinflammatory effect of
C-reactive protein on human endothelial cells. Circulation 2000;102:
2165– 8.
25. Landmesser U, Hornig B, Drexler H. Endothelial function: a critical
determinant in atherosclerosis? Circulation 2004;109:II27–33.
26. Harrison DG, Cai H. Endothelial control of vasomotion and nitric
oxide production. Cardiol Clin 2003;21:289 –302.
27. Yasunari K, Maeda K, Watanabe T, Nakamura M, Yoshikawa J,
Asada A. Comparative effects of valsartan versus amlodipine on left
ventricular mass and reactive oxygen species formation by monocytes
in hypertensive patients with left ventricular hypertrophy. J Am Coll
Cardiol 2004;43:2116 –23.
28. London GM, Marchais SJ, Guerin AP, Metivier F, Adda H, Pannier
B. Inflammation, arteriosclerosis, and cardiovascular therapy in hemodialysis patients. Kidney Int Suppl 2003;84:S88 –93.
29. Palmon LC, Reichek N, Yeon SB, et al. Intramural myocardial
shortening in hypertensive left ventricular hypertrophy with normal
pump function. Circulation 1994;89:122–31.
30. de Simone G, Devereux RB, Koren MJ, Mensah GA, Casale PN,
Laragh JH. Midwall left ventricular mechanics. An independent
predictor of cardiovascular risk in arterial hypertension. Circulation
1996;93:259 – 65.
31. Aurigemma GP, Silver KH, Priest MA, Gaasch WH. Geometric
changes allow normal ejection fraction despite depressed myocardial
shortening in hypertensive left ventricular hypertrophy. J Am Coll
Cardiol 1995;26:195–202.
32. Rosen BD, Edvardsen T, Lai S, et al. Left ventricular concentric
remodeling is associated with decreased global and regional systolic
function: the Multi-Ethnic Study of Atherosclerosis. Circulation
2005;112:984 –91.
33. Gardner JD, Brower GL, Janicki JS. Gender differences in cardiac
remodeling secondary to chronic volume overload. J Card Fail 2002;
8:101–7.
34. Weinberg EO, Thienelt CD, Katz SE, et al. Gender differences in
molecular remodeling in pressure overload hypertrophy. J Am Coll
Cardiol 1999;34:264 –73.
35. Tamura T, Said S, Gerdes AM. Gender-related differences in myocyte
remodeling in progression to heart failure. Hypertension 1999;33:
676 – 80.
36. Rosen BD, Edvardsen T, Lai S, et al. Left ventricular concentric
remodeling is associated with decreased global and regional systolic
function: the Multi-Ethnic Study of Atherosclerosis. Circulation
2005;112:984 –91.
37. Salton CJ, Chuang ML, O’Donnell CJ, et al. Gender differences and
normal left ventricular anatomy in an adult population free of hypertension. A cardiovascular magnetic resonance study of the Framingham Heart Study Offspring cohort. J Am Coll Cardiol 2002;39:
1055– 60.
38. Hees PS, Fleg JL, Lakatta EG, Shapiro EP. Left ventricular remodeling with age in normal men versus women: novel insights using
three-dimensional magnetic resonance imaging. Am J Cardiol 2002;
90:1231– 6.
39. Gottdiener JS, Arnold AM, Aurigemma GP, et al. Predictors of
congestive heart failure in the elderly: the Cardiovascular Health
Study. J Am Coll Cardiol 2000;35:1628 –37.
40. Edvardsen T, Gerber BL, Garot J, Bluemke DA, Lima JA, Smiseth
OA. Quantitative assessment of intrinsic regional myocardial deformation by Doppler strain rate echocardiography in humans: validation
against three-dimensional tagged magnetic resonance imaging. Circulation 2002;106:50 – 6.
APPENDIX
For a list of multivariate regression models,
please see the online version of this article.