Inflammatory Markers and the Risk of Coronary Heart Disease in Men and Women
http://www.100md.com
《新英格兰医药杂志》
ABSTRACT
Background Few studies have simultaneously investigated the role of soluble tumor necrosis factor (TNF-) receptors types 1 and 2 (sTNF-R1 and sTNF-R2), C-reactive protein, and interleukin-6 as predictors of cardiovascular events. The value of these inflammatory markers as independent predictors remains controversial.
Methods We examined plasma levels of sTNF-R1, sTNF-R2, interleukin-6, and C-reactive protein as markers of risk for coronary heart disease among women participating in the Nurses' Health Study and men participating in the Health Professionals Follow-up Study in nested case–control analyses. Among participants who provided a blood sample and who were free of cardiovascular disease at baseline, 239 women and 265 men had a nonfatal myocardial infarction or fatal coronary heart disease during eight years and six years of follow-up, respectively. Using risk-set sampling, we selected controls in a 2:1 ratio with matching for age, smoking status, and date of blood sampling.
Results After adjustment for matching factors, high levels of interleukin-6 and C-reactive protein were significantly related to an increased risk of coronary heart disease in both sexes, whereas high levels of soluble TNF- receptors were significant only among women. Further adjustment for lipid and nonlipid factors attenuated all associations; only C-reactive protein levels remained significant. The relative risk among all participants was 1.79 for those with C-reactive protein levels of at least 3.0 mg per liter, as compared with those with levels of less than 1.0 mg per liter (95 percent confidence interval, 1.27 to 2.51; P for trend <0.001). Additional adjustment for the presence or absence of diabetes and hypertension moderately attenuated the relative risk to 1.68 (95 percent confidence interval, 1.18 to 2.38; P for trend = 0.008).
Conclusions Elevated levels of inflammatory markers, particularly C-reactive protein, indicate an increased risk of coronary heart disease. Although plasma lipid levels were more strongly associated with an increased risk than were inflammatory markers, the level of C-reactive protein remained a significant contributor to the prediction of coronary heart disease.
Inflammation plays an essential role in the development of insulin resistance and type 2 diabetes mellitus, the initiation and progression of atherosclerotic lesions, and plaque disruption.1,2 Interleukin-6 and tumor necrosis factor (TNF-) are inflammatory cytokines and the main inducers of the secretion of C-reactive protein in the liver.3 C-reactive protein is a marker of low-grade inflammation, and recent studies suggest that this protein has a role in the pathogenesis of atherosclerotic lesions in humans.4 The effects of TNF- are mediated by two receptors, type 1 and type 2 (TNF-R1 and TNF-R2), which circulate in soluble forms (sTNF-R1 and sTNF-R2, respectively) and can be measured with greater sensitivity and reliability than can TNF- itself.5 The soluble receptors may attenuate the bioactivity of TNF- but may also serve as slow-release reservoirs and promote inflammation in the absence of free TNF ligand.6
Nonetheless, only a few studies have examined the relationship between levels of sTNF-R1, sTNF-R2, and interleukin-6 and the risk of coronary heart disease.7,8,9,10 The predictive value of C-reactive protein for screening and its causal relationship to coronary heart disease remain matters of controversy.11,12,13,14,15,16,17 We prospectively examined the association between inflammatory markers and the risk of coronary heart disease and the role of potential mediators among men and women in a nested case–control analysis.
Methods
Study Population
The Nurses' Health Study (NHS) and the Health Professionals Follow-up Study (HPFS) are prospective cohort investigations respectively involving 121,700 female U.S. registered nurses who were 30 to 55 years old at baseline in 1976 and 51,529 U.S. male health professionals who were 40 to 75 years old at baseline in 1986. Information about health and disease is assessed biennially, and information about diet is obtained every four years by means of self-administered questionnaires.18,19 From 1989 through 1990, a blood sample was requested from all participants in the NHS, and 32,826 women provided one. Similarly, between 1993 and 1995, a blood sample was provided as requested by 18,225 men in the HPFS. Participants who provided blood samples were similar to those who did not, albeit the men who provided samples were somewhat younger than those who did not. In the NHS, among women without cardiovascular disease or cancer before 1990, we identified 249 women who had a nonfatal myocardial infarction or fatal coronary heart disease between the date of blood drawing and June 1998. In the HPFS, we identified 266 men who had a nonfatal myocardial infarction or fatal coronary heart disease between the date of blood drawing and the return of the 2000 questionnaire. Using risk-set sampling,20 we randomly selected controls in a 2:1 ratio who were matched for age, smoking status, and date of blood sampling from the subgroup of participants who were free of cardiovascular disease at the time coronary disease was diagnosed in the case patients. Within the NHS cohort, an additional matching criterion was fasting status at the time of blood sampling.
Assessment of Coronary Heart Disease
Study physicians who were unaware of the participant's exposure status confirmed the diagnosis of myocardial infarction on the basis of the criteria of the World Health Organization (symptoms plus either diagnostic electrocardiographic changes or elevated levels of cardiac enzymes). Deaths were identified from state vital records and the National Death Index or reported by the participant's next of kin or the postal system. Fatal coronary heart disease was confirmed by an examination of hospital or autopsy records, by the listing of coronary heart disease as the cause of death on the death certificate, if coronary heart disease was the underlying and most plausible cause, and if evidence of previous coronary heart disease was available.
Assessment of Other Factors
Anthropometric, lifestyle, and dietary data were derived from the questionnaire administered in 1990 to women and 1994 to men, with missing information substituted from previous questionnaires. Body-mass index was calculated as the weight in kilograms divided by the square of the height in meters. Average nutrient intake was computed with the use of a semiquantitative food-frequency questionnaire. Physical activity was expressed in terms of metabolic equivalent (MET)–hours. The questionnaires and the validity and reproducibility of measurements have been described previously.18,21
Measurement of Biochemical Variables
Blood samples from women were collected in tubes treated with liquid sodium heparin, and those from men were collected in EDTA-treated tubes. The tubes were then placed on ice packs, stored in Styrofoam containers, returned to our laboratory by overnight courier, centrifuged, and divided into aliquots for storage in liquid-nitrogen freezers (–130°C or colder).
The levels of C-reactive protein were determined by means of a highly sensitive immunoturbidimetric assay with the use of reagents and calibrators from Denka Seiken; this assay has a day-to-day variability of 1 to 2 percent. Levels of sTNF-R1, sTNF-R2, and interleukin-6 were measured by means of enzyme-linked immunosorbent assays (R&D Systems), which have a day-to-day variability of 3.5 to 9.0 percent. Levels of inflammatory markers were largely unaffected by transport conditions and reproducible within subjects over time.22,23 Total, high-density lipoprotein (HDL), and directly obtained low-density lipoprotein (LDL) cholesterol and triglycerides were measured according to standard methods with the use of reagents from Roche Diagnostics and Genzyme. Study samples were sent to the laboratory for analysis in randomly ordered batches, and the laboratory personnel were unaware of a sample's case–control status.
The study protocol was approved by the institutional review board of the Brigham and Women's Hospital and the Human Subjects Committee Review Board of Harvard School of Public Health.
Exclusions
After the exclusion of participants with missing data on biomarker levels, our data sets consisted of 708 women (239 patients and 469 controls) and 794 men (265 patients and 529 controls). The assay for interleukin-6 required slightly more plasma than we originally reserved for this assay among women. Therefore, analyses involving interleukin-6 were restricted to the subgroup of 676 women for whom interleukin-6 levels were available.
Statistical Analysis
We analyzed the two cohorts separately. Inflammatory markers were divided into quintiles, from the lowest to highest levels, on the basis of the sex-specific distributions among the controls. With risk-set sampling, the odds ratio derived from the logistic regression directly estimates the hazard ratio and, thus, the relative risk.20 We analyzed the association between biomarker levels and the risk of coronary heart disease using both conditional and unconditional logistic regression, with adjustment for matching factors. Because both analyses provided essentially the same results, we present the results of unconditional logistic regression, which parallel the results in the subgroup analyses.
In our multivariable model, we further adjusted for parental history of coronary heart disease before the age of 60 years (yes vs. no), alcohol intake (nondrinker, 0.1 to 4.9 g per day, 5.0 to 14.9 g per day, 15.0 to 29.9 g per day, or at least 30.0 g per day), body-mass index (less than 20, 20 to 24, 25 to 29, 30 to 34, or 35 or more), physical activity (in quintiles from lowest to highest level), ratio of total to HDL cholesterol (in quintiles from lowest to highest ratio), and use of postmenopausal hormone therapy (yes vs. no — for women only). Finally, we also added a history of diabetes (yes vs. no) and hypertension (yes vs. no) at baseline to the model to assess the effect of these potential mediators. Baseline was defined as the year blood was drawn.
Correlation coefficients were calculated with the use of age-adjusted Spearman partial-correlation coefficients. To test for linear trend, we used the median levels of inflammatory markers in the control categories as a continuous variable. To pool the estimates of relative risk for men and women, we used the weighted average of estimates according to the random-effects model of DerSimonian and Laird.24
All P values are two-tailed, and P values below 0.05 were considered to indicate statistical significance. All analyses were performed with the use of SAS software, version 8.2 (SAS Institute).
Results
Baseline Characteristics
Women in whom coronary heart disease developed during follow-up had significantly higher baseline levels of sTNF-R1 and sTNF-R2 than did control women; however, the levels did not differ significantly between men in whom coronary heart disease developed during follow-up and men in the control group (Table 1). In the case of both men and women, patients had significantly higher baseline levels of interleukin-6 and C-reactive protein than controls.
Table 1. Baseline Characteristics of Women and Men in Whom Coronary Heart Disease Developed during Follow-up and Matched Controls.
The levels of sTNF-R1 and sTNF-R2 showed a high degree of correlation with each other (Table 2). The correlation with and between the other inflammatory markers was moderate and ranged from 0.27 for sTNF-R1 and C-reactive protein to 0.45 for interleukin-6 and C-reactive protein. The levels of inflammatory markers were moderately inversely associated with HDL cholesterol levels.
Table 2. Age-Adjusted Spearman Partial-Correlation Coefficients between Selected Cardiovascular Risk Factors among 469 Control Women and 529 Control Men.
Main Effects
After adjustment for matching factors, women in the highest quintile of each inflammatory marker, as compared with women in the lowest quintile, had a significantly increased risk of coronary heart disease — by a factor of 1.95 to 2.57 — with significant trends across quintiles (Table 3). After additional adjustment for the presence or absence of a parental history of coronary heart disease before the age of 60 years, alcohol intake, level of physical activity, the ratio of total to HDL cholesterol, body-mass index, and the use or nonuse of postmenopausal hormone therapy, these associations were attenuated and no longer significant, except for C-reactive protein (model 2 in Table 3). Additional adjustment for the presence or absence of diabetes and hypertension, which are potentially in the causal pathway, further reduced the association for all inflammatory markers.
Table 3. Relative Risks of Coronary Heart Disease during Follow-up, According to the Quintile of Plasma Levels of Inflammatory Markers at Baseline.
Among men, we did not find an association between the levels of soluble TNF- receptors and the risk of coronary heart disease (Table 3). Men in the highest quintile of interleukin-6 had a 57 percent increase in the risk of coronary heart disease, as compared with men in the lowest quintile, after adjustment for matching factors, although this association was not significant and was further attenuated after multivariable adjustment. However, we found a significant association between C-reactive protein levels and the risk of coronary heart disease. Multivariable adjustment and adjustment for the presence or absence of hypertension and diabetes moderately attenuated this relationship; after accounting for these variables, men in the highest quintile of C-reactive protein, as compared with those in the lowest quintile, had a relative risk of coronary heart disease of 2.55 (95 percent confidence interval, 1.40 to 4.65; P for trend = 0.02).
For comparison, in the final multivariable-adjusted model (including the presence or absence of diabetes and hypertension and C-reactive protein levels), the relative risk of coronary heart disease for the highest quintile of the ratio of total to HDL cholesterol, as compared with the lowest quintile, was 4.33 (95 percent confidence interval, 2.11 to 8.90; P for trend <0.001) in women and 3.29 (95 percent confidence interval, 1.84 to 5.90; P for trend <0.001) in men.
Subgroup Analyses
Overall, we found no significant interactions between various low and high cardiovascular risk groups and the association of biomarkers with the risk of coronary heart disease, although the association of C-reactive protein was generally stronger in low-risk subgroups. For example, in the multivariable-adjusted model (excluding the presence or absence of hypertension and diabetes), the relative risk in the highest as compared with the lowest quintile of C-reactive protein was 2.53 among women with a body-mass index of less than 25 (95 percent confidence interval, 1.04 to 6.18; P for trend = 0.02) and 6.25 among men with a body-mass index of less than 25 (95 percent confidence interval, 2.28 to 17.1; P for trend = 0.005). Similarly, among participants with LDL cholesterol levels of less than 130 mg per deciliter (3.4 mmol per liter), the corresponding relative risks were 3.54 (95 percent confidence interval, 1.19 to 10.5; P for trend = 0.01) for women and 2.52 (95 percent confidence interval, 1.09 to 5.83; P for trend = 0.04) for men. Among participants without hypertension, the corresponding relative risks were 1.87 (95 percent confidence interval, 0.77 to 4.56; P for trend = 0.02) for women and 3.01 (95 percent confidence interval, 1.41 to 6.44; P for trend = 0.02) for men.
Clinical Cutoff Points for C-Reactive Protein
We further categorized the study participants, on the basis of recently proposed cutoff points for C-reactive protein, as having low levels (less than 1.0 mg per liter), moderate levels (1.0 to 2.9 mg per liter), and high levels (at least 3.0 mg per liter).25 In these analyses, participants with high levels of C-reactive protein, as compared with those with low levels, had a relative risk of coronary heart disease of approximately 1.8 after adjustment for covariates (including body-mass index and lipid levels) (Table 4). When we pooled the risk estimates for men and women, the final multivariable-adjusted relative risk (including adjustment for the presence or absence of diabetes and hypertension) was 1.68 in the group with high levels of C-reactive protein, as compared with the group with low levels (95 percent confidence interval, 1.18 to 2.38; P for trend = 0.008) (Table 4). This is similar to the pooled estimate (relative risk, 1.48; 95 percent confidence interval, 1.08 to 2.04; P for trend = 0.03) after we controlled for covariates from the Framingham risk score,26 including age, presence or absence of hypertension and diabetes, ratio of total to HDL cholesterol, and smoking status.
Table 4. Relative Risks of Coronary Heart Disease during Follow-up According to the Baseline Level of C-Reactive Protein.
We found a gradient of risk of coronary heart disease within each increasing category of C-reactive protein and ratio of total to HDL cholesterol (Figure 1). This finding supports the hypothesis that the levels of C-reactive protein may predict risk beyond the information afforded by lipid levels. However, despite the independent associations, the gradient of risk associated with lipid levels was greater than that for C-reactive protein levels.
Figure 1. Multivariable-Adjusted Relative Risk of Coronary Heart Disease among Women (Panel A) and Men (Panel B), According to the Baseline Level of C-Reactive Protein (CRP) and the Quintile of the Ratio of Total to HDL Cholesterol.
Data on women are from the Nurses' Health Study and include eight years of follow-up, and data on men are from the Health Professionals Follow-up Study and include six years of follow-up. The model was adjusted for age, smoking status, date of blood sampling, presence or absence of a parental history of coronary heart disease before the age of 60 years, alcohol intake, level of physical activity, and body-mass index. Among women, the multivariable model was also adjusted for fasting status at the time of blood sampling and the use or nonuse of postmenopausal hormone therapy. In each panel, the subjects in quintile 1 who had a CRP level of less than 1.0 mg per liter served as the reference group.
Additional Analyses
When we stratified our analysis according to the time to an event in two-year intervals, the relative risk of coronary heart disease associated with C-reactive protein levels remained relatively stable over time (data not shown). When we repeated our main analyses after excluding participants with C-reactive protein levels of at least 10.0 mg per liter, we found essentially the same results. C-reactive protein levels may be affected by hormone therapy.10 However, results were similar when we used quintiles of C-reactive protein based on levels in women in the control group who reported never using hormones.
Discussion
In these two nested case–control studies, we found that high plasma levels of C-reactive protein were associated with an increased risk of coronary heart disease among women and men without previous cardiovascular disease. Elevated plasma levels of sTNF-R1 and sTNF-R2 were related to an increased risk among women, but not men. We found only a moderate suggestion of increased risk associated with elevated levels of interleukin-6. For all markers, associations were substantially attenuated and — with the exception of C-reactive protein — no longer significant after adjustment for cardiovascular risk factors, particularly body-mass index and the presence or absence of diabetes and hypertension. These findings are consistent with a role of these inflammatory markers in the elevated risk of cardiovascular events that is associated with type 2 diabetes and hypertension.
TNF- and interleukin-6 are the main inducers of hepatic production of acute-phase proteins, including C-reactive protein.3 These inflammatory markers are associated with biologic and environmental risk factors for cardiovascular events, including components of the metabolic syndrome (obesity, insulin resistance, diabetes, hypertension, and low HDL cholesterol levels), and lifestyle factors, such as smoking, abstinence from alcohol, and physical inactivity.27,28,29
Compelling evidence suggests that inflammation causally contributes to several precursors of cardiovascular disease. TNF- and interleukin-6 can cause insulin resistance in animal models, and plasma levels of C-reactive protein and interleukin-6 have been shown to predict type 2 diabetes in humans.30,31 The increased cytokine synthesis in obesity may promote insulin resistance and impaired glucose uptake, type 2 diabetes, and ultimately, coronary heart disease.30 In line with these hypotheses, we found that plasma levels of interleukin-6 and C-reactive protein, in particular, were related to the risk of coronary heart disease and that the risks were attenuated after adjustment for the presence or absence of diabetes and hypertension.
TNF- has a limited half-life and is difficult to measure in large-scale epidemiologic studies.5,6 In a nested case–control study, Ridker et al. reported a multivariable-adjusted relative risk of recurrent coronary events of 2.5 (95 percent confidence interval, 1.3 to 5.1) among men whose TNF- levels exceeded the 95th percentile, as compared with men with lower levels.32 Cesari et al. reported a relative risk of of coronary events of 1.79 (95 percent confidence interval, 1.18 to 2.71) among elderly participants without cardiovascular disease who had the highest of three levels of TNF-, as compared with those who had the lowest levels.8 The value of assessing circulating levels of TNF- is unknown, since such levels can be very low and unstable. The levels of soluble TNF- receptors may be more stable and may better reflect longer-term average circulating levels of TNF-, although data on the role of soluble TNF- receptors in coronary heart disease are scarce.7,33 It is unclear why we found a difference in risk between men and women associated with elevated levels of soluble TNF- receptors; however, others also have found differences between women and men with respect to lipids34 and in the overall prediction of risk.35 Similarly, mechanisms of insulin sensitivity, rather than inflammation, may contribute more to the risk of coronary heart disease in women than men.
Findings of an association between interleukin-6 levels and the risk of coronary heart disease have been inconsistent.8,10,36 In our study, this association was substantially reduced and no longer significant after multivariable adjustment.
C-reactive protein is the most extensively studied inflammatory marker in prospective settings. In an early meta-analysis of 11 prospective studies, the relative risk of coronary heart disease in subjects with the highest of three C-reactive protein levels, as compared with those with the lowest levels, was 2.0 (95 percent confidence interval, 1.6 to 2.5) among population-based studies.37 Eleven other prospective studies have since been published. In an updated meta-analysis, Danesh et al. reported an overall odds ratio of 1.58 (95 percent confidence interval, 1.48 to 1.68) among subjects with the highest of three levels of C-reactive protein, as compared with subjects with the lowest level.16 This risk estimate is similar to that in our comparisons of C-reactive protein levels of at least 3.0 mg per liter with those of less than 1.0 mg per liter. However, the degree of adjustment for traditional cardiovascular risk factors differed markedly among the studies included in the meta-analysis.
An important question is whether knowing the level of C-reactive protein adds materially to risk prediction. In the Women's Health Study, Ridker et al. reported that the level of C-reactive protein was a stronger predictor than the LDL cholesterol level and that it added to the information provided by the Framingham risk score.12,38 Comparing C-reactive protein levels of at least 3.0 mg per liter with those of less than 1.0 mg per liter, they reported a relative risk of 1.5 (95 percent confidence interval, 1.2 to 1.9) after adjustment for the Framingham risk score and the presence or absence of diabetes.38
In the Atherosclerosis Risk in Communities Study, Ballantyne et al. reported a relative risk of coronary heart disease of 1.72 (95 percent confidence interval, 1.24 to 2.39) among subjects with a C-reactive protein level of at least 3.0 mg per liter, as compared with subjects with a level of less than 1.0 mg per liter (adjusted for components of the Framingham risk score, including the presence or absence of diabetes).14 In the Monitoring Trends and Determinants in Cardiovascular Disease (MONICA) study, comparing C-reactive protein levels of at least 3.0 mg per liter with those of less than 1.0 mg per liter, Koenig et al. reported a hazard ratio of 2.21 (95 percent confidence interval, 1.49 to 3.27), adjusted for the Framingham risk score.13 In contrast, in the Rotterdam Study, measuring the level of C-reactive protein did not improve the prediction of coronary events beyond that afforded by the Framingham risk score, with an odds ratio of 1.2 (95 percent confidence interval, 0.6 to 2.2) among participants in the highest quartile of C-reactive protein, as compared with those in the lowest quartile.39
In our analysis, the pooled relative risk among men and women classified according to clinical cutoff points for the levels of C-reactive protein was 1.48 (95 percent confidence interval, 1.08 to 2.04; P for trend = 0.03) after we accounted for covariates in the Framingham risk score, including the presence or absence of diabetes. Our results are similar to those of Ridker et al.38 and Ballantyne et al.,14 as well as those of the recent meta-analysis by Danesh et al.,16 a fact that suggests that after adjustment for the Framingham risk score, the relative risk associated with a clinical cutoff point of at least 3.0 mg per liter, as compared with a cutoff of less than 1.0 mg per liter, is probably moderately less than previously suggested in the guidelines for the clinical assessment of inflammatory markers issued by the American Heart Association and the Centers for Disease Control and Prevention (relative risk, 1.5 vs. approximately 2.0).25 Nevertheless, our findings support the theory that the level of C-reactive protein provides an additional measure of the risk of coronary heart disease beyond that afforded by the Framingham risk score.
Our study has some limitations. As with any observational study design, there is the possibility of unmeasured confounding. However, we controlled for most known cardiovascular risk factors. Though we obtained only a single blood sample at baseline, previous studies have shown the levels of biomarkers to be relatively stable over time.22,23 Since the ranges of anthropometric variables in our cohorts were quite broad, the biologic relationships found should be widely generalizable. Though we excluded men and women with missing data on blood levels, generalizability should be minimally affected because the participants were similar to those who did not provide blood samples.
Although the Framingham risk score is a tool for estimating the 10-year risk of coronary heart disease among healthy subjects,26 it does not include other well-established risk factors, such as body-mass index, alcohol intake, level of physical activity, or the presence or absence of a parental history of coronary heart disease.40 Therefore, to examine the role of inflammatory markers in coronary heart disease, we used an etiologic approach in our main analyses, to take into account the pathophysiology of coronary heart disease and include the major cardiovascular risk factors, beyond those included in the Framingham risk score, for comparison.
Our questionnaires did not include questions on the use of hydroxymethylglutarylcoenzyme A reductase inhibitors (statins) because these drugs were not widely used at time of blood sampling. However, the reported use of cholesterol-lowering drugs was generally low in both cohorts.
In conclusion, our findings suggest that high levels of C-reactive protein are associated with an increased risk of coronary heart disease among men and women and that the level of C-reactive protein is a significant marker of the risk of coronary heart disease, even after careful multivariable adjustment. Though all other associations were attenuated after multivariable adjustment, high levels of sTNF-R1 and sTNF-R2 may be also associated with an increased risk and deserve further exploration in other populations. From a clinical standpoint, although the ratio of total to HDL cholesterol was more strongly associated with the risk of coronary heart disease than were the levels of inflammatory markers, the level of C-reactive protein was still a significant contributor to the prediction of coronary heart disease.
Supported by grants (HL35464, CA55075, AA11181, and HL34594) from the National Institutes of Health and by a grant from Merck Research Laboratories. Dr. Pischon is a Jetson Lincoln fellow, supported in part by an unrestricted gift from Mr. Lincoln. Ms. Pai is supported by an institutional training grant (HL07575) from the National Heart, Lung, and Blood Institute.
Dr. Cannuscio was an employee of Merck at the time the research was conducted. Dr. Manson is listed as a coinventor of a patent filed by Brigham and Women's Hospital related to inflammatory markers and diabetes mellitus. Dr. Rimm reports having received grant support from Merck.
We are indebted to Alan Paciorek, Helena Ellis, and Jeanne Sparrow for coordinating the collection of samples and for laboratory management, and to Lydia Liu for programming review.
Source Information
From the Departments of Epidemiology (J.K.P., T.P., J.E.M., S.E.H., K.J., G.C.C., M.J.S., E.B.R.) and Nutrition (T.P., M.J.S., E.B.R.), Harvard School of Public Health; the Division of Preventive Medicine (J.K.P., J.E.M.) and Channing Laboratory (T.P., J.M., S.E.H., G.C.C., M.J.S., E.B.R.), Department of Medicine, Brigham and Women's Hospital and Harvard Medical School; the Department of Oral Health Policy and Epidemiology, Harvard School of Dental Medicine (K.J.); the Department of Laboratory Medicine, Children's Hospital (N.R.); and the Department of Pathology, Harvard Medical School (N.R.) — all in Boston; the Department of Epidemiology, German Institute of Human Nutrition, Potsdam-Rehbruecke, Germany (T.P.); and Merck Research Laboratories, West Point, Pa. (C.C.C.).
Ms. Pai and Dr. Pischon contributed equally to this article.
Address reprint requests to Dr. Rimm at the Departments of Epidemiology and Nutrition, Harvard School of Public Health, 665 Huntington Ave., Boston, MA 02115, or at erimm@hsph.harvard.edu.
References
Libby P. Inflammation in atherosclerosis. Nature 2002;420:868-874.
Pradhan AD, Ridker PM. Do atherosclerosis and type 2 diabetes share a common inflammatory basis? Eur Heart J 2002;23:831-834.
Yudkin JS, Kumari M, Humphries SE, Mohamed-Ali V. Inflammation, obesity, stress and coronary heart disease: is interleukin-6 the link? Atherosclerosis 2000;148:209-214.
Blake GJ, Ridker PM. Inflammatory biomarkers and cardiovascular risk prediction. J Intern Med 2002;252:283-294.
Diez-Ruiz A, Tilz GP, Zangerle R, Baier-Bitterlich G, Wachter H, Fuchs D. Soluble receptors for tumour necrosis factor in clinical laboratory diagnosis. Eur J Haematol 1995;54:1-8.
Aderka D. The potential biological and clinical significance of the soluble tumor necrosis factor receptors. Cytokine Growth Factor Rev 1996;7:231-240.
Benjafield AV, Wang XL, Morris BJ. Tumor necrosis factor receptor 2 gene (TNFRSF1B) in genetic basis of coronary artery disease. J Mol Med 2001;79:109-115.
Cesari M, Penninx BW, Newman AB, et al. Inflammatory markers and onset of cardiovascular events: results from the Health ABC study. Circulation 2003;108:2317-2322.
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;342:836-843.
Pradhan AD, Manson JE, Rossouw JE, et al. Inflammatory biomarkers, hormone replacement therapy, and incident coronary heart disease: prospective analysis from the Women's Health Initiative observational study. JAMA 2002;288:980-987.
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;336:973-979.
Ridker PM, Rifai N, Rose L, Buring JE, Cook NR. Comparison of C-reactive protein and low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events. N Engl J Med 2002;347:1557-1565.
Koenig W, Lowel H, Baumert J, Meisinger C. C-reactive protein modulates risk prediction based on the Framingham Score: implications for future risk assessment: results from a large cohort study in southern Germany. Circulation 2004;109:1349-1353.
Ballantyne CM, Hoogeveen RC, Bang H, et al. Lipoprotein-associated phospholipase A2, high-sensitivity C-reactive protein, and risk for incident coronary heart disease in middle-aged men and women in the Atherosclerosis Risk in Communities (ARIC) study. Circulation 2004;109:837-842.
Packard CJ, O'Reilly DS, Caslake MJ, et al. Lipoprotein-associated phospholipase A2 as an independent predictor of coronary heart disease: West of Scotland Coronary Prevention Study Group. N Engl J Med 2000;343:1148-1155.
Danesh J, Wheeler JG, Hirschfield GM, et al. C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease. N Engl J Med 2004;350:1387-1397.
Hackam DG, Anand SS. Emerging risk factors for atherosclerotic vascular disease: a critical review of the evidence. JAMA 2003;290:932-940.
Rimm EB, Giovannucci EL, Stampfer MJ, Colditz GA, Litin LB, Willett WC. Reproducibility and validity of an expanded self-administered semiquantitative food frequency questionnaire among male health professionals. Am J Epidemiol 1992;135:1114-1126.
Colditz GA, Manson JE, Hankinson SE. The Nurses' Health Study: 20-year contribution to the understanding of health among women. J Womens Health 1997;6:49-62.
Prentice RL, Breslow NE. Retrospective studies and failure time models. Biometrika 1978;65:153-158.
Willett W, Lenart E. Reproducibility and validity of food-frequency questionnaires. In: Willett W, ed. Nutritional epidemiology. New York: Oxford University Press, 1998:101-47.
Pai JK, Curhan GC, Cannuscio CC, Rifai N, Ridker PM, Rimm EB. Stability of novel plasma markers associated with cardiovascular disease: processing within 36 hours of specimen collection. Clin Chem 2002;48:1781-1784.
Pischon T, Hankinson SE, Hotamisligil GS, Rifai N, Willett WC, Rimm EB. Habitual dietary intake of n-3 and n-6 fatty acids in relation to inflammatory markers among US men and women. Circulation 2003;108:155-160.
Takkouche B, Cadarso-Suarez C, Spiegelman D. Evaluation of old and new tests of heterogeneity in epidemiologic meta-analysis. Am J Epidemiol 1999;150:206-215.
Pearson TA, Mensah GA, Alexander RW, et al. Markers of inflammation and cardiovascular disease: application to clinical and public health practice: a statement for healthcare professionals from the Centers for Disease Control and Prevention and the American Heart Association. Circulation 2003;107:499-511.
Wilson PW, D'Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97:1837-1847.
Ridker PM, Buring JE, Cook NR, Rifai N. C-reactive protein, the metabolic syndrome, and risk of incident cardiovascular events: an 8-year follow-up of 14 719 initially healthy American women. Circulation 2003;107:391-397.
Pischon T, Hankinson SE, Hotamisligil GS, Rifai N, Rimm EB. Leisure-time physical activity and reduced plasma levels of obesity-related inflammatory markers. Obes Res 2003;11:1055-1064.
Imhof A, Froehlich M, Brenner H, Boeing H, Pepys MB, Koenig W. Effect of alcohol consumption on systemic markers of inflammation. Lancet 2001;357:763-767.
Pradhan AD, Manson JE, Rifai N, Buring JE, Ridker PM. C-reactive protein, interleukin 6, and risk of developing type 2 diabetes mellitus. JAMA 2001;286:327-334.
Hotamisligil GS. The role of TNFalpha and TNF receptors in obesity and insulin resistance. J Intern Med 1999;245:621-625.
Ridker PM, Rifai N, Pfeffer M, Sacks F, Lepage S, Braunwald E. Elevation of tumor necrosis factor-alpha and increased risk of recurrent coronary events after myocardial infarction. Circulation 2000;101:2149-2153.
Cesari M, Penninx BW, Newman AB, et al. Inflammatory markers and cardiovascular disease (The Health, Aging and Body Composition Study). Am J Cardiol 2003;92:522-528.
Richey Sharrett A, Coady SA, Folsom AR, Couper DJ, Heiss G. Smoking and diabetes differ in their associations with subclinical atherosclerosis and coronary heart disease -- the ARIC Study. Atherosclerosis 2004;172:143-149.
Jousilahti P, Vartiainen E, Tuomilehto J, Puska P. Sex, age, cardiovascular risk factors, and coronary heart disease: a prospective follow-up study of 14 786 middle-aged men and women in Finland. Circulation 1999;99:1165-1172.
Ridker PM, Rifai N, Stampfer MJ, Hennekens CH. Plasma concentration of interleukin-6 and the risk of future myocardial infarction among apparently healthy men. Circulation 2000;101:1767-1772.
Danesh J, Whincup P, Walker M, et al. Low grade inflammation and coronary heart disease: prospective study and updated meta-analyses. BMJ 2000;321:199-204.
Ridker PM, Cook N. Clinical usefulness of very high and very low levels of C-reactive protein across the full range of Framingham Risk Scores. Circulation 2004;109:1955-1959.
van der Meer IM, de Maat MP, Kiliaan AJ, van der Kuip DA, Hofman A, Witteman JC. The value of C-reactive protein in cardiovascular risk prediction: the Rotterdam Study. Arch Intern Med 2003;163:1323-1328.
Pearson TA, Blair SN, Daniels SR, et al. AHA guidelines for primary prevention of cardiovascular disease and stroke: 2002 update: Consensus Panel guide to comprehensive risk reduction for adult patients without coronary or other atherosclerotic vascular diseases. Circulation 2002;106:388-391.(Jennifer K. Pai, M.H.S., )
Background Few studies have simultaneously investigated the role of soluble tumor necrosis factor (TNF-) receptors types 1 and 2 (sTNF-R1 and sTNF-R2), C-reactive protein, and interleukin-6 as predictors of cardiovascular events. The value of these inflammatory markers as independent predictors remains controversial.
Methods We examined plasma levels of sTNF-R1, sTNF-R2, interleukin-6, and C-reactive protein as markers of risk for coronary heart disease among women participating in the Nurses' Health Study and men participating in the Health Professionals Follow-up Study in nested case–control analyses. Among participants who provided a blood sample and who were free of cardiovascular disease at baseline, 239 women and 265 men had a nonfatal myocardial infarction or fatal coronary heart disease during eight years and six years of follow-up, respectively. Using risk-set sampling, we selected controls in a 2:1 ratio with matching for age, smoking status, and date of blood sampling.
Results After adjustment for matching factors, high levels of interleukin-6 and C-reactive protein were significantly related to an increased risk of coronary heart disease in both sexes, whereas high levels of soluble TNF- receptors were significant only among women. Further adjustment for lipid and nonlipid factors attenuated all associations; only C-reactive protein levels remained significant. The relative risk among all participants was 1.79 for those with C-reactive protein levels of at least 3.0 mg per liter, as compared with those with levels of less than 1.0 mg per liter (95 percent confidence interval, 1.27 to 2.51; P for trend <0.001). Additional adjustment for the presence or absence of diabetes and hypertension moderately attenuated the relative risk to 1.68 (95 percent confidence interval, 1.18 to 2.38; P for trend = 0.008).
Conclusions Elevated levels of inflammatory markers, particularly C-reactive protein, indicate an increased risk of coronary heart disease. Although plasma lipid levels were more strongly associated with an increased risk than were inflammatory markers, the level of C-reactive protein remained a significant contributor to the prediction of coronary heart disease.
Inflammation plays an essential role in the development of insulin resistance and type 2 diabetes mellitus, the initiation and progression of atherosclerotic lesions, and plaque disruption.1,2 Interleukin-6 and tumor necrosis factor (TNF-) are inflammatory cytokines and the main inducers of the secretion of C-reactive protein in the liver.3 C-reactive protein is a marker of low-grade inflammation, and recent studies suggest that this protein has a role in the pathogenesis of atherosclerotic lesions in humans.4 The effects of TNF- are mediated by two receptors, type 1 and type 2 (TNF-R1 and TNF-R2), which circulate in soluble forms (sTNF-R1 and sTNF-R2, respectively) and can be measured with greater sensitivity and reliability than can TNF- itself.5 The soluble receptors may attenuate the bioactivity of TNF- but may also serve as slow-release reservoirs and promote inflammation in the absence of free TNF ligand.6
Nonetheless, only a few studies have examined the relationship between levels of sTNF-R1, sTNF-R2, and interleukin-6 and the risk of coronary heart disease.7,8,9,10 The predictive value of C-reactive protein for screening and its causal relationship to coronary heart disease remain matters of controversy.11,12,13,14,15,16,17 We prospectively examined the association between inflammatory markers and the risk of coronary heart disease and the role of potential mediators among men and women in a nested case–control analysis.
Methods
Study Population
The Nurses' Health Study (NHS) and the Health Professionals Follow-up Study (HPFS) are prospective cohort investigations respectively involving 121,700 female U.S. registered nurses who were 30 to 55 years old at baseline in 1976 and 51,529 U.S. male health professionals who were 40 to 75 years old at baseline in 1986. Information about health and disease is assessed biennially, and information about diet is obtained every four years by means of self-administered questionnaires.18,19 From 1989 through 1990, a blood sample was requested from all participants in the NHS, and 32,826 women provided one. Similarly, between 1993 and 1995, a blood sample was provided as requested by 18,225 men in the HPFS. Participants who provided blood samples were similar to those who did not, albeit the men who provided samples were somewhat younger than those who did not. In the NHS, among women without cardiovascular disease or cancer before 1990, we identified 249 women who had a nonfatal myocardial infarction or fatal coronary heart disease between the date of blood drawing and June 1998. In the HPFS, we identified 266 men who had a nonfatal myocardial infarction or fatal coronary heart disease between the date of blood drawing and the return of the 2000 questionnaire. Using risk-set sampling,20 we randomly selected controls in a 2:1 ratio who were matched for age, smoking status, and date of blood sampling from the subgroup of participants who were free of cardiovascular disease at the time coronary disease was diagnosed in the case patients. Within the NHS cohort, an additional matching criterion was fasting status at the time of blood sampling.
Assessment of Coronary Heart Disease
Study physicians who were unaware of the participant's exposure status confirmed the diagnosis of myocardial infarction on the basis of the criteria of the World Health Organization (symptoms plus either diagnostic electrocardiographic changes or elevated levels of cardiac enzymes). Deaths were identified from state vital records and the National Death Index or reported by the participant's next of kin or the postal system. Fatal coronary heart disease was confirmed by an examination of hospital or autopsy records, by the listing of coronary heart disease as the cause of death on the death certificate, if coronary heart disease was the underlying and most plausible cause, and if evidence of previous coronary heart disease was available.
Assessment of Other Factors
Anthropometric, lifestyle, and dietary data were derived from the questionnaire administered in 1990 to women and 1994 to men, with missing information substituted from previous questionnaires. Body-mass index was calculated as the weight in kilograms divided by the square of the height in meters. Average nutrient intake was computed with the use of a semiquantitative food-frequency questionnaire. Physical activity was expressed in terms of metabolic equivalent (MET)–hours. The questionnaires and the validity and reproducibility of measurements have been described previously.18,21
Measurement of Biochemical Variables
Blood samples from women were collected in tubes treated with liquid sodium heparin, and those from men were collected in EDTA-treated tubes. The tubes were then placed on ice packs, stored in Styrofoam containers, returned to our laboratory by overnight courier, centrifuged, and divided into aliquots for storage in liquid-nitrogen freezers (–130°C or colder).
The levels of C-reactive protein were determined by means of a highly sensitive immunoturbidimetric assay with the use of reagents and calibrators from Denka Seiken; this assay has a day-to-day variability of 1 to 2 percent. Levels of sTNF-R1, sTNF-R2, and interleukin-6 were measured by means of enzyme-linked immunosorbent assays (R&D Systems), which have a day-to-day variability of 3.5 to 9.0 percent. Levels of inflammatory markers were largely unaffected by transport conditions and reproducible within subjects over time.22,23 Total, high-density lipoprotein (HDL), and directly obtained low-density lipoprotein (LDL) cholesterol and triglycerides were measured according to standard methods with the use of reagents from Roche Diagnostics and Genzyme. Study samples were sent to the laboratory for analysis in randomly ordered batches, and the laboratory personnel were unaware of a sample's case–control status.
The study protocol was approved by the institutional review board of the Brigham and Women's Hospital and the Human Subjects Committee Review Board of Harvard School of Public Health.
Exclusions
After the exclusion of participants with missing data on biomarker levels, our data sets consisted of 708 women (239 patients and 469 controls) and 794 men (265 patients and 529 controls). The assay for interleukin-6 required slightly more plasma than we originally reserved for this assay among women. Therefore, analyses involving interleukin-6 were restricted to the subgroup of 676 women for whom interleukin-6 levels were available.
Statistical Analysis
We analyzed the two cohorts separately. Inflammatory markers were divided into quintiles, from the lowest to highest levels, on the basis of the sex-specific distributions among the controls. With risk-set sampling, the odds ratio derived from the logistic regression directly estimates the hazard ratio and, thus, the relative risk.20 We analyzed the association between biomarker levels and the risk of coronary heart disease using both conditional and unconditional logistic regression, with adjustment for matching factors. Because both analyses provided essentially the same results, we present the results of unconditional logistic regression, which parallel the results in the subgroup analyses.
In our multivariable model, we further adjusted for parental history of coronary heart disease before the age of 60 years (yes vs. no), alcohol intake (nondrinker, 0.1 to 4.9 g per day, 5.0 to 14.9 g per day, 15.0 to 29.9 g per day, or at least 30.0 g per day), body-mass index (less than 20, 20 to 24, 25 to 29, 30 to 34, or 35 or more), physical activity (in quintiles from lowest to highest level), ratio of total to HDL cholesterol (in quintiles from lowest to highest ratio), and use of postmenopausal hormone therapy (yes vs. no — for women only). Finally, we also added a history of diabetes (yes vs. no) and hypertension (yes vs. no) at baseline to the model to assess the effect of these potential mediators. Baseline was defined as the year blood was drawn.
Correlation coefficients were calculated with the use of age-adjusted Spearman partial-correlation coefficients. To test for linear trend, we used the median levels of inflammatory markers in the control categories as a continuous variable. To pool the estimates of relative risk for men and women, we used the weighted average of estimates according to the random-effects model of DerSimonian and Laird.24
All P values are two-tailed, and P values below 0.05 were considered to indicate statistical significance. All analyses were performed with the use of SAS software, version 8.2 (SAS Institute).
Results
Baseline Characteristics
Women in whom coronary heart disease developed during follow-up had significantly higher baseline levels of sTNF-R1 and sTNF-R2 than did control women; however, the levels did not differ significantly between men in whom coronary heart disease developed during follow-up and men in the control group (Table 1). In the case of both men and women, patients had significantly higher baseline levels of interleukin-6 and C-reactive protein than controls.
Table 1. Baseline Characteristics of Women and Men in Whom Coronary Heart Disease Developed during Follow-up and Matched Controls.
The levels of sTNF-R1 and sTNF-R2 showed a high degree of correlation with each other (Table 2). The correlation with and between the other inflammatory markers was moderate and ranged from 0.27 for sTNF-R1 and C-reactive protein to 0.45 for interleukin-6 and C-reactive protein. The levels of inflammatory markers were moderately inversely associated with HDL cholesterol levels.
Table 2. Age-Adjusted Spearman Partial-Correlation Coefficients between Selected Cardiovascular Risk Factors among 469 Control Women and 529 Control Men.
Main Effects
After adjustment for matching factors, women in the highest quintile of each inflammatory marker, as compared with women in the lowest quintile, had a significantly increased risk of coronary heart disease — by a factor of 1.95 to 2.57 — with significant trends across quintiles (Table 3). After additional adjustment for the presence or absence of a parental history of coronary heart disease before the age of 60 years, alcohol intake, level of physical activity, the ratio of total to HDL cholesterol, body-mass index, and the use or nonuse of postmenopausal hormone therapy, these associations were attenuated and no longer significant, except for C-reactive protein (model 2 in Table 3). Additional adjustment for the presence or absence of diabetes and hypertension, which are potentially in the causal pathway, further reduced the association for all inflammatory markers.
Table 3. Relative Risks of Coronary Heart Disease during Follow-up, According to the Quintile of Plasma Levels of Inflammatory Markers at Baseline.
Among men, we did not find an association between the levels of soluble TNF- receptors and the risk of coronary heart disease (Table 3). Men in the highest quintile of interleukin-6 had a 57 percent increase in the risk of coronary heart disease, as compared with men in the lowest quintile, after adjustment for matching factors, although this association was not significant and was further attenuated after multivariable adjustment. However, we found a significant association between C-reactive protein levels and the risk of coronary heart disease. Multivariable adjustment and adjustment for the presence or absence of hypertension and diabetes moderately attenuated this relationship; after accounting for these variables, men in the highest quintile of C-reactive protein, as compared with those in the lowest quintile, had a relative risk of coronary heart disease of 2.55 (95 percent confidence interval, 1.40 to 4.65; P for trend = 0.02).
For comparison, in the final multivariable-adjusted model (including the presence or absence of diabetes and hypertension and C-reactive protein levels), the relative risk of coronary heart disease for the highest quintile of the ratio of total to HDL cholesterol, as compared with the lowest quintile, was 4.33 (95 percent confidence interval, 2.11 to 8.90; P for trend <0.001) in women and 3.29 (95 percent confidence interval, 1.84 to 5.90; P for trend <0.001) in men.
Subgroup Analyses
Overall, we found no significant interactions between various low and high cardiovascular risk groups and the association of biomarkers with the risk of coronary heart disease, although the association of C-reactive protein was generally stronger in low-risk subgroups. For example, in the multivariable-adjusted model (excluding the presence or absence of hypertension and diabetes), the relative risk in the highest as compared with the lowest quintile of C-reactive protein was 2.53 among women with a body-mass index of less than 25 (95 percent confidence interval, 1.04 to 6.18; P for trend = 0.02) and 6.25 among men with a body-mass index of less than 25 (95 percent confidence interval, 2.28 to 17.1; P for trend = 0.005). Similarly, among participants with LDL cholesterol levels of less than 130 mg per deciliter (3.4 mmol per liter), the corresponding relative risks were 3.54 (95 percent confidence interval, 1.19 to 10.5; P for trend = 0.01) for women and 2.52 (95 percent confidence interval, 1.09 to 5.83; P for trend = 0.04) for men. Among participants without hypertension, the corresponding relative risks were 1.87 (95 percent confidence interval, 0.77 to 4.56; P for trend = 0.02) for women and 3.01 (95 percent confidence interval, 1.41 to 6.44; P for trend = 0.02) for men.
Clinical Cutoff Points for C-Reactive Protein
We further categorized the study participants, on the basis of recently proposed cutoff points for C-reactive protein, as having low levels (less than 1.0 mg per liter), moderate levels (1.0 to 2.9 mg per liter), and high levels (at least 3.0 mg per liter).25 In these analyses, participants with high levels of C-reactive protein, as compared with those with low levels, had a relative risk of coronary heart disease of approximately 1.8 after adjustment for covariates (including body-mass index and lipid levels) (Table 4). When we pooled the risk estimates for men and women, the final multivariable-adjusted relative risk (including adjustment for the presence or absence of diabetes and hypertension) was 1.68 in the group with high levels of C-reactive protein, as compared with the group with low levels (95 percent confidence interval, 1.18 to 2.38; P for trend = 0.008) (Table 4). This is similar to the pooled estimate (relative risk, 1.48; 95 percent confidence interval, 1.08 to 2.04; P for trend = 0.03) after we controlled for covariates from the Framingham risk score,26 including age, presence or absence of hypertension and diabetes, ratio of total to HDL cholesterol, and smoking status.
Table 4. Relative Risks of Coronary Heart Disease during Follow-up According to the Baseline Level of C-Reactive Protein.
We found a gradient of risk of coronary heart disease within each increasing category of C-reactive protein and ratio of total to HDL cholesterol (Figure 1). This finding supports the hypothesis that the levels of C-reactive protein may predict risk beyond the information afforded by lipid levels. However, despite the independent associations, the gradient of risk associated with lipid levels was greater than that for C-reactive protein levels.
Figure 1. Multivariable-Adjusted Relative Risk of Coronary Heart Disease among Women (Panel A) and Men (Panel B), According to the Baseline Level of C-Reactive Protein (CRP) and the Quintile of the Ratio of Total to HDL Cholesterol.
Data on women are from the Nurses' Health Study and include eight years of follow-up, and data on men are from the Health Professionals Follow-up Study and include six years of follow-up. The model was adjusted for age, smoking status, date of blood sampling, presence or absence of a parental history of coronary heart disease before the age of 60 years, alcohol intake, level of physical activity, and body-mass index. Among women, the multivariable model was also adjusted for fasting status at the time of blood sampling and the use or nonuse of postmenopausal hormone therapy. In each panel, the subjects in quintile 1 who had a CRP level of less than 1.0 mg per liter served as the reference group.
Additional Analyses
When we stratified our analysis according to the time to an event in two-year intervals, the relative risk of coronary heart disease associated with C-reactive protein levels remained relatively stable over time (data not shown). When we repeated our main analyses after excluding participants with C-reactive protein levels of at least 10.0 mg per liter, we found essentially the same results. C-reactive protein levels may be affected by hormone therapy.10 However, results were similar when we used quintiles of C-reactive protein based on levels in women in the control group who reported never using hormones.
Discussion
In these two nested case–control studies, we found that high plasma levels of C-reactive protein were associated with an increased risk of coronary heart disease among women and men without previous cardiovascular disease. Elevated plasma levels of sTNF-R1 and sTNF-R2 were related to an increased risk among women, but not men. We found only a moderate suggestion of increased risk associated with elevated levels of interleukin-6. For all markers, associations were substantially attenuated and — with the exception of C-reactive protein — no longer significant after adjustment for cardiovascular risk factors, particularly body-mass index and the presence or absence of diabetes and hypertension. These findings are consistent with a role of these inflammatory markers in the elevated risk of cardiovascular events that is associated with type 2 diabetes and hypertension.
TNF- and interleukin-6 are the main inducers of hepatic production of acute-phase proteins, including C-reactive protein.3 These inflammatory markers are associated with biologic and environmental risk factors for cardiovascular events, including components of the metabolic syndrome (obesity, insulin resistance, diabetes, hypertension, and low HDL cholesterol levels), and lifestyle factors, such as smoking, abstinence from alcohol, and physical inactivity.27,28,29
Compelling evidence suggests that inflammation causally contributes to several precursors of cardiovascular disease. TNF- and interleukin-6 can cause insulin resistance in animal models, and plasma levels of C-reactive protein and interleukin-6 have been shown to predict type 2 diabetes in humans.30,31 The increased cytokine synthesis in obesity may promote insulin resistance and impaired glucose uptake, type 2 diabetes, and ultimately, coronary heart disease.30 In line with these hypotheses, we found that plasma levels of interleukin-6 and C-reactive protein, in particular, were related to the risk of coronary heart disease and that the risks were attenuated after adjustment for the presence or absence of diabetes and hypertension.
TNF- has a limited half-life and is difficult to measure in large-scale epidemiologic studies.5,6 In a nested case–control study, Ridker et al. reported a multivariable-adjusted relative risk of recurrent coronary events of 2.5 (95 percent confidence interval, 1.3 to 5.1) among men whose TNF- levels exceeded the 95th percentile, as compared with men with lower levels.32 Cesari et al. reported a relative risk of of coronary events of 1.79 (95 percent confidence interval, 1.18 to 2.71) among elderly participants without cardiovascular disease who had the highest of three levels of TNF-, as compared with those who had the lowest levels.8 The value of assessing circulating levels of TNF- is unknown, since such levels can be very low and unstable. The levels of soluble TNF- receptors may be more stable and may better reflect longer-term average circulating levels of TNF-, although data on the role of soluble TNF- receptors in coronary heart disease are scarce.7,33 It is unclear why we found a difference in risk between men and women associated with elevated levels of soluble TNF- receptors; however, others also have found differences between women and men with respect to lipids34 and in the overall prediction of risk.35 Similarly, mechanisms of insulin sensitivity, rather than inflammation, may contribute more to the risk of coronary heart disease in women than men.
Findings of an association between interleukin-6 levels and the risk of coronary heart disease have been inconsistent.8,10,36 In our study, this association was substantially reduced and no longer significant after multivariable adjustment.
C-reactive protein is the most extensively studied inflammatory marker in prospective settings. In an early meta-analysis of 11 prospective studies, the relative risk of coronary heart disease in subjects with the highest of three C-reactive protein levels, as compared with those with the lowest levels, was 2.0 (95 percent confidence interval, 1.6 to 2.5) among population-based studies.37 Eleven other prospective studies have since been published. In an updated meta-analysis, Danesh et al. reported an overall odds ratio of 1.58 (95 percent confidence interval, 1.48 to 1.68) among subjects with the highest of three levels of C-reactive protein, as compared with subjects with the lowest level.16 This risk estimate is similar to that in our comparisons of C-reactive protein levels of at least 3.0 mg per liter with those of less than 1.0 mg per liter. However, the degree of adjustment for traditional cardiovascular risk factors differed markedly among the studies included in the meta-analysis.
An important question is whether knowing the level of C-reactive protein adds materially to risk prediction. In the Women's Health Study, Ridker et al. reported that the level of C-reactive protein was a stronger predictor than the LDL cholesterol level and that it added to the information provided by the Framingham risk score.12,38 Comparing C-reactive protein levels of at least 3.0 mg per liter with those of less than 1.0 mg per liter, they reported a relative risk of 1.5 (95 percent confidence interval, 1.2 to 1.9) after adjustment for the Framingham risk score and the presence or absence of diabetes.38
In the Atherosclerosis Risk in Communities Study, Ballantyne et al. reported a relative risk of coronary heart disease of 1.72 (95 percent confidence interval, 1.24 to 2.39) among subjects with a C-reactive protein level of at least 3.0 mg per liter, as compared with subjects with a level of less than 1.0 mg per liter (adjusted for components of the Framingham risk score, including the presence or absence of diabetes).14 In the Monitoring Trends and Determinants in Cardiovascular Disease (MONICA) study, comparing C-reactive protein levels of at least 3.0 mg per liter with those of less than 1.0 mg per liter, Koenig et al. reported a hazard ratio of 2.21 (95 percent confidence interval, 1.49 to 3.27), adjusted for the Framingham risk score.13 In contrast, in the Rotterdam Study, measuring the level of C-reactive protein did not improve the prediction of coronary events beyond that afforded by the Framingham risk score, with an odds ratio of 1.2 (95 percent confidence interval, 0.6 to 2.2) among participants in the highest quartile of C-reactive protein, as compared with those in the lowest quartile.39
In our analysis, the pooled relative risk among men and women classified according to clinical cutoff points for the levels of C-reactive protein was 1.48 (95 percent confidence interval, 1.08 to 2.04; P for trend = 0.03) after we accounted for covariates in the Framingham risk score, including the presence or absence of diabetes. Our results are similar to those of Ridker et al.38 and Ballantyne et al.,14 as well as those of the recent meta-analysis by Danesh et al.,16 a fact that suggests that after adjustment for the Framingham risk score, the relative risk associated with a clinical cutoff point of at least 3.0 mg per liter, as compared with a cutoff of less than 1.0 mg per liter, is probably moderately less than previously suggested in the guidelines for the clinical assessment of inflammatory markers issued by the American Heart Association and the Centers for Disease Control and Prevention (relative risk, 1.5 vs. approximately 2.0).25 Nevertheless, our findings support the theory that the level of C-reactive protein provides an additional measure of the risk of coronary heart disease beyond that afforded by the Framingham risk score.
Our study has some limitations. As with any observational study design, there is the possibility of unmeasured confounding. However, we controlled for most known cardiovascular risk factors. Though we obtained only a single blood sample at baseline, previous studies have shown the levels of biomarkers to be relatively stable over time.22,23 Since the ranges of anthropometric variables in our cohorts were quite broad, the biologic relationships found should be widely generalizable. Though we excluded men and women with missing data on blood levels, generalizability should be minimally affected because the participants were similar to those who did not provide blood samples.
Although the Framingham risk score is a tool for estimating the 10-year risk of coronary heart disease among healthy subjects,26 it does not include other well-established risk factors, such as body-mass index, alcohol intake, level of physical activity, or the presence or absence of a parental history of coronary heart disease.40 Therefore, to examine the role of inflammatory markers in coronary heart disease, we used an etiologic approach in our main analyses, to take into account the pathophysiology of coronary heart disease and include the major cardiovascular risk factors, beyond those included in the Framingham risk score, for comparison.
Our questionnaires did not include questions on the use of hydroxymethylglutarylcoenzyme A reductase inhibitors (statins) because these drugs were not widely used at time of blood sampling. However, the reported use of cholesterol-lowering drugs was generally low in both cohorts.
In conclusion, our findings suggest that high levels of C-reactive protein are associated with an increased risk of coronary heart disease among men and women and that the level of C-reactive protein is a significant marker of the risk of coronary heart disease, even after careful multivariable adjustment. Though all other associations were attenuated after multivariable adjustment, high levels of sTNF-R1 and sTNF-R2 may be also associated with an increased risk and deserve further exploration in other populations. From a clinical standpoint, although the ratio of total to HDL cholesterol was more strongly associated with the risk of coronary heart disease than were the levels of inflammatory markers, the level of C-reactive protein was still a significant contributor to the prediction of coronary heart disease.
Supported by grants (HL35464, CA55075, AA11181, and HL34594) from the National Institutes of Health and by a grant from Merck Research Laboratories. Dr. Pischon is a Jetson Lincoln fellow, supported in part by an unrestricted gift from Mr. Lincoln. Ms. Pai is supported by an institutional training grant (HL07575) from the National Heart, Lung, and Blood Institute.
Dr. Cannuscio was an employee of Merck at the time the research was conducted. Dr. Manson is listed as a coinventor of a patent filed by Brigham and Women's Hospital related to inflammatory markers and diabetes mellitus. Dr. Rimm reports having received grant support from Merck.
We are indebted to Alan Paciorek, Helena Ellis, and Jeanne Sparrow for coordinating the collection of samples and for laboratory management, and to Lydia Liu for programming review.
Source Information
From the Departments of Epidemiology (J.K.P., T.P., J.E.M., S.E.H., K.J., G.C.C., M.J.S., E.B.R.) and Nutrition (T.P., M.J.S., E.B.R.), Harvard School of Public Health; the Division of Preventive Medicine (J.K.P., J.E.M.) and Channing Laboratory (T.P., J.M., S.E.H., G.C.C., M.J.S., E.B.R.), Department of Medicine, Brigham and Women's Hospital and Harvard Medical School; the Department of Oral Health Policy and Epidemiology, Harvard School of Dental Medicine (K.J.); the Department of Laboratory Medicine, Children's Hospital (N.R.); and the Department of Pathology, Harvard Medical School (N.R.) — all in Boston; the Department of Epidemiology, German Institute of Human Nutrition, Potsdam-Rehbruecke, Germany (T.P.); and Merck Research Laboratories, West Point, Pa. (C.C.C.).
Ms. Pai and Dr. Pischon contributed equally to this article.
Address reprint requests to Dr. Rimm at the Departments of Epidemiology and Nutrition, Harvard School of Public Health, 665 Huntington Ave., Boston, MA 02115, or at erimm@hsph.harvard.edu.
References
Libby P. Inflammation in atherosclerosis. Nature 2002;420:868-874.
Pradhan AD, Ridker PM. Do atherosclerosis and type 2 diabetes share a common inflammatory basis? Eur Heart J 2002;23:831-834.
Yudkin JS, Kumari M, Humphries SE, Mohamed-Ali V. Inflammation, obesity, stress and coronary heart disease: is interleukin-6 the link? Atherosclerosis 2000;148:209-214.
Blake GJ, Ridker PM. Inflammatory biomarkers and cardiovascular risk prediction. J Intern Med 2002;252:283-294.
Diez-Ruiz A, Tilz GP, Zangerle R, Baier-Bitterlich G, Wachter H, Fuchs D. Soluble receptors for tumour necrosis factor in clinical laboratory diagnosis. Eur J Haematol 1995;54:1-8.
Aderka D. The potential biological and clinical significance of the soluble tumor necrosis factor receptors. Cytokine Growth Factor Rev 1996;7:231-240.
Benjafield AV, Wang XL, Morris BJ. Tumor necrosis factor receptor 2 gene (TNFRSF1B) in genetic basis of coronary artery disease. J Mol Med 2001;79:109-115.
Cesari M, Penninx BW, Newman AB, et al. Inflammatory markers and onset of cardiovascular events: results from the Health ABC study. Circulation 2003;108:2317-2322.
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;342:836-843.
Pradhan AD, Manson JE, Rossouw JE, et al. Inflammatory biomarkers, hormone replacement therapy, and incident coronary heart disease: prospective analysis from the Women's Health Initiative observational study. JAMA 2002;288:980-987.
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;336:973-979.
Ridker PM, Rifai N, Rose L, Buring JE, Cook NR. Comparison of C-reactive protein and low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events. N Engl J Med 2002;347:1557-1565.
Koenig W, Lowel H, Baumert J, Meisinger C. C-reactive protein modulates risk prediction based on the Framingham Score: implications for future risk assessment: results from a large cohort study in southern Germany. Circulation 2004;109:1349-1353.
Ballantyne CM, Hoogeveen RC, Bang H, et al. Lipoprotein-associated phospholipase A2, high-sensitivity C-reactive protein, and risk for incident coronary heart disease in middle-aged men and women in the Atherosclerosis Risk in Communities (ARIC) study. Circulation 2004;109:837-842.
Packard CJ, O'Reilly DS, Caslake MJ, et al. Lipoprotein-associated phospholipase A2 as an independent predictor of coronary heart disease: West of Scotland Coronary Prevention Study Group. N Engl J Med 2000;343:1148-1155.
Danesh J, Wheeler JG, Hirschfield GM, et al. C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease. N Engl J Med 2004;350:1387-1397.
Hackam DG, Anand SS. Emerging risk factors for atherosclerotic vascular disease: a critical review of the evidence. JAMA 2003;290:932-940.
Rimm EB, Giovannucci EL, Stampfer MJ, Colditz GA, Litin LB, Willett WC. Reproducibility and validity of an expanded self-administered semiquantitative food frequency questionnaire among male health professionals. Am J Epidemiol 1992;135:1114-1126.
Colditz GA, Manson JE, Hankinson SE. The Nurses' Health Study: 20-year contribution to the understanding of health among women. J Womens Health 1997;6:49-62.
Prentice RL, Breslow NE. Retrospective studies and failure time models. Biometrika 1978;65:153-158.
Willett W, Lenart E. Reproducibility and validity of food-frequency questionnaires. In: Willett W, ed. Nutritional epidemiology. New York: Oxford University Press, 1998:101-47.
Pai JK, Curhan GC, Cannuscio CC, Rifai N, Ridker PM, Rimm EB. Stability of novel plasma markers associated with cardiovascular disease: processing within 36 hours of specimen collection. Clin Chem 2002;48:1781-1784.
Pischon T, Hankinson SE, Hotamisligil GS, Rifai N, Willett WC, Rimm EB. Habitual dietary intake of n-3 and n-6 fatty acids in relation to inflammatory markers among US men and women. Circulation 2003;108:155-160.
Takkouche B, Cadarso-Suarez C, Spiegelman D. Evaluation of old and new tests of heterogeneity in epidemiologic meta-analysis. Am J Epidemiol 1999;150:206-215.
Pearson TA, Mensah GA, Alexander RW, et al. Markers of inflammation and cardiovascular disease: application to clinical and public health practice: a statement for healthcare professionals from the Centers for Disease Control and Prevention and the American Heart Association. Circulation 2003;107:499-511.
Wilson PW, D'Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97:1837-1847.
Ridker PM, Buring JE, Cook NR, Rifai N. C-reactive protein, the metabolic syndrome, and risk of incident cardiovascular events: an 8-year follow-up of 14 719 initially healthy American women. Circulation 2003;107:391-397.
Pischon T, Hankinson SE, Hotamisligil GS, Rifai N, Rimm EB. Leisure-time physical activity and reduced plasma levels of obesity-related inflammatory markers. Obes Res 2003;11:1055-1064.
Imhof A, Froehlich M, Brenner H, Boeing H, Pepys MB, Koenig W. Effect of alcohol consumption on systemic markers of inflammation. Lancet 2001;357:763-767.
Pradhan AD, Manson JE, Rifai N, Buring JE, Ridker PM. C-reactive protein, interleukin 6, and risk of developing type 2 diabetes mellitus. JAMA 2001;286:327-334.
Hotamisligil GS. The role of TNFalpha and TNF receptors in obesity and insulin resistance. J Intern Med 1999;245:621-625.
Ridker PM, Rifai N, Pfeffer M, Sacks F, Lepage S, Braunwald E. Elevation of tumor necrosis factor-alpha and increased risk of recurrent coronary events after myocardial infarction. Circulation 2000;101:2149-2153.
Cesari M, Penninx BW, Newman AB, et al. Inflammatory markers and cardiovascular disease (The Health, Aging and Body Composition Study). Am J Cardiol 2003;92:522-528.
Richey Sharrett A, Coady SA, Folsom AR, Couper DJ, Heiss G. Smoking and diabetes differ in their associations with subclinical atherosclerosis and coronary heart disease -- the ARIC Study. Atherosclerosis 2004;172:143-149.
Jousilahti P, Vartiainen E, Tuomilehto J, Puska P. Sex, age, cardiovascular risk factors, and coronary heart disease: a prospective follow-up study of 14 786 middle-aged men and women in Finland. Circulation 1999;99:1165-1172.
Ridker PM, Rifai N, Stampfer MJ, Hennekens CH. Plasma concentration of interleukin-6 and the risk of future myocardial infarction among apparently healthy men. Circulation 2000;101:1767-1772.
Danesh J, Whincup P, Walker M, et al. Low grade inflammation and coronary heart disease: prospective study and updated meta-analyses. BMJ 2000;321:199-204.
Ridker PM, Cook N. Clinical usefulness of very high and very low levels of C-reactive protein across the full range of Framingham Risk Scores. Circulation 2004;109:1955-1959.
van der Meer IM, de Maat MP, Kiliaan AJ, van der Kuip DA, Hofman A, Witteman JC. The value of C-reactive protein in cardiovascular risk prediction: the Rotterdam Study. Arch Intern Med 2003;163:1323-1328.
Pearson TA, Blair SN, Daniels SR, et al. AHA guidelines for primary prevention of cardiovascular disease and stroke: 2002 update: Consensus Panel guide to comprehensive risk reduction for adult patients without coronary or other atherosclerotic vascular diseases. Circulation 2002;106:388-391.(Jennifer K. Pai, M.H.S., )