Objectives: To assess the association between systemic C-reactive protein (CRP) and incident coronary heart disease (CHD) in community-dwelling elderly people.
Design: A French population-based multicenter prospective cohort study.
Setting: Three cities in France: Bordeaux in the southwest, Dijon in the northeast, and Montpellier in the southeast.
Participants: After 4 years of follow-up, a case-cohort study was designed including 1,004 subjects randomly selected from the initial cohort of 9,294 subjects free of CHD at baseline and 174 subjects who developed first CHD events during follow-up.
Measurements: Hazard ratios (HRs) were estimated using a Cox proportional hazard model adapted for the case-cohort design using a CRP level less than 1 mg/L as the reference category.
Results: Of the random sample, 24.3% had a CRP level less than 1.0 mg/L, 45.8% had a CRP level of 1.0 to 2.9 mg/L, and 29.9% had a CRP level of 3.0 to 10.0 mg/L. The HRs for CHD, adjusted for age, sex, and study center, were 1.69 (95% confidence interval (CI)=1.04-2.75) for CRP from 1.0 to 2.9 mg/L and 2.32 (95% CI=1.41-3.82) for CRP from 3.0 to 10.0 mg/L (P for trend <.001). After additional adjustment for smoking, body mass index, diabetes mellitus, systolic blood pressure, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglycerides, statin use, and antihypertensive treatment, a baseline CRP of 3.0 to 10.0 mg/L remained associated with risk of CHD (HR=1.87, 95% CI=1.09-3.25), although CRP did not improve the discriminative ability of a predicting model based on traditional risk factors (receiver operating characteristic curves from 0.740 to 0.749).
Conclusion: CRP is an independent CHD risk marker but does not improve CHD risk prediction in community-dwelling elderly people.