Background: C-reactive protein (CRP) has been associated with increased risk of coronary heart disease (CHD) and stroke, but much of the research had focused on middle-aged populations with, limited prospective, population-based, longitudinal data. In this study, we examined data from an elderly population and described the distribution of CRP concentrations and the prevalence of elevated CRP levels (>3 mg/l), examined the association between CRP levels and incidence of CHD or ischemic stroke, and assessed the potential interaction of CRP with sex or race on the incidence of CHD or ischemic stroke.Methods: Baseline CRP levels were measured in a cohort of 57 13 participants 265 years of age from the Cardiovascular Health Study (CHS) using a high-sensitivity assay. The cohort included 3859 (68%) subjects free of cardiovascular disease and 1104 (19%) with existing CVD. Data were collected from 1989-1990 or 1992- 1993 to June 30, 1997. SAS 9.10 software was used for analyses and statistical tests included t test, ANOVA, χ2 Kaplan-Meier method, Log-rank test, and Cox proportional hazards regression. Results: CRP distribution was highly skewed toward higher values, thus necessitating the use of the median and log transformation of the mean. For all participants, the median of CRP concentrations was 1.92 mg/l; the geometric mean was1.97 mg/l. Thirty percent of participants had CRP values >3 mg/l. Among subjects with prevalent CHD and those free of CHD at baseline the median CRP levels were 2.32 mg/l and 1.75 mg/l, respectively. The prevalence of elevated CRP levels was 36% in participants with baseline CHD and 26% in those free of CHD; it was higher in women than in men (32% vs. 27%, respectively), in blacks than in whites (42% vs. 28%, respectively), in subjects taking versus not taking cardiovascular medicines (35% vs. 22%, respectively). The mean CRP were similar among participants with and without initial statin uses (P = 0.3 155). For CHD participants, 37% of statin users and 36% of nonusers had elevated CRP levels. During 8 years of follow-up, 270 incident CHD events and 245 incident ischemic strokes occurred. Incidence rate of CHD and ischemic stroke was 10.7 and 9.7 per 1000 person-years, respectively. The relative risk (RR) of CHD and ischemic stroke for CRP >3 mg/l compared with P for sex-CRP interaction, 0.7638; P for race-CRP interaction, 0.4428). Similarly, no effect modification was observed by sex and race in the association of CRP with ischemic stroke (P for sex-CRP interaction, 0.1721 ; P for race-CRP interaction, 0.5486). Conclusions: CRP levels were higher among prevalent CHD subjects than among those without CHD. Women, blacks, and CV drug users had elevated CRP levels. Elevated CRP was associated with increased 8-year risk of CHD and ischemic stroke. Neither sex nor race modified the association between CRP and CHD or ischemic stroke. Future studies will be needed to explore new CRP thresholds for the elderly, and to examine if reduction of CRP levels using pharmacological agents reduces the risk of CHD or stroke.