C-reactive protein (CRP) has been shown to predict cardiovascular disease. Whether predictions differ across risk factor strata and for short and long-term follow-up has not been clearly examined. The purpose of this report is to assess the relation between CRP and the development of myocardial infarction (MI) over a 20-year period in men in the Honolulu Heart Program. Subjects were aged 48 to 70 years and free of prevalent disease at the time when CRP levels were measured and follow-up began. Using a case-control design, 369 cases of MI were compared with 1,348 control subjects. After risk factor adjustment, the odds of an MI rose with increasing levels of CRP as early as 5 years into follow-up (P = 0.009). Associations appeared to persist beyond this time, but after 15 years, effects became modest. Adverse effects of an elevated CRP level were observed in middle-aged men (< or =55 years), in men without hypertension or diabetes, and in those who were nonsmokers (P < 0.05). Although positive effects were also observed in those who were hypertensive and smoking at the time of CRP measurement, findings suggest that in clinically healthy men, atherosclerosis could have origins more closely linked with inflammation than with other processes.