Efficient use of cholesterol measurements to screen for coronary heart disease in the elderly is not well defined. The purpose of this report is to examine such screening based on national guidelines in a sample of older men. Since relations between cholesterol and coronary heart disease are better established in those who are younger, screening in the elderly will also consider levels of cholesterol that existed earlier in life. Data are from a prospective study of 1,170 men enrolled in the Honolulu Heart Program who were followed over a 12-year period for coronary heart disease. Follow-up began from 1980 to 1982, when cholesterol levels were determined in men who were aged 61 to 81 years. Past cholesterol levels were measured 10 years earlier (1970-1972). During the course of follow-up, coronary heart disease developed in 117 of the men. Risk of disease rose significantly (P = 0.003) with increases in past cholesterol levels (1970-1972) but not with more recent levels (1980-1982). For men with current cholesterol levels that were desirable (<5.2 mmol/L [200 mg/dl], as defined by guidelines from the National Cholesterol Education Program), disease incidence continued to rise with increasing past cholesterol levels (P<0.001). Accounting for high-density lipoprotein cholesterol and other screening factors did little to alter these findings. We conclude that desirable cholesterol levels in the elderly may not be a marker of a healthy risk profile if past cholesterol levels were high. Screening for coronary heart disease in the elderly could be improved by considering past cholesterol levels, rather than just a single measurement in later life.