The study describes rates of coronary heart disease death and nonfatal coronary events over five years in a cohort of random sample population and relates them to levels of three major risk factors. It is based on a longitudinal follow-up of survey conducted initially in 1990 in all geographic regions of Turkey. Two-thirds of the original cohort aged 20 years or over 2259 adults comprising 1146 women was followed up by physical examination and an ECG recording at rest. New coronary events were defined to include myocardial infarction and stable angina with or without associated myocardial ischemia developed during the follow-up period. Overall annual death rate was nine per 1000 adults. Coronary deaths numbered 55 (of which 26 were women) representing 4.1 per year and were found high in women. New coronary events were registered in 37 men and 32 women (annual rates 7.2 and 5.8 per 1000, respectively). Among male participants aged initially 40 years or over, high systolic blood pressure (> or = 130 mmHg) at baseline significantly predicted coronary death (age-adjusted risk ratio (RR) 3.3) while high cholesterol concentrations (> or = 5.2 mmol l-1) predicted new coronary events alone (RR almost 2). In women systolic pressure again strongly predicted coronary death (age-adjusted RR 3.9), whereas abnormal cholesterol levels discriminated for coronary death and new coronary events (RR around 2.3 for each). High diastolic pressure (> or = 85 mmHg) was of predictive value for the combined outcome of coronary death and events in women (RR 1.9) but not in men. Multivariate analysis by logistic regression identified systolic blood pressure in men as significant independent predictor of coronary death, while total cholesterol concentration in both genders and systolic blood pressure in men were independent predictors of the combined outcome of coronary death or nonfatal coronary events. It was concluded that known major risk factors act in similar magnitudes commensurate with the specific risk increments also in populations with essentially low cholesterol levels. The relatively high coronary morbidity and mortality in Turkish women approaching that in men may be accounted for by an inherent greater risk burden.