In the WHO sponsored trial of clofibrate and its follow-up, about 15,000 men were observed for a mean period of 13.2 years. As expected, incidence of heart attacks (HA) was directly related to serum cholesterol, blood pressure (BP) and cigarette smoking. The previously reported lower incidence of HA in men receiving clofibrate compared with controls was most noticeable in hypertensive heavy smokers (P less than 0.01). BP was slightly lower in smokers than non-smokers (P less than 0.01). The difference in BP was greater in the trial visit before HA. Smokers also had higher plasma fibrinogen levels (P less than 0.05). The combination of reduced diastolic BP, and therefore myocardial perfusion pressure, with an increased thrombogenic tendency, might explain the high incidence of HA in smokers. Clofibrate apparently reduced fibrinogen levels, which might account for its specially good effect in preventing HA in smokers. However, the ill effects of smoking are still evident at a lower level in the men taking clofibrate and the drug is no substitute for giving up the habit. Clofibrate is not recommended for widespread use, except in subjects with marked lipid and/or blood fibrinogen abnormalities after dietary measures have been tried.