Most epidemiological and intervention studies in patients with coronary artery disease have focused on men, the assumption being that such data can be extrapolated to women. However, there is little evidence to support this belief. We have completed a fifteen-year follow-up of 15,399 adults, including 8262 women, who lived in Renfrew and Paisley and were aged 45-64 years when screened between 1972 and 1976. We identified 490 deaths from coronary heart disease (CHD) in women and 878 in men. Women were more likely to have high cholesterol, to be obese, and to come from lower social classes than men, but they smoked less and had similar blood pressures. The relative risk--top to bottom quintile (95% Cl)--of cholesterol for coronary death after adjustment for all other risk markers was slightly greater in women (1.77 [1.45,2.16]) than in men (1.56 [1.32, 1.85]), but absolute and attributable risk were lower. Thus, women in the top quintile for cholesterol had lower coronary mortality (6.1 deaths per thousand patient years) than men in the bottom quintile (6.8 deaths per thousand patient years). Moreover, it was estimated that there would have been only 103 (21%) fewer CHD deaths in women, yet 211 (24%) fewer in men, if mortality had been the same for women and men in the lowest quintiles of cholesterol. Trends showing similar relative risks in these women, but lower absolute and attributable risks than in men, were present for smoking, diastolic blood pressure, and social class. There was no relation between obesity and coronary death after adjustment for other risks. Our results suggest that some other factors protect women against CHD. The potential for women to reduce their risk of CHD by changes in lifestyle may be less than for men.