In 1986 we reported that high levels of plasma dehydroepiandrosterone sulfate (DHEAS) reduced the risk of fatal cardiovascular disease (CVD) in 242 men and increased the risk in 289 women from the Rancho Bernardo cohort who were followed up for 12 years. We report here an update on the epidemiology of DHEAS and CVD based on a 19-year follow-up of 1,029 men and 942 women aged 30-88 years from the same cohort. In cross-sectional analyses, DHEAS levels decreased with age in both sexes and were lower in women than men. Men who were overweight were more likely to have low DHEAS levels; women who had hypercholesterolemia or hypertension or were nonusers of estrogen therapy had higher DHEAS levels. Alcohol intake and cigarette smoking were associated with higher DHEAS levels in both sexes. All differences were no longer statistically significant after adjusting for alcohol intake. All participants were followed for vital status. After 19 years there were 254 CVD deaths in men and 199 CVD deaths in women. DHEAS was not associated with CVD or ischemic heart disease (IHD) deaths in age-adjusted analyses where the comparison group was individuals without CVD or IHD death. In contrast, when the comparison group was survivors, multiply adjusted models showed a statistically significant, modestly reduced risk of fatal CVD (RR = 0.85) in men and a nonsignificant increased risk of fatal CVD (RR = 1.11) in women.