The long- and short-term relation of risk factors to sudden cardiac death (SCD) is examined in the Framingham heart study of 2,011 men and 2,534 women aged 35-70 at the fourth biennial exam. Risk factor measurements over the first four biennial exams were averaged and analyzed as predictors of the long-term occurrence of SCD over the ensuing 28 years using Kaplan-Meier survival curves and the Cox proportional hazards method. The relation of risk factors to the short-term risk of SCD was examined by relating risk factors at each biennium to incidence over the ensuing 2 years, using pooled logistic regression analyses. Over the 28 years of follow-up, 171 men and 80 women experienced SCDs. Women had a lower incidence than men at all ages, and even after adjusting for known risk factors, their SCD rate was only 32% of that in men. In the short term, women have an SCD rate that is 23% of that in men. Most of the modifiable or constitutional risk factors, including glucose intolerance, systolic blood pressure, body mass index, and cigarette smoking have a greater long-term than short-term net effect. This is less apparent in women. Electrocardiographic abnormalities such as left ventricular hypertrophy, intraventricular block, and nonspecific repolarization abnormality were better short-term predictors. In men, preexisting coronary heart disease conferred a 3.3-fold (risk factor-adjusted) increased long-term risk of SCD and 5.3-fold increased short-term risk. In women, the long-term risk is 1.9 and short-term risk is 2.8.(ABSTRACT TRUNCATED AT 250 WORDS)