The known risk factors of atherosclerotic diseases may be involved in the development of a subarachnoid hemorrhage. We studied the morbidity and mortality due to subarachnoid hemorrhage among 42,862 men and women aged 20-69 years who had participated in a large health survey in Finland. During a mean follow-up of 12 years, 102 non-fatal and 85 fatal cases of subarachnoid hemorrhage were observed. The total incidence was 37 per 100,000 person-years. Smoking and hypertension were positively associated and body mass index was inversely associated with the risk of subarachnoid hemorrhage. These associations were not confounded by age or each other. No statistically significant association with risk was detected for serum cholesterol level, hematocrit content, known heart disease, or diabetes. The risk was especially elevated among lean hypertensive subjects and lean smoking subjects. The age-adjusted relative risks of subarachnoid hemorrhage for lean, hypertensive smokers were 18.3 (95% confidence interval (CI), 7.8-42.7) among women and 6.7 (95% CI, 2.3-19.7) among men as compared to the risk among subjects without these risk factors. We conclude that modifiable risk factors are predictive of subarachnoid hemorrhage, for which reason subarachnoid hemorrhage may in part be preventable. Leanness combined with arterial hypertension and/or smoking, in particular, poses a substantially elevated risk.