The Boston Collaborative Drug Surveillance Program obtained medical histories on 29 premenopausal women with a discharge diagnosis of stroke. Of these women, 14 were otherwise healthy with no known predisposing illnesses. Eleven of the 14 patients were taking oral contraceptives just before admission compared with seven of 56 otherwise healthy control women (13%). The relative risk estimate for stroke among oral contraceptive users compared with nonusers is 26 (lower 90% one-sided confidence bound = 7.0). Cigarette smoking was only weakly associated with stroke in this group of women.