To test the hypothesis that female prevalence is greater than expected among reported cases of torsades de pointes associated with cardiovascular drugs that prolong cardiac repolarization. A MEDLINE search of the English-language literature for the period of 1980 through 1992, using the terms torsade de pointes, polymorphic ventricular tachycardia, atypical ventricular tachycardia, proarrhythmia, and drug-induced ventricular tachycardia, supplemented by pertinent references (dating back to 1964) from the reviewed articles and by personal communications with researchers involved in this field. Ninety-three articles were identified describing at least one case of polymorphic ventricular tachycardia (with gender specified) associated with quinidine, procainamide hydrochloride, disopyramide, amiodarone, sotalol hydrochloride, bepridil hydrochloride, or prenylamine. A total of 332 patients were included in the analysis following application of prospectively defined criteria (eg, corrected QT [QTc] interval of 0.45 second or greater while receiving drug). Clinical and electrocardiographic descriptors were extracted for analysis. Expected female prevalence for torsades de pointes associated with quinidine, procainamide, disopyramide, and aminodarone was conservatively estimated from gender-specific data reported for antiarrhythmic drug prescriptions in 1986, as derived from the National Disease and Therapeutic Index, a large pharmaceutical database; expected female prevalence for torsades de pointes associated with sotalol, bepridil, and prenylamine was assumed to be 50% or less since these agents are prescribed for male-predominant cardiovascular conditions. Women made up 70% (95% confidence interval, 64% to 75%) of the 332 reported cases of cardiovascular-drug-related torsades de pointes, and a female prevalence exceeding 50% was observed in 20 (83%) of 24 studies having at least four included cases. When analyzed according to various descriptors, women still constituted the majority (range, 51% to 94% of torsades de pointes cases), irrespective of the presence or absence of underlying coronary artery or rheumatic heart disease, left ventricular dysfunction, type of underlying arrhythmia, hypokalemia, hypomagnesemia, bradycardia, concomitant digoxin treatment, or level of QTc at baseline or while receiving drug. When cases of torsades de pointes were analyzed by individual drug, observed female prevalence was always greater than expected, representing a statistically significant difference (P < .05) for all agents except procainamide. These findings strongly suggest that women are more prone than men to develop torsades de pointes during administration of cardiovascular drugs that prolong cardiac repolarization. The pathophysiological basis for, and therapeutic implications of, this gender disparity should be further investigated.