OBJECTIVE To compare the hypnotic effects of a bedtime dose of zolpidem, triazolam, and placebo. METHODS "Double-blind, randomized, placebo- and active-controlled, parallel-group" trial. METHODS Six Canadian hospitals. METHODS 357 patients (aged 19 to 71 years) hospitalized the night before a surgical procedure. METHODS At bedtime, each patient received either zolpidem 10 mg, triazolam 0.25 mg, or placebo, and was allowed to sleep for a maximum of 8 hours. METHODS Outcome measures were subjective in nature and included a morning questionnaire, visual analog scales, and observation forms by study personnel. All continuous variables were analyzed by analysis of variance. All categorical data were compared using the Cochran-Mantel-Haenszel (CMH) test, and the percentage of patients asleep was compared using a CMH chi-square analysis. When significant overall treatment effects were observed, pairwise comparisons were undertaken. Compared with the placebo group, the following parameters were significantly (p < 0.001) different in the zolpidem and triazolam groups: sleep latency was shorter, total sleep time was longer, patients fell asleep more easily, and the number of patients awake 2 hours after drug administration was lower. There were no differences between any groups in next-morning somnolence or ability to concentrate. Both drugs were well tolerated, with adverse event incidence rates nearly identical to placebo. CONCLUSIONS In patients suffering from transient insomnia, a single dose of zolpidem 10 mg was as effective as triazolam 0.25 mg, and both were more effective than placebo and were well tolerated.