Fifty-three adults hospitalized with Shigella dysentery were empirically treated with trimethoprim (200 mg) twice/day for 5 days, a single dose of trimethoprim (600 mg), or placebo in a randomized double-blind trial. During the first 24 hr of therapy, there was a reduction in the number of stools in 18/21 (86%) of patients treated with the 5-day regimen (trimethoprim-5) and 13/15 (87%) of patients treated with a single dose (trimethoprim-1), compared with 7/17 (41%) of the placebo group (P < 0.025, both comparisons). The mean number of stools passed in the first 24 hr of therapy was 10.6, 10.8, and 21.3 stools in the trimethoprim-5, trimethoprim-1, and placebo groups, respectively. The mean (±SD) change in number of stools from baseline among treated patients during the first 24 hr was -4.9 (6.6) and -6.3 (6.3) for the trimethoprim-5 and trimethoprim-1 groups, respectively, compared with an increase of +2.4 (14.8) for the placebo group. There was a clinical failure at 48 hr in 9% of the trimethoprim-5 patients and 13% of trimethoprim-1 patients compared with 70% of placebo patients (P < 0.005, both comparisons). Although we were unable to demonstrate a difference in efficacy between the two dosage schedules of trimethoprim, we conclude that both treatment regimens are effective for the treatment of Shigella dysentery.