We compared the efficacy of amoxicillin with that of the combination drug sulfamethoxazole and trimethoprim in reducing recurrences of acute otitis media (AOM) in a single-blind, randomized, placebo-controlled trial involving 96 children. Each of the children had had three or more episodes of AOM in the preceding 6 months, and 97% (93/96) of them still had unilateral or bilateral effusion at the beginning of the study. During the 6-month study period, 9 (27%) of 33 of the children in the amoxicillin group developed 9 episodes of AOM, 9 (27%) of 33 of the children in the sulfamethoxazole and trimethoprim group experienced 11 episodes of AOM, and 19 (63%) of 30 of the children in the placebo group developed 25 episodes. Young age and day-care attendance characterized children for whom prophylaxis was more efficacious. Overall persistence of middle-ear effusion was shorter in treated children only as a consequence of the reduced number of new episodes of AOM.