We performed a randomized controlled trial of amoxicillin plus clavulanate versus cefaclor for treatment of acute otitis media. Total daily doses given in three divided doses were 40 mg/kg amoxicillin plus 10 mg/kg clavulanate, and 40 mg/kg cefaclor. Pathogens were eradicated from the middle ear exudate after 3 to 6 days of therapy in 35 (97%) of 36 patients given amoxicillin-clavulanate compared with 24 (75%) of 32 given cefaclor (P = 0.028). When analysis was restricted to patients with positive urine or serum drug assays during therapy, pathogens were eliminated in 33 (97%) of 34 patients given amoxicillin-clavulanate compared with 21 (75%) of 28 given cefaclor (P = 0.026). Bacterial isolates associated with bacteriologic failure of cefaclor therapy were Streptococcus pneumoniae (two patients), beta-lactamase-negative Haemophilus influenzae (four), and beta-lactamase-positive Branhamella catarrhalis (two). The single failure with amoxicillin-clavulanate therapy was associated with non-beta-lactamase-producing H. influenzae isolated from the middle ear exudate. We conclude that cefaclor is less efficacious than amoxicillin-clavulanate for the treatment of acute otitis media.