We undertook a prospective double-blind controlled study to compare the efficacy of a drug that usually has no antistaphylococcal activity (amoxicillin trihydrate) with the efficacy of the same drug with an addition of a beta-lactamase inhibitor (amoxicillin plus clavulanic acid [Augmentin]) in the treatment of nonbullous impetigo. Fifty-one culture-positive patients, aged 6 months to 9 years, were included, 26 in the amoxicillin group and 25 in the Augmentin group. The study groups were clinically and bacteriologically comparable at the start of the study. Staphylococcus aureus was isolated from all patients and beta-hemolytic streptococcus from 14 (29%). All staphylococci were sensitive to Augmentin but resistant to amoxicillin. Forty-nine patients completed the study. The clinical response was significantly better among the Augmentin recipients (marked improvement in 71% and 95% of patients after 2 and 5 days, respectively; no new lesions during the treatment course) than among the amoxicillin recipients (marked improvement in 44% and 68% of patients after 2 and 5 days, respectively; new lesions appeared in 20% of patients). Recurrence within 3 weeks occurred in 12 (26%) of 49 patients, and no difference was observed between the two groups. We conclude that S aureus is common in nonbullous impetigo, and that at least in some cases it plays an important role in the course of the disease that can be altered by specific therapy.