Little information is available on the epidemiologic characteristics of antibiotic-associated diarrhea (AAD) in children. The authors' aim was to evaluate the incidence of AAD in an outpatient pediatric population and to identify risk factors. Children aged 1 month to 15.4 years treated with oral antibiotics for a proven or suspected infection were enrolled from an ambulatory pediatric practice during an 11-month period. Parents recorded the daily frequency and characteristics of stools using a diary during the antibiotic treatment and for 1 week after it was stopped. An episode of diarrhea was defined by at least 3 soft or liquid stools/d for at least 2 consecutive days. Risk factors for AAD-age, type of antibiotic treatment, type of combined treatment, and site of infection-were analyzed. Of 650 children included, 11% had an episode of AAD, lasting a mean of 4.0 +/- 3.0 days, beginning a mean of 5.3 +/- 3.5 days after the start of antibiotic treatment. No child was hospitalized because of AAD. The incidence of AAD was higher in children less than 2 years (18%) than in those more than 2 years (3%; P < 0.0001). The incidence of AAD was particularly high after administration of certain antibiotics (amoxicillin/clavulanate, 23%; P = 0.003 compared with other antibiotics). The type of combined treatment and site of infection did not influence the onset of AAD. Antibiotic-associated diarrhea was common in these outpatient children, especially for those aged less than 2 years and after the prescription of certain antibiotics, particularly, the combination of amoxicillin/clavulanate.