OBJECTIVE: To compare short course (two day) versus traditional (five day) therapy for treatment of diarrheal disease due to Shigella sonnei. DESIGN: Prospective, randomized, double-blind, placebo-controlled. SUBJECTS AND METHODS: Patients at least 6 months of age presenting to Children's Hospital of Pittsburgh with a history of fever and diarrhea (at least three loose or watery, stools a day) or bloody diarrhea or diarrhea and a known contact with culture-proven Shigella infection. Patients were randomized to receive either two days of cefixime (8mg/kg/day) given once daily followed by three days of placebo or five days of cefixime. Families were instructed to maintain daily diaries; phone follow-up was done on days 3, 7 and 14 after enrollment. Follow-up bacterial cultures were obtained on day 7 to assess for bacteriologic cure. There were standardized definitions for cure, improvement, failure and relapse. RESULTS: Forty-seven patients were enrolled. Five were eliminated from analysis because their stool cultures were not positive for Shigella. There were 42 evaluable patients, 24 in the 2 day group and 18 in the 5 day group. Patients ranged in age from 6 months to 17 years. 52% of the subjects were male. Symptoms were improved or had resolved by day 3 of therapy in all patients. There were 8 clinical relapses: 5 of 24 (21%) patients in the 2 day group and 3 of 18 (17%) patients in the 5 day group, P=NS. There were 13 bacteriological failures defined as patients with positive cultures at the 7 day follow-up: 11 of 21 (46%) in the 2 day group and 2 of 15 (13%) in the 5 day group, P < 0.02. The secondary attack rate was 38% in the 2 day group vs 24% in the 5 day group (P=NS). CONCLUSION: Two and five day treatment courses with cefixime for treatment of diarrheal disease due to Shigella sonnel result in similar rates of clinical cure and clinical relapses; however, there was a higher rate of bacteriologic failures with short course therapy.