Objective: To compare the efficacy of cefixime with that of pivamdinocillin in the treatment of adults with acute dysentery caused by Shigella infection. Design: Randomized, double-blind clinical trial. Setting: A diarrhea treatment center in Dhaka, Bangladesh. Patients: 30 men with dysentery lasting 72 hours or less. Interventions: Patients were randomly assigned to receive either 400 mg of cefixime every 24 hours (n = 15) or 400 mg of pivamdinocillin every 6 hours (n = 15) for 5 days. All patients were hospitalized for 6 days. Patients in whom initial drug therapy failed received alternative antimicrobial therapy. Measurements: Physical examinations were done and symptoms were recorded daily, and body temperatures were measured every 6 hours. Stools were counted and examined for consistency and for the presence of blood and mucus. Therapy failed if symptoms of dysentery persisted for more than 72 hours or if, on study day 5, a patient had six stools, one watery or bloody-mucoid stool, or an oral temperature higher than 37.8°C. Bacteriologic failure of therapy occurred if Shigella could be isolated from a stool sample on or after study day 3. Results: Therapy failed in seven (47%) patients given cefixime but in none of the patients given pivamdinocillin (P = 0.06). Patients given cefixime had longer duration of fever (median, 6 hours compared with 0 hours, P = 0.019), longer duration of the period with dysenteric stools (median, 4 days compared with 1 day, P = .001), and more stools during the 6 study days (median, 65 compared with 28, P = 0.002) than patients treated with pivamdinocillin. Bacteriologic failure of therapy occurred in 60% of patients (9 of 15) given cefixime and 13% of those (2 of 15) given pivamdinocillin (P = 0.009). Conclusion: Cefixime is ineffective in treating shigellosis in adults when used in the standard recommended dosage.