Seventy-seven patients were prospectively enrolled in a randomized clinical trial to compare two antimicrobial regimens for the treatment of post-cesarean section endometritis. The two groups were not significantly different with respect to age, race, gravidity, parity, hours in labor, cesarean section indication, preoperative or postoperative hemoglobin/hematocrit, pretreatment white blood cell count or pretreatment temperature. Pretreatment urine, blood and endometrial cultures were obtained. One or more organisms was recovered from the endometrium in 90% of the patients using a double-lumen sampling device. The most frequent endometrial isolates were Peptostreptococcus and Bacteroides species, followed by Gardnerella vaginalis and enterococci. Thirty (81%) of 37 patients receiving ampicillin/sulbactam and 33 (83%) of 40 receiving gentamicin and clindamycin responded to therapy. There were 14 (18%) treatment failures, 7 in each group. Five (36%) of the 14 clinical failures were due to septic pelvic thrombophlebitis, 2 (14%) of the 14 failures were complications of intraabdominal abscesses, and the remaining 7 patients responded after a change in their antibiotic regimen. We conclude that ampicillin/sulbactam and clindamycin/gentamicin are similarly effective for the treatment of post-cesarean section endometritis.