In a randomized study the clinical and bacteriologic effectiveness of imipenem was compared with the classical combination of netilmicin with clindamycin in patients who had surgery for an intraperitoneal infection, localized or generalized, with positive bacteriologic findings of the specimen taken at surgery. Excluded were all patients who received other antibiotics before surgery, or who died within 3 days after antibiotic therapy was started. Imipenem was given at a dose of 500 mg t.i.d., clindamycin 600 mg t.i.d., and netilmicin according to serum levels. The diagnoses ranged from postoperative peritonitis, gallbladder empyema, perforated gastroduodenal ulcer, small bowel perforation with and without obstruction, and perforated appendicitis to perforation of the colon. The bacteriologic work-up included examination of the primary specimen (aerobic and anaerobic), the urine, feces, and serologic testing for Candida albicans once or twice a week and after the course of antibiotic therapy. In addition, pH measurements of abscesses and drainage fluids were performed. Ninety-three patients entered the study. Forty-seven patients were treated with imipenem (test group), and 46 patients were treated with the combination therapy (control group). The two groups did not show significant differences in age, sex, diagnostic groups, risk factors, primary bacteriology, and duration of therapy (mean: 6.7 days). Thirty-eight patients (80.9%) treated with imipenem were cured, six patients (12.8%) were improved, and there were three (6.4%) failures. The respective numbers for the control group were 31 (67.4%), 10 (21.7%), and 5 (10.9%). The mean duration of hospitalization was 19 days for the test group and 24.5 days for the control group. There were four wound infections in the test group and 11 wound infections in the control group. Imipenem is at least as effective in the adjuvant therapy of intra-abdominal infections as the combination of netilmicin with clindamycin.