Fifty-six patients with severe intra-abdominal infection have been included in a controlled randomized open study. Imipenem/cilastatin was administered to 28 patients. The remaining 28 patients received ceftazidime, including 7 patients who received a combination of ceftazidime with metronidazole. The difference in cure rates did not reach statistical difference, although 3 failures were reported in the ceftazidime group and none in the imipenem/cilastatin group. Based on 100 pretreatment antibiograms, far more organisms were susceptible to imipenem (99) than to ceftazidime (80) (p < 0.001). The presence of a pathogen resistant to therapy in the follow-up culture was observed more frequently in the ceftazidime group (p < 0.05). Tolerance and safety profile were food in both groups. It is concluded that imipenem/cilastatin is an efficient empiric monotherapy in severe intra-abdominal infections.