An open, prostective, randomized multicenter clinical trial randomly allocated successive patient who were scheduled for a surgical procedure for serious intra-abdominal infections to receive either treatment CM (ceftazidine plus metronidazole) or monotherapy with treatment IC (imipenem/cilastatin). Out of 90 eligible patients, 87 were clinically evaluable of which 71 were clinically and bacteriologically evaluable (CBE). Cases allocated to each treatment group were comparable as to age, sex, diagnostic group distribution, mean APACHE II scores, and bacteriologic evaluability. Among the 87 clinically evaluable patients, there were 4 (9.1%) and 2 (4.7%) treatment failures among those who received treatments CM and IC respectively (p=0.486). For all eligible patients, the mean fever days was 2.07, mean treatment days was 6.01, and mean hospital days was 11.57, and was not significantly different between the two treatment groups. Among clinically evaluable cases, the mean APACHE II scores of patients with succesful outcomes (5.8) was very significantly lower (p=0.000) than that of patients whose treatment failed (13.8). This was also true for CBE cases. Logistic regression analysis showed that among six variables (diagnostic group, APACHE II score, antibiotic used, fever days, hospital days and treatment days) only the APACHE II score signficantly contributed to treatment failure (p=0.001).