One hundred patients with intraabdominal infections were assigned randomly in double-blind fashion to receive either the combination of tobramycin plus clindamycin (TM/C) or moxalactam (MOX) alone. Fifty patients comprised each group, but one patient in each group died of infection before 48 hours treatment. In the remaining 98 patients, the average age was 62 years, initial serum albumin was 3.0 mg/dl, serum creatinine was 1.5 mg/dl, and over half of the patients were nutritionally deficient by the prognostic nutritional index criteria. In approximately one-half of the patients, the source of infection was perforated colon or perforated appendix. There were no significant differences in demographic factors between these groups, except that those who were given TM/C were older, while those who were given MOX had a more serious long-term prognosis due to underlying disease. The average length of treatment was 11 days, and the average hospitalization time was 24 days. Clinical response to therapy was identical, since 74% of the TM/C patients and 76% of the MOX patients had satisfactory responses. Bacteria persisted at the site of infection in 63% of the TM/C patients and in 65% of the MOX patients, with the most common isolate being Staphylococcus epidermidis. Pseudomonas infections were the most difficult to cure in both groups. The two regimens differed only in side effects; TM/C was a more frequent (p less than 0.05) cause of nephrotoxicity, and elevated prothrombin time/partial thromboplastin time (PT/PTT) was more frequently (p less than 0.05) observed in MOX. All PT/PTT elevations responded to injections of vitamin K, and no serious bleeding occurred. Choice between these regimens depends on the risk of renal versus hematologic side effects, rather than efficacy.