In this controlled, randomized clinical trial we compared piperacillin and carbenicillin in the treatment of complicated urinary tract infections. 24 patients received piperacillin 150 mg/kg/day for 7.2 ± 2.75 days and 17 patients received carbenicillin 200 mg/kg/day for 7.5 ± 2.90 days. Patients were evaluated for clinical and bacteriologic responses and tolerance to therapy. Although the clinical cure rate significantly favored carbenicillin treatment (p < 0.01), the sum of the percentages of cases with clinical cure and clinical improvement were similar between groups: 91.6% for piperacillin and 88.2% for carbenicillin. The bacteriologic cure rates for piperacillin and carbenicillin patients (54.1 and 47.0%, respectively) were not significantly different (p > 0.05). The low cure rates in our study were probably the result of uncorrected/uncorrectable genitourinary tract abnormalities. Superinfections developed in 12.5 and 17.6% of piperacillin and carbenicillin patients, respectively, and were due to Klebsiella pneumonia, Proteus mirabilis, Citrobacter diversus, and Pseudomonas aeruginosa. Overall, side effects were mild, reversible, and did not require discontinuation of treatment. However, carbenicillin caused elevations in liver enzymes more frequently than piperacillin (p < 0.05). Based on our data, we recommend reserving piperacillin monotherapy for patients who are poor candidates for aminoglycosides, or are on severe sodium restriction, and have serious complicated urinary tract infections due to susceptible organisms. We do not recommend piperacillin alone for empiric treatment of complicated urinary tract infections.