In this open, controlled, randomized multi-clinic trial, monotherapy with imipenem/cilastatin was compared to amikacin plus piperacillin as empiric antibacterial therapy in 210 neutropenic cancer patients. Of patients randomized, 53 (25%) had bacteriologically documented infections and of those 30 had septicemia. A further 80 patients (38%) were evaluable for clinical efficacy but did not have documented infections. Seventy-seven patients (37%) were non-evaluable due to effective antibiotic treatment before the trial, early institution of other antibiotics during the trial, verified non-bacterial infections, no neutropenia or other reasons. There were no significant differences in terms of efficacy between imipenem/cilastatin and amikacin plus piperacillin but a consistent trend towards higher rates of clinical cure or improvement and of elimination of causative pathogens was noted in the imipenem/cilastatin group. In patients who were severely neutropenic (less than 0.1 x 10(9) granulocytes/l), similar cure rates were obtained in the two treatment groups--again with a tendency towards better results in the imipenem/cilastatin group. Among evaluable patients with septicemia, one patient in the imipenem/cilastatin group had persistent Staphylococcus aureus bacteremia during treatment. Five patients in the amikacin plus piperacillin group had persistent bacteremia during treatment; all but one (a Pseudomonas aeruginosa) caused by strains resistant to amikacin or piperacillin. Clinical and laboratory adverse effects were mild in the imipenem/cilastatin group although nausea was significantly more common than in the amikacin plus piperacillin group. Among patients on amikacin plus piperacillin, one died in renal failure, possibly related to treatment. Drug-related serious adverse events were reported in two additional amikacin plus piperacillin patients; one with drug fever and one with hearing loss. Microbiological adverse effects occurred in similar frequencies in the two groups. It is concluded that imipenem/cilastatin is a promising candidate for monotherapy of bacterial infections in neutropenic cancer patients.