An optimum antimicrobial regimen for bacterial infection after orthotopic liver transplantation has not been identified. In this prospective 4 year study of patients undergoing liver transplantation, patients were randomized to receive either piperacillin-tazobactam (112 patient episodes) or ciprofloxacin plus amoxicillin (105 patient episodes) for empirical treatment of infective episodes in the first 3 months after transplant. Metronidazole was added to the ciprofloxacin-amoxicillin regimen where anaerobic infection was suspected. Patient groups were comparable with respect to clinical, biochemical and haematological parameters. At the 72 h primary efficacy end-point, the overall response rate for the intention-to-treat group was 74/112 (66.1%) for piperacillin-tazobactam and 63/105 (60.0%) for ciprofloxacin plus amoxicillin (P=0.399); the corresponding figures for the per-protocol (PP) group were 73/82 (89.0%) (piperacillin-tazobactam) and 61/80 (76.3%) (ciprofloxacin plus amoxicillin) (P=0.038). At the end-of-study assessment, 58.9% of episodes in the piperacillin-tazobactam group had a successful clinical outcome, compared with 50.5% in the ciprofloxacin plus amoxicillin group (P=0.222); the corresponding figures for the PP group were 83.5% (piperacillin-tazobactam) and 68.8% (ciprofloxacin plus amoxicillin) (P=0.038). Staphylococci and aerobic Gram-negative bacilli were the predominant pathogens in both groups. Bacteria resistant to the study drugs were encountered, including methicillin-resistant Staphylococcus aureus, vancomycin-resistant Enterococcus faecium and multiply-resistant Klebsiella spp. Empirical monotherapy with piperacillin-tazobactam is an effective treatment for infective episodes in liver transplant patients.