We compared intravenously administered ciprofloxacin with ceftazidime in a randomized, double-blind study. Patients received ciprofloxacin 200 mg intravenously every 12 hours or ceftazidime 2 g intravenously every eight hours, with placebo infusions to maintain blinding. Therapy with metronidazole was added for suspected or documented intra-abdominal infection. Thirty-two of the 57 ciprofloxacin-treated patients were evaluable for determination of efficacy and had 41 bacterial isolates from 34 sites. Thirty-six of the 56 ceftazidime patients were evaluable for determination of efficacy and had 41 bacterial isolates from 38 sites. Seven of 35 bacteremic patients had no identifiable primary focus. The most commonly isolated pathogens were Escherichia coli, Streptococcus pneumoniae, Staphylococcus aureus, and Klebsiella pneumoniae. Cure rates and bacteriologic eradication rates were comparable. Nine patients did not improve. Patients with treatment failures in the ciprofloxacin group included a quadriplegic patient with relapse of urinary tract infection with bacteremia (K. pneumoniae). Another patient with bacteremia (Pseudomonas aeruginosa), pneumonia (Proteus vulgaris), and urinary tract infection (P. vulgaris, Providencia sp.) died on the first treatment day. The third patient (S. aureus and Streptococcus pneumoniae pneumonia) had a new aspiration pneumonia develop on the third day; the pneumococcus also persisted. Undrained S. pneumoniae empyema caused the fourth ciprofloxacin treatment failure, and the fifth patient had a relapse of S. aureus pneumonia with bacteremia. One ceftazidime-treated patient died of pneumococcal pneumonia on the first day. Another had persistent Staphylococcus epidermidis and Listeria bacteremia despite 48 hours of treatment. Two other patients had pneumonia (S. aureus and P. aeruginosa, respectively) and completed full courses of ceftazidime therapy without improvement. Five patients had pneumococcal bacteremia; four patients were cured: one of two patients in the ciprofloxacin group and three of three patients in the ceftazidime group. Significant increases in the number of platelets (four patients with ciprofloxacin treatment, one patient with ceftazidime treatment) and declines in the number of platelets (one patient with ciprofloxacin treatment, one patient with ceftazidime treatment) were observed. Intravenously administered ciprofloxacin is comparable with ceftazidime and is a safe and effective antibiotic for the treatment of patients with serious infections, including bacteremia.