The present study compared oral versus intraperitoneal (ip) ciprofloxacin (ciproxin) as primary treatment of bacterial peritonitis in patients receiving continuous ambulatory peritoneal dialysis (CAPD) in a randomized, prospective trial. A total of 54 episodes in 46 patients were recruited for study. After excluding nonbacterial episodes and those not treated according to protocol, 48 episodes evenly divided between the two treatment arms were eligible for analysis. The primary cure rate was 41.7% and 66.7%, respectively, in the oral and ip treatment group. Half of those who failed or relapsed were due to infection with resistant, mostly gram-positive bacteria, which accounted for 79% of culture-positive episodes. Of the gram-positive isolates 42.3% were either resistant or intermediately susceptible to ciproxin compared to 16.7% of gram-negative isolates. The high level of bacterial resistance to ciproxin and treatment failure rate were related to the previous exposure to fluoroquinolones. Inadequate trough peritoneal drug levels also accounted for the failures in the ip but not the oral treatment group. We conclude that oral ciproxin is ineffective as a primary treatment of CAPD peritonitis in patients previously exposed to fluoroquinolones and that when administered ip, a dose of 50 mg/L instead of 25 mg/L of ciproxin should be used as maintenance in order to achieve adequate trough peritoneal drug levels.