The objective of this study was to compare prophylaxis for Staphylococcus aureus infections in peritoneal dialysis patients using 600 mg cyclic oral rifampin for 5 days every 3 months versus mupirocin calcium ointment 2% applied daily to the exit site. The study design was a prospective randomized trial, controlling for S aureus nasal carriage. Eighty-two continuous ambulatory and continuous cyclic peritoneal dialysis patients (54% male, 71 % white, 34% insulin-dependent, mean prestudy time on peritoneal dialysis 1.2 years) were randomly assigned to cyclic rifampin (n = 41 patients) or daily exit site mupirocin prophylaxis (n = 41 patients). Mean follow-up was 1 year. S aureus catheter infection rates were 0.13/yr with mupirocin and 0.15/yr with rifampin (P = NS). Both rates were significantly lower than the center's historical rate (the period between 1983 and 1992) of 0.46/yr prior to the study (P < 0.001). S aureus peritonitis rates were 0.04/yr with mupirocin and 0.02/yr with rifampin (P = NS), both significantly lower than the center's historical rate of 0.16/yr (P < 0.02). Catheter loss due to S aureus infections was 0.02/yr with mupirocin and 0/yr with rifampin (P = NS), both significantly lower than the center's historical rate of 0.12/yr (P < 0.001). There were no side effects in patients using mupirocin, but 12% were unable to continue rifampin due to side effects. We conclude that mupirocin ointment at the exit site and cyclic oral rifampin are equally effective in reducing S aureus catheter infections. In addition, rifampin or mupirocin significantly reduced S aureus peritonitis and catheter loss due to S aureus infections. Mupirocin at the exit site provides an excellent alternative prophylaxis for S aureus infections, particularly in patients who cannot tolerate oral rifampin therapy.