A total of 127 patients from 8 hospitals were randomized into 1 of 2 exit-site care regimes to evaluate their effect on rate of exit-site infection (ESI). Group 1 used povidone iodine and nonocclusive dressings changed 2 to 3 times weekly; Group 2 simply cleansed the exit site with nondisinfectant soap and water. Incidence, cause, duration, and treatment of ESI and peritonitis (P) were noted. Groups were analysed for age, sex, end-stage renal disease (ESRD), catheter, and systems. Total cumulative follow up time was 95.6 years. There was a significantly higher rate (p = 0.0183) of ESI in Group 2 (soap and water). The mean rate of ESI was 0.27 episodes/patient year for Group 1 versus 0.71 episodes/patient year for Group 2. Rates of P for the two groups were not significantly different (p greater than 0.50): 0.446 episodes/year for Group 1 versus 0.574 episodes/year for Group 2. S. aureus was responsible for 83% of ESI in Group 1 and 67% of ESI in Group 2. Protective dressing with a disinfectant is associated with significantly less ESI than minimum care. However, further research in exit-site care aimed specifically at reducing S. aureus infection is still required.