Peritonitis has been a leading complication of long-term therapy with continuous ambulatory peritoneal dialysis (CAPD). This study was designed to evaluate the risk of peritonitis and technique failure according to the initial CAPD connection technique. Patients from all U.S. facilities starting CAPD therapy at home between January 1 and June 30, 1989 were followed for up to 21 months on the initial CAPD connection technique to change in technique or dialytic modality, to transplantation, death or loss to follow-up. Patients were grouped into standard connection techniques (SCT) (N = 1,133), Y-set (N = 1,067), standard UV set (N = 916) and O-set (N = 167). The time to first peritonitis episode was analyzed actuarially and by using the Cox proportional hazards model which adjusted for age, sex, race, cause of ESRD, CAPD program size and ESRD therapy prior to CAPD. Peritonitis occurred on average at 9.0 month intervals with SCT, 15.0 months with Y-set, 13.4 with standard UV and 9.4 with O-set. The relative risk (RR by Cox analysis) of first peritonitis compared to SCT was 0.60 (40% lower) for the Y-set (P less than 0.01), 0.75 for standard UV (P less than 0.01), and similar to SCT (RR = 0.96) for the O-set (NS), all else being equal. Analysis time to second (N = 1,271) peritonitis episode gave similar results as did analysis of time to CAPD technique failure. Significantly higher RR of peritonitis and technique failure was observed for younger and black patients. These findings suggest the utilization of connection techniques with superior results.