In the United States, chronic peritoneal dialysis take-on has declined among incident ESRD patients. Although increasing age, co-morbidity, and body size may explain part of this decline, other factors likely contribute. Among incident ESRD patients in the United States, we found that peritoneal dialysis take-on significantly decreased from 11% in 1996 to 1997 to 7% in 2002 to 2003 (P < 0.001 for the trend). This decrease remained after adjusting for patient demographics, case-mix, and laboratory data, suggesting the involvement of other factors. This decline in utilization occurred during a time of improving outcomes for incident peritoneal dialysis patients, measured as reduced hazards for death or for the need to transfer to hemodialysis. In contrast, among patients initially treated with hemodialysis, the 12-month adjusted hazards for death or transfer to peritoneal dialysis slightly worsened or were unchanged over this same period. Therefore, the decline in peritoneal dialysis take-on cannot be entirely explained by increasing age, co-morbidity and body size of incident ESRD patients. The decline in utilization has occurred at a time when the early outcomes of CPD patients have improved.