The influence of the type of dialysis on survival of patients with end-stage renal disease (ESRD) is controversial. To compare survival among patients with ESRD receiving peritoneal dialysis (PD) or hemodialysis (HD), we conducted a prospective cohort study in a single center from April 1995 to March 2005. During that period, 454 patients (161 women, 293 men; mean age: 61.7 +/- 14.4 years; 46.6% with diabetic nephropathy) were started on HD therapy, and 120 patients (40 women, 80 men; mean age: 54.5 +/- 11.3 years; 16.7% with diabetic nephropathy) were started on PD therapy; all patients were followed for at least 3 years. The 3-year survival rates were 65% for the HD patients and 81% for the PD patients (p < 0.05). The causes of death in patients undergoing HD were 52% cardiovascular 25% infectious diseases, and 12% cancer; in patients undergoing PD, the causes were 36% infectious diseases, 24% cardiovascular, and 6% cancer Median time from initiation of dialysis to study enrollment was 90 days for HD patients and 180 days for PD patients. Although patients in this study were not randomly assigned to their initial type of dialysis therapy, survival rate was found to be dependent on dialysis type. Moreover, this study suggests the importance of early referral and evaluation of risk factors in individual patients before they are started on dialysis therapy.