The present study evaluated end-stage renal disease (ESRD) patient survival in Lombardy, Italy, and the United States (U.S.) using data from two registries, the Lombardy Dialysis and Transplant Registry (RLDT) and the U.S. Renal Data System (USRDS), respectively. For this purpose, 4,196 white patients (2,900 from the USRDS Case Mix Severity Study and all 1296 from RLDT) who started renal replacement therapy in 1986 and 1987 were studied. Compared to Lombardy patients, those in the USA were significantly older (mean age 59.9 +/- 16.4 vs. 55.9 +/- 14.7 years), had a lower proportion of males (53.7 vs. 62.1%), a greater proportion with diabetic nephropathy (29.9 vs. 9.7%) and a significantly greater proportion of patients with the recorded comorbid conditions (heart disease, peripheral vascular disease, cirrhosis, cachexia, malignancy). U.S. patients were less frequently treated with peritoneal dialysis (PD) by day 30 of ESRD (21.2 vs. 30.7). Survival was compared in the Cox proportional hazard regression model, using age, sex, comorbid conditions and early modality of treatment as explanatory covariates. Overall, 48% of the 4196 patients died during the 48 to 72 months follow-up to 12/31/91. Per 100 patient-years the gross death rate for USRDS patients was 28.7 compared to 13.0 of RLDT patients. The unadjusted death relative risk for RLDT was 0.439, that is, 56% lower death rate compared to USRDS patients. Age, sex, diabetic status, each of the recorded comorbid conditions and treatment modality were significantly related to survival and included in the model. The Cox cumulative survival adjusted for all these explanatory covariates survival was for U.S. patients 84.4% at one year, 67.0% at two years and 33.4% at five years, and for RLDT patients 88.3% at one year, 75.9% at two years and 45.9% at five years. The relative mortality risk (RR) for the patients treated in Lombardy adjusted for all the reported covariates was 29% lower than for US patients (RR = 0.71; P < 0.0001). This comparative risk varied significantly by age (P < 0.0001) and was 65 percent lower for Lombardy compared to U.S. patients in the age range 25 to 44 years (RR = 0.35) and about 20% lower for patients over age 65 years (RR = 0.80). This relative risk was mainly related to hemodialysis and was not statistically significant for PD patients. The observed lower mortality risk in Lombardy was less pronounced when adjusted for demographic and comorbid covariates, but was still large and therefore suggests the need for further studies regarding treatment related factors and unmeasured patient factors, particularly in hemodialysis patients.