Survival of patients undergoing renal replacement therapy in one center with special emphasis on racial differences.

This study compared racial differences in end-stage renal disease (ESRD) in 550 patients starting renal replacement therapy at a large academic dialysis center between January 1, 1990, and December 31, 1993, with follow-up through December 31, 1994. Patient groups were compared with respect to cause of ESRD, comorbid factors at the start of dialysis therapy, choice of modality, transplantation rate, and survival. Fifty-eight percent of the patients were white and 42% were African-American. There was a similar distribution of causes of ESRD between races. African-American patients were less likely to choose peritoneal dialysis as initial therapy (11.6% v 29.3%; P < 0.001) and were less likely to change dialysis modality. Transplantation rates were significantly different between African-American and white patients (9.3% v 27.6%; P < 0.001). African-Americans less frequently received living-related, living-nonrelated, and cadaveric renal transplants. Given differences in transplantation rates and in survival of transplanted patients versus patients on dialysis, survival analysis was performed without censoring for transplantation. A multivariate Cox proportional hazards model was formed, and the following were identified as being significant independent predictors of survival: age, race, age-race interaction, serum albumin at the start of dialysis, activity level at the start of dialysis, and presence of congestive heart failure and cancer. Age had little effect on survival among African-American patients, while it was a significant predictor of survival in white patients. In the group of patients starting dialysis before the age of 30 years, African-American patients had a significantly increased mortality risk compared with white patients. However, white patients older than 50 years had a higher mortality risk; this risk difference increased with age. Racial differences in mortality among older white patients could not be explained by differences in comorbid conditions, transplantation rates, or withdrawal from dialysis.