Serum albumin: a predictor of long-term outcome in peritoneal dialysis patients.

We report a long-term follow-up of 71 peritoneal dialysis patients who participated in a prior study that evaluated the serum albumin concentration as a predictor of short-term morbidity. In this study the use of the original serum albumin level to predict failure of peritoneal dialysis or death was investigated. Sixty-nine of the original 71 study patients were followed for 2 years from the time of enrollment in the initial study. Patients who died within this 2-year period had a significantly lower serum albumin concentration during the original study period (2.7 +/- 0.7; n = 8) than those remaining on peritoneal dialysis (3.6 +/- 0.4, n = 31), those transferred to hemodialysis (3.4 +/- 0.5; n = 13), or those receiving renal transplants (3.6 +/- 0.4 g/dL; n = 17) (P < 0.05 for all comparisons v the group that died). A low serum albumin during the original study period was not predictive of patients who transferred to hemodialysis. We conclude that the stable outpatient serum albumin concentration in peritoneal dialysis patients is a powerful predictor of mortality as well as of short-term morbidity.