Serum albumin: a marker for morbidity in peritoneal dialysis patients.

This study was designed to determine if the serum albumin is a marker for morbidity or mortality in peritoneal dialysis (PD) patients. The impact of a low serum albumin on the risk of hospitalization, peritonitis, or death was examined in 71 patients. Blood urea nitrogen (BUN), cholesterol, age, and the presence or absence of diabetes were also examined. In independent analyses, the serum albumin was lower (32.7 +/- 5.6 v 36.3 +/- 4.3 g/L, P < 0.01), the diagnosis of diabetes was more frequent (41% v 7%, P < 0.01), and the number of episodes of peritonitis were greater (2.0 +/- 1.6 v 0.7 +/- 1.3, P < 0.01) in the group of patients hospitalized compared with those not hospitalized. When diabetics were excluded from analysis, the serum albumin remained significantly lower in hospitalized patients. Stepwise logistic regression analysis, excluding the 10 patients hospitalized only for treatment of peritonitis, confirmed that only a low serum albumin and the diagnosis of diabetes were independent predictors of increased morbidity as evident by the increased frequency of hospitalization. Every 10 g/L decrease in the serum albumin increased the odds ratio for hospitalization by 5.2. The diagnosis of diabetes resulted in a 10-fold increase in the odds ratio. We conclude that a low serum albumin serves as a marker of morbidity in PD patients, primarily as a marker of increased risk for hospitalization. The diagnosis of diabetes also greatly increases the likelihood of hospitalization. Peritonitis is a cause for hospitalization, but not an independent risk factor.