Serum albumin in patients on continuous ambulatory peritoneal dialysis - Predictors and correlations with outcomes

Serum albumin (SA) is a powerful predictor of patient morbidity and mortality in hemodialysis, but data are limited for continuous ambulatory peritoneal dialysis (CAPD). SA was monitored in 76 new CAPD patients over 222 6-month periods and mean SA was correlated with morbidity and mortality during those periods. The influence of initial SA on duration of technique survival was also investigated. To determine which factors best predict SA, correlations with patient demographics and with 6-month measurements of dialytic dose, protein intake, and peritoneal transport were sought. Mean SA overall was 34.1 ± 3.3 g/ L, and mean initial SA was 33.4 ± 3.1 g/L. Mean SA was lower in diabetics and in those aged 65 or over. Mean SA tended to increase during the first year on CARD, and this increase was maintained, except in patients aged 65 or over, where it tended to revert to initial values. SA correlated with hospital days (r = -0.20; P<0.005), fatigue index (r= -0.20; P<0.005), nerve conduction (P<0.001), and a variety of laboratory values, and lower SA was associated with technique failure (P<0.03) and death (P<0.07). Initial, as well as ongoing, SA was predictive of technique failure (P<0.05) and Cox proportional hazards regression showed that this predictive power was independent of age, sex, diabetes, and other factors (P = 0.05). The strongest predictors of low SA by stepwise multiple regression were diabetes, a higher dialysate-to-plasma creatinine equilibration ratio, older age, lower body weight, and shorter time on CAPD. Neither protein catabolic rate nor dialytic dose was predictive of SA. SA correlates with morbidity and mortality in CAPD and is primarily influenced by factors that are difficult to alter. Strategies to correct low SA require investigation.