RESUMO
The impact of serum albumin at start of continuous ambulatory peritoneal dialysis (SA1), serum albumin before death (SA2), and change in serum albumin during continuous ambulatory peritoneal dialysis (delta SA) were prospectively studied in 41 continuous ambulatory peritoneal dialysis (CAPD) patients with a follow-up of 19 +/- 11.6 months. For analysis, patients were divided into two groups at each measurement point: SA1 or SA2 > 3.0 g/dL or delta SA < 1.0 g/dL (group I), and SA1 or SA2 < 3.0 g/dL or delta SA > 1.0 g/dL (group II). On log rank test, the mortality rate was significantly higher in group II patients in relation to SA1 (p < or = 0.0001), SA2 (p = 0.0002), and delta SA (p = 0.001). On univariate Cox proportional hazard regression analysis, SA1 (p = 0.0001), SA2 (p = 0.0001), delta SA (p = 0.002), and episodes of peritonitis (p = 0.0001) were significant. On multivariate analysis, SA2 (p = 0.003) was significantly related to patient mortality. SA2 (r = 0.8; p = 0.0001), but not delta SA, was related to SA1. We conclude that SA2 is the best predictor of patient mortality on CAPD. SA2 is strongly related to SA1. Thus protein restriction in the pre-dialysis stage should be advised cautiously to avoid consequent hypoalbuminemia.