We retrospectively reviewed 167 consecutive peritoneal dialysis patients with regard to serum albumin (Alb), mortality and abdominal complications. In addition, 25 patients were studied with serial measurements of urea kinetics. The patients were divided into four groups based on their dialysis index (DI) and normalized protein catabolic rate (NPCR) (Table I). 12/167 patients were identified with abdominal catastrophes. Before these complications occurred, the M Alb in this group was 2.67 + 0.24 (compared to age, sex and disease matched controls of 3.55 + .11 P < .05). Six of these patients died from abdominal complications. In the 26 patients with serial urea kinetic studies, 4/11 patients in group IV died (low NPCR and low DI) (P < .05 compared to Group I, II or III). We conclude that urea kinetic modeling is predictive of outcome in those patients with presumed poor nutrition and inadequate dialysis and that abdominal catastrophes are more common in those patients with poor nutrition. Prospective interventional studies should be designed in an attempt to improve the poor outcome in this group of patients.