Comorbidity, urea kinetics (Kt/V and normalized protein catabolic rate), dietary protein, total calorie intake, and plasma albumin were measured in 97 continuous ambulatory peritoneal dialysis patients followed prospectively for 30 months. Comorbid disease was graded severe in 12 patients, intermediate in 29, and absent in 56. At entry to the study comorbidity was associated with increased age (P = 0.001), lower dietary protein (P = 0.015) and calorie intake (P = 0.02), and a lower plasma creatinine (P = 0.026). Trends toward lower Kt/V and albumin were not significant, and normalized protein catabolic rate was unaffected. Ability of these measures to predict mortality was assessed by univariate and multivariate analysis using Cox's proportional hazard model. On univariate analysis, comorbidity (P < 0.0001), age (P = 0.0001), Kt/V (P = 0.009), plasma albumin (P = 0.009), calorie intake (P = 0.035), and dietary protein intake (P = 0.03) predicted outcome, whereas normalized protein catabolic rate did not (P = 0.46). Multivariate analysis indicated that comorbidity (P = 0.0003) and age (P = 0.0085) were the only independent predictors of outcome. The addition of plasma albumin and Kt/V increased the significance of the Cox model. Further analysis of comorbidity demonstrated the relative importance of vascular disease and left ventricular dysfunction. This study illustrates the profound influence of comorbid disease on mortality in continuous ambulatory peritoneal dialysis patients and suggests that it causes suppression of appetite independent of the dialysis dose.