Comorbidity, age, dialysis dose (KT/V(urea)), plasma albumin, and peritoneal function (D/P(creat) were measured cross-sectionally in 228 continuous ambulatory peritoneal dialysis (CAPD) patients, who were then followed up for a mean of two years. Comorbidity, utilizing a semi-quantitative score described previously, was the most powerful predictor of mortality in both univariate and multivariate analysis. Using univariate analysis, all the variables predicted outcome with statistical significance, mortality being associated with lower KT/V and plasma albumin and a higher D/P(creat). On multivariate analysis only comorbidity, age, and KT/V remained independent predictors. Data was further analyzed on the basis of type of comorbid condition. In those patients without comorbid disease (n = 127) neither KT/V, albumin nor D/P(creat) predicted outcome. In patients with clinical evidence of ischemic heart disease the KT/V was a significant predictor of favorable outcome. In those with clinical evidence of left ventricular function, mortality was significantly and independently associated with low plasma albumin, high D/P(creat), and KT/V. It is suggested that the concept of treatment adequacy in CAPD patients must include both measures of dialysis dose and peritoneal function, particularly in the context of the patient's comorbidity.