Nineteen patients treated by continuous ambulatory peritoneal dialysis (CAPD) were studied according to clinical outcome parameters: insomnia, asthenia, pruritus, arterial hypertension, anorexia, nausea and/or vomiting, anemia, and rate of hospitalization. Using clinical scores, three groups were defined: poor clinical outcome (P), intermediate (I), and good (G). The quantity of treatment by PD was evaluated monthly with urea kinetic tests (weekly Kt/V, weekly urea clearance/1.73 m2 of body surface area (BSA), index of dialysis by Teehan), and with the weekly creatinine clearance/1.73 m2 of BSA. The metabolic index was analyzed: normalized protein catabolic rate (NPCR), serum albumin (Alb) and prealbumin, and reabsorption of glucose. There was good correlation between clinical scores and quantity of dialysis. The Alb was lower in group P. Group G was differentiated from group I and from group P by quantification tests and NPCR, with lower levels as follows: weekly Kt/V = 2.06, urea clearance 70 L/week/1.73 m2, index of dialysis = 0.87, and creatinine clearance = 60 L/week/1.73 m2. We conclude that the qualitative clinical approach is not sufficient to predict deleterious signs, and the quantitative approach is predictive of the good clinical outcome and good nutritional status. We think that levels proposed to now are insufficient, and we suggest the following: weekly urea clearance > 70 L, weekly Kt/V > 2, weekly creatinine clearance > 60 L, and index of dialysis > 0.85.