This study was performed to evaluate the use of the glucose concentration versus glucose absorption in the dialysate to estimate ultrafiltration efficacy. In 26 patients (16 men, 10 women) on continuous ambulatory peritoneal dialysis, a total of 128 dwells were investigated. Each dwell lasted 5 hours, using 2 L of peritoneal dialysis fluid containing 140 mmol/L glucose. After the dwell, the removed peritoneal dialysis fluid was weighted. The concentrations of creatinine and glucose and the total amount of glucose were analyzed. The mean concentration of dialysate glucose for the whole group after a 5-hour dwell was 38.9 +/- 8.9 (SD) mmol/L. The ultrafiltration volume was inversely correlated to the glucose absorbed (r = -0.59, p = 0.002), but was not significantly correlated to the glucose concentration in the dialysate (r = 0.40, p = 0.057). The ultrafiltration volume was not related to the creatinine concentration in the dialysate or the body weight. The average dialysate glucose did not change during a one-year observation time. Patients with diabetes mellitus had significantly lower glucose absorption than nondiabetics (p = 0.045), and patients with low ultrafiltration efficacy had significantly greater glucose absorption than those with normal and high ultrafiltration (p = 0.038). There was no difference between those who had those who had not suffered from peritonitis. The results indicate that the glucose absorption from the dialysis fluid is the best parameter to calculate the peritoneal ultrafiltration efficacy and not the glucose concentration currently used.