Current guidelines for peritoneal dialysis adequacy are based on kinetics of small water-soluble molecules and do not consider the role of other compounds such as middle molecules and protein-bound solutes. Information on the elimination characteristics of the latter solutes by peritoneal dialysis is limited. Moreover, their relation with uremic symptoms remains unclear. The aim of the present study was (1) to investigate the relative contribution of residual renal function to the overall clearances of beta2-microglobulin (beta2m), a middle molecule, and p-cresol, a protein-bound solute, in adults on peritoneal dialysis as compared to small water-soluble molecules and (2) to evaluate relations between serum levels and uremic symptoms. We performed a cross-sectional observational study, including 30 nonanuric peritoneal dialysis patients. Total, peritoneal, and renal clearances were calculated for urea nitrogen (60 D), creatinine (113 D), phosphate (96 D), beta2m (11.8 kD), and p-cresol (108 D). All patients were asked to complete a uremic symptom questionnaire. Declining total clearances (L/week/1.73 m2) were measured for urea nitrogen, creatinine, phosphate, beta2m, and p-cresol, respectively: 97.3 +/- 4.6, 98.9 +/- 6.1, 64.0 +/- 3.4, 23.1 +/- 2.6, and 17.5 +/- 2.3 (Friedman test P < 0.001). Conversely, the contribution of residual renal function (%) to the respective solute clearances increased significantly: 31.6 +/- 3.2, 51.0 +/- 4.0, 42.4 +/- 4.0, 68.0 +/- 5.4, 61.9 +/- 4.6 (Friedman test P < 0.001). The serum level of p-cresol, but of none of the other solutes examined, correlated significantly with the symptom score (Pearson r= 0.48, P= 0.008). During peritoneal dialysis p-cresol behaves like beta2m, probably due to its protein binding. The total clearance of both molecules is significantly lower as compared to water-soluble solutes and mainly depends on residual renal function. Our data further suggest that protein-bound solutes are involved in the pathophysiology of uremic symptoms.