Low sodium peritoneal dialysate has been reported to enhance sodium loss and alleviate signs of fluid overload in continuous ambulatory peritoneal dialysis patients. To elucidate the mechanisms involved, we compared ultrafiltration and solute kinetics using low sodium dialysate (LNaD; 105 mEq/liter sodium, 2.5% glucose, 348 mOsm/liter), conventional dialysate with equal osmolality (CD1.5; 132 mEq/liter sodium, 1.5% glucose, 348 mOsm/liter) and conventional dialysate with equal glucose concentration (CD2.5; 132 mEq/liter sodium, 2.5% glucose, 403 mOsm/liter). A 2 liter, six hour exchange of each dialysate was performed on separate days in 10 chronic peritoneal dialysis patients. Transperitoneal solute diffusion was assessed by calculating the permeability-area product (PA) of the peritoneal membrane from the dependence of plasma and dialysate solute concentrations on tie. Net fluid removed using LNaD of 190 +/- 90 (SEM) ml was similar to that using CD2.5 (250 +/- 90 ml) but higher (P < 0.01) than that using CD1.5 (-200 +/- 60 ml). Sodium loss was higher using LNaD (72 +/- 11 mEq, P < 0.01) and CD2.5 (41 +/- 12 mEq, P < 0.05) than using CD1.5 (-18 +/- 8 mEq). Changes in plasma sodium concentration were small during each dwell and were not different among the study dialysates. PA values for urea (23.4 +/- 1.6 ml/min), creatinine (10.0 +/- 1.0 ml/min), and glucose (10.3 +/- 1.3 ml/min) were similar when determined in each dialysate. The PA value for sodium (7.6 +/- 1.5 ml/min) could only be accurately determined in LNaD. We conclude that: (1) net fluid removed is greater using LNaD than CD1.5 despite similar osmolalities because LNaD has a higher glucose concentration and glucose is a more effective osmotic solute than sodium; (2) sodium loss when using LNaD is enhanced by both diffusion and convection; and (3) sodium diffuses across the peritoneum slower than urea, creatinine and glucose. These data suggest that LNaD alleviates signs of fluid overload by increasing net fluid removal and enhancing sodium loss.