♦ Background: Icodextrin induces ultrafiltration (UF) during long-dwell exchanges by creating a difference in oncotic pressure between the peritoneal cavity and plasma; however, the mechanisms governing intra-patient and inter-patient variability in UF when icodextrin is used remain largely unexplained. In the present study, we show theoretically that differences in peritoneal residual volume (V R ) have a more profound effect on UF with icodextrin use than with glucose use. This phenomenon is attributed to a differential effect of V R on oncotic, rather than osmotic, pressure between the peritoneal cavity and plasma. ♦ Methods: The three-pore model was used to calculate the effect on UF of V R between 150 mL and 1200 mL when 7.5% icodextrin (ICO) or 3.86% glucose solution is used at the end of a 12-hour dwell in the four patient transport groups (that is, fast to slow). Oncotic (with ICO) and osmotic (with glucose) pressure differences averaged over the entire dwell were also calculated. ♦ Results: As expected, at a nominal V R of 300 mL, UF with glucose differed substantially between the four patient transport groups (2 -804 mL), whereas UF with ICO did not (556 -573 mL). When V R was increased to 1200 mL from 150 mL, the concentrations of the oncotic and osmotic agents at the start of the dwell with an infusion volume of 2 L decreased to 4.9% from 7.0% with ICO and to 2.5% from 3.6% with glucose. The decrease in UF on average was greater with ICO [to 252 mL from 624 mL: that is, a reduction of 372 mL (60%)] than with glucose [to 292 mL from 398 mL: that is, a reduction of 106 mL (