Net fluid absorption under membrane transport models of peritoneal dialysis

The effect of oncotic pressure and lymphatic flow on intraperitoneal dialysate volumes in peritoneal dialysis is investigated under each of two membrane transport models: one assuming a homogeneous single-pore membrane and the other a heteroporous three-pore membrane. In both cases, solute and fluid removal are assumed to occur via a mass transport model in which the peritoneum acts like a synthetic membrane separating two well-mixed compartments (body and dialysate). The homoporous mass transport model of Pyle and Popovich and the three-pore model of Rippe et al., although conceptually different, are shown to be equivalent mathematically. This feature allows one to apply the analytical solutions of Vonesh et al. to either model. It also enables one to apply parameter estimates from one model to another; for example, one can apply the lumped sum reflection coefficients of the three-pore model to a homoporous membrane model. A comparison is made between the use of empirically estimated rejection coefficients computed under the homoporous membrane model of Pyle and Popovich versus lumped-sum reflection coefficients calculated in accordance with the three-pore model of Rippe et al. The two models predict similar drain volumes provided the exchange is conducted using glucose as the osmotic agent. However, one does see a significantly different contribution of protein oncotic pressure and lymphatic drainage to fluid absorption under the two sets of osmotic reflection coefficients. Moreover, for a simulated exchange employing an osmotic agent with a molecular weight of 20,000 daltons, the use of reflection coefficients calculated under the three-pore model yields net ultrafiltration values which are more consistent and physiological than results obtained using the empirically estimated rejection coefficients. Since estimates of 'lymphatic flow' will vary according to the quantity and quality of input parameter values (i.e., hydrostatic pressure, protein concentrations, osmotic reflection coefficients), it would be better to label these estimates as the sum of lymphatic and unmodeled net fluid absorption.