Simple Models for Fluid Transport during Peritoneal Dialysis

Abstract

Peritoneal fluid transport can be predicted using different simplified formulas. To evaluate three such models, fluid transport was studied in 38 single six hour dwell studies using standard glucose 1.36% (n=9), 2.27% (n=9) and 3.86% (n=20) dialysis fluids as well as amino acid 2.70% fluid (n=8) in 33 patients on continuous ambulatory peritoneal dialysis (CAPD). Dialysate volume and the peritoneal absorption rate were measured using radioiodinated serum albumin (RISA) as a marker. The dialysate volume over dwell time curves were examined using three mathematical models of fluid transport for solutions with a crystalloid osmotic agent: Model P based on phenomenologically derived exponential function of time (Pyle, 1981), Model OS based on linear relationship between the rate of net volume change, Qv, to the difference of osmolality in dialysate and blood, and Model G based on linear relationship between Qv and the difference of glucose concentration in dialysate and blood. All these models provided a good description of the measured dialysate volume over time curves, however the descriptions with Models OS and G for glucose 3.86% fluid were slightly but significantly less precise. The coefficients of Model OS were stable in time, but the coefficients of Model G and P dependend in general on the time period used for their estimation, especially for glucose 3.86% dialysis fluid. The evaluation of dwell studies with solutions containing amino acid 2.70% (instead of glucose) as osmotic agent, using Model OS and P, showed that the transport coefficients were stable in time and both models provided equally precise descriptions. These results suggested that all three models can be used but models P and OS can be preferred for pratical applications such as predictions of fluid transport with alternative cristalloid osmotic agents. Furthermore, we found that the peritoneal barrier for fluid transport may change transiently during exchanges with the standard glucose - based dialysis fuid, whereas such changes were not observed with the amino acid-based fluid. This discrepancy may be due to a different composition of the dialysis fluids, including osmotic agent, buffer and pH.