John T. Daugirdas

Background A kinetic model for beta-2-microglobulin removal and generation was used to explore the impact of adding hemodiafiltration on predialysis and time-averaged serum values. Methods: The model was tested on data from the HEMO study and on a sample of patients undergoing high-flux hemodialysis. The impact of hemodiafiltration on beta-2-microglobulin levels was evaluated by modeling four randomized studies of hemodiafiltration vs. hemodialysis. The impact of residual kidney function on beta-2-microglobulin was tested by comparing results of previously reported measured data with model predictions. Results: In the low-flux and high-flux arms of the HEMO study, measured median beta-2-microglobulin reduction ratios could be matched by dialyzer clearances of 5.2 and 26 ml/min, respectively. Median pre-dialysis serum beta-2-microglobulin levels were matched if generation rates of beta-2-microglobulin were set to approximately 240 mg/day. In another group of patients treated with dialyzers with increased beta-2-microglobulin clearances, measured cross-dialyzer clearances (57±28 ml/min) were used as inputs. In these studies the kinetic model estimates of intradialysis and early postdialysis serum beta-2-microglobulin levels were similar to median measured values. The model was able to estimate the changes in pre-dialysis serum beta-2-microglobulin in each of four published randomized comparisons of hemodiafiltration with hemodialysis, although the model predicted a greater decrease in predialysis serum beta-2-microglobulin with hemodiafiltration than was reported in two of the studies. The predicted impact of residual kidney clearance on pre-dialysis serum beta-2-microglobulin concentrations was similar to that reported in one published observational study. Modeling predicted that postdilution hemodiafiltration using 25 L/4h replacement fluid would lower serum time-averaged concentration of beta-2-microglobulin by about 18.2%, similar to the effect of 1.50 ml/min residual kidney glomerular filtration rate. Conclusions: A two-pool kinetic model of beta-2-microglobulin yielded values of reduction ratio and pre-dialysis serum concentration that were consistent with measured values with various hemodiafiltration and hemodialysis treatment regimens