Comparison of measured vs kinetic-model predicted phosphate removal during hemodialysis and hemodiafiltration.

Background: To what extent hemodiafiltration (HDF) improves management of hyperphosphatemia over hemodialysis (HD) is a subject of ongoing investigation.

Methods: We modified a previously described phosphate kinetic model to include incorporation of EUDIAL recommended equations for hemodiafiltration (HDF) clearance. We used the model to predict the recovery of phosphate from spent dialysate/hemofiltrate and compared this with averaged data from five published studies. Mean study average predialysis serum phosphate was 1.81 ± 0.20 mmol/L. Session length was close to 240 min per treatment. All HDF was done postdilution, at an average rate of 65 ± 24 mL/min.

Results: Measured mean phosphate removal was 1039 ± 136 mg (33.5 ± 4.41 mmol, slightly lower than the model-predicted mean value of 1092 ± 127 mg (35.3 ± 4.09 mmol). The measured ratio of phosphate removal with HDF compared with HD averaged 1.15 ± 0.22, ranging from 1.01 to 1.44. Using mean study input parameters for patient size and treatment characteristics, the predicted ratio of phosphate removal with HDF compared with HD averaged 1.095 ± 0.029, ranging from 1.05 to 1.13.

Conclusions: Addition of EUDIAL-recommended convective clearance equations to a phosphate kinetic model predicts a 10% or greater benefit in terms of phosphate removal for HDF compared with HD at typical dialysis and hemodiafiltration treatment settings. These predictions are similar to the HDF advantage reported in the literature in studies where phosphate removal has been measured in spent dialysate.