Comparison of modeled versus reported phosphate removal and modeled versus postdialysis serum phosphate levels in conventional hemodialysis.

Background: We compared predictions of phosphate removal by a 2-pool kinetic model with measured phosphate removal in spent dialysate as reported by others.

Methods: Twenty-six studies were identified that reported phosphate removal in 35 groups of patients. In almost all studies, patients were dialyzed for close to 4 h (range 3 to 6 h). For each study, group mean values of predialysis serum phosphate, body size, dialyzer K₀ A urea, blood and dialysate flow rates, and session lengths were input into the kinetic model. Predictions of group mean phosphate removal and postdialysis serum phosphate were compared with reported measured values.

Results: Mean (by patient group) predicted phosphate removal was 931 ± 170 mg/treatment, somewhat higher (p < 0.001) than the reported measured value, 900 mg ± 287. The ratio of predicted/measured removal averaged 1.15 ± 0.427. In 5/35 patient groups (3/26 studies) the predicted/measured phosphate removal was greater than 1.50. If these groups were excluded, the mean measured phosphate removal was 990 mg versus 966 predicted, with a ratio of predicted/measured removal averaging 0.993. Measured group mean postdialysis serum phosphate values (reported in 25/35) were 2.64 ± 0.54, not significantly different from predicted (2.60 ± 0.24 mg/dl, p = NS).

Conclusions: For conventional 4-h dialysis treatments, phosphate removal and postdialysis serum phosphate values predicted by a 2-pool kinetic model are similar to reported measured values.