Differences between KT/V measured during dialysis and KT/V predicted from manufacturer clearance data

We retrospectively analyzed data from 3,863 dialysis treatments in 329 end-stage renal disease patients over a period of 33 months to evaluate the accuracy of in vitro KT/V estimated by manufacturer's urea clearance data in relation to in vivo measured KT/V. In 1,087 urea clearances measured, mean actual clearance was 87% of predicted. At all blood flows, actual clearances were significantly lower than predicted (8-16% lower than predicted). In 2,807 KT/V measurements, predicted KT/V was 1.238 +/- 0.005 whereas the mean of actual measured KT/V was 16% lower or 1.024 +/- 0.005 (P less than 0.0001). At different blood flows and with different dialyzers, predicted KT/V overestimated actual values. With increasing numbers of reuse, actual/predicted clearance ratios and actual/predicted KT/V ratios progressively dropped. Prescribing dialysis treatments using manufacturer's in vitro generated clearance data can lead to marked underdialysis of patients.