Kt/V: patients do not get what the physician prescribes.

Monthly urea kinetic modeling is performed [service Kt/V (urea)] to ensure that dialysis prescriptions provide patients a Kt/V greater than or equal to 1 and yield a protein catabolic rate (PCR) greater than or equal to 0.8. The frequency with which the dialysis prescription (physician's order +/- 5%, p +/- 5%) was achieved was calculated by three methods: 1) CompuMod (3 ureas; computer derived), 2) Jindal-Goldstein, and 3) Daugirdas, (2 and 3% reduction of urea). Ten patients were followed serially over 1 month for a total of 120 dialyses. Mean Kt/V values for each method were: prescription, 1.54 +/- 0.36; service, 1.40 +/- t0.63; CompuMod, 1.33 +/- 0.27; Jindal-Goldstein, 1.55 +/- 0.24; and Daugirdas, 1.33 +/- 0.23. The percentages of dialyses within the p +/- 5% were 12.4%, CompuMod; 12.8%, Jindal-Goldstein and 14.3%, Daugirdas. The percentages above p +/- 5% were 20.4%, CompuMod; 47%, Jindal-Goldstein; and 21.4%, Daugirdas. The percentages below p +/- 5% were 67.3%, CompuMod; 40.2%, Jindal-Goldstein; and 64.3%, Daugirdas. The CompuMod and Daugirdas methods of assessment of Kt/V were significantly lower (p less than 0.001) than the prescribed Kt/V, whereas the Jindal-Goldstein estimate was not. The authors conclude that dialysis patients rarely achieve their prescribed Kt/V. The service Kt/V, therefore, is not a useful parameter for prescribing dialysis therapy. The CompuMod and Daugirdas methods are the best estimates of the Kt/V, while the Jindal-Goldstein equation overestimates the Kt/V. The need for frequent urea kinetic modelling is stressed. An online urea monitor for each dialysis would be the ideal solution.