A new approach to establishing individualized maintenance doses of phenytoin

A new method for arriving at individualized doses of phenytoin (PHT) is described. This method, designated the eta method, operates on two measurements of unbound plasma PHT. One plasma sample is collected 48 h after a single, oral PHT dose (approximately 4.5 mg/kg). From the single dose, single sample measurements of plasma unbound PHT individual estimates of intrinsic unbound PHT clearance (CLuint) can be made. With the use of CLuint and weight-adjusted mean population values of Vmax predicted plasma concentrations of unbound PHT for 11 subjects exhibited a mean prediction error (mpe) of 0.07 mg/L with 95% confidence limits of -0.11 to 0.26 mg/L and an error coefficient of variation [root mean squared error (rmse)/mean unbound concentration at steady state (Cuss)] of 0.4. The second plasma unbound PHT is collected at steady state. From CLuint and one Cuss measurement, Vmax and Km can be directly calculated. Doses calculated by the eta method to produce plasma unbound PHT concentrations of approximately 1.0 mg/L were identical to those calculated by the Bayesian method to yield plasma total PHT concentrations of approximately 10 mg/L. When those doses were administered to six subjects, the plasma unbound PHT concentrations predicted by the eta method exhibited a mpe of -0.13 mg/L, with 95% confidence limits ranging from -0.33 to 0.06 mg/L and an error coefficient of variation of 0.2. The plasma PHT concentrations predicted by the Bayesian method exhibited an mpe of -1.30 mg/L, with 95% confidence limits of -4.02 to 1.42 mg/L and an error coefficient of variation of 0.3. Values of Vmax and doses of PHT calculated to produce a Cuss of 1.0 mg/L calculated by the eta method were nearly identical to Vmax values and doses of PHT calculated by Bayesian double feedback (based on two different Css values) to produce a Css of 10.0 mg/L.