|
It is difficult to adjust phenytoin dosage in clinical situation because of its non-linear kinetics in therapeutic range. So two or more dose-steady state concentration data is needed for relatively accurate dosage adjustment by calculating individual Vmax and Km value. But because of phenytoin´s slow elimination rate, 2-3weeks or more of time is needed to reach steady state. So it is difficult to wait until the achievement of two or more does-steady state concentration data without any attempt to adjust phenytoin dosage. Therefore many methods have been advocated to aid dosage adjustment based on single phenytoin concentration measured at steady state. The predictive abilities of the two methods, Orbit graph method and Richens ＆ Dunlop nomogram method, were retropectively evaluated using data from 22 neurology outpatient of Seoul National University Hospital, whose liver function and renal function were within normal limit. Orbit graph method is a graphical approach to the Bayesian method proposed by Vozeh et al, and Richens ＆ Dunlop nomogram is a method for dosage adjustment by calculating Vmax value from fixed Km value, 3.8㎍/ml. Predictive performance of each method was evaluated by using mean error(ME) as a measure of bias, mean absolute error(MAE) as a measure of precision, root mean squared error(RMSE) as a composit measure of bias and precision, and correlation coefficient(r) as a traditional approach. As a result, Orbit graghmethod(:ME＝－0.64㎍/㎖, MAE＝2.27㎍/㎖. RMSE＝2.68㎍/㎖. r＝0.86. n＝21) was not significantly different from Richens ＆ Dunlop nomogram method (:ME＝－0.90㎍/㎖, MAE＝2.09㎍/㎖, RMSE＝2.70㎍/㎖, r＝0.91, n＝21). So it is assumed that both methods can be used for early phenytoin dosage adjustment with similar accuracy when there is only one dose-steady state concentration data.
|