Prediction of steady-state trough serum procainamide concentrations after administration of a sustained-release...

Four methods of predicting steady-state trough serum procainamide concentrations (SPC) were compared in 15 patients receiving sustained-release procainamide (Procan-SR) therapy. All methods were based on a one-compartment pharmacokinetic model. Method 1 utilized nine initial measured SPC and individualized pharmacokinetic parameters for prediction of the steady-state SPC. Method 2 utilized three SPC and individualized pharmacokinetic parameters. Method 3 utilized two SPC and an individualized apparent elimination rate constant plus other average pharmacokinetic parameters. Method 4 utilized all averaged pharmacokinetic parameters (required no initial SPC). The predicted and measured SPCs for each method were analyzed by linear regression. Regression equations and correlation coefficients (r) for Methods 1, 2, 3, and 4 were as follows: predicted SPC = 0.72 measured SPC + 1.60 (r = 0.86), predicted SPC = 0.67 measured SPC + 1.74 (r = 0.82), predicted SPC = 0.13 measured SPC + 2.57 (r = 0.58), and predicted SPC = 0.15 measured SPC + 2.47 (r = 0.52), respectively. The precision, as measured by the mean squared prediction error (95% confidence interval) for Methods 1, 2, 3, and 4 was 1.44 (0.61, 2.27), 1.75 (0.93, 2.57), 2.81 (1.3, 4.49), and 3.68 (1.12, 6.26), respectively. (Eighty-five percent of the predictions were within +/- 1.5 micrograms/ml of the measured SPC by Methods 1 and 2, as compared with 69% by Methods 3 and 4.) Bias, as measured by the mean prediction error (95% confidence interval) for Methods 1, 2, 3, and 4 were -0.44 (-0.90, 0.03), -0.34 (-0.87, 0.18), -0.43 (-1.09, 0.24), and -0.98 (-1.66, -0.31).(ABSTRACT TRUNCATED AT 250 WORDS)