DIRECT COMPARISON OF 3 METHODS FOR PREDICTING DIGOXIN CONCENTRATIONS

Three methods of determining digoxin population pharmacokinetic parame ters were compared for their abilities to predict 118 measured serum d igoxin concentrations (SDCs) in 49 patients. NONMEM software (version IV) was used to generate a residual and a weighted residual for each m easured-concentration-predicted-concentration pair. Prediction error a nalysis was done by a maximum likelihood technique that accounted for several within-patient measures. Data analysis also included graphic o bservation of weighted residuals (WRES) and calculation of the mean WR ES and median absolute prediction error. A further parallel analysis w as also carried out on subpopulations with and without concurrent quin idine and congestive heart failure (CHF). Method III was without bias in all subpopulations studied and had the smallest WRES in all populat ions. Method I was without bias in the overall population, however, it underpredicted SDCs in patients receiving quinidine and in those with CHE Method II underpredicted SDCs in the overall population, those re ceiving quinidine, and in patients without CHE There were no between-m ethod differences in precision as assessed by absolute prediction erro r.