ABSTRACT
Pharmacokinetic values derived by three nonlinear least-squares regression computer programs for sets of serum drug concentration data were compared. The three programs selected for comparison were the MS-DOS-based programs PCNONLIN and ADAPT and the Macintosh program BOOMER. Data on serum recainam hydrochloride concentration in 10 patients given an i.v. loading dose followed by a maintenance infusion were fitted by the appropriate models in each program. Samples had been subjected to reverse-phase isocratic high-performance liquid chromatography with ultraviolet light detection; intraday and interday coefficients of variation were less than 8% over the range of concentrations measured. A two-compartment model was used for all regressions. Each program was given identical initial estimates, and the simplex minimization algorithm of each program was used to fit the model to the data. An identical weighting scheme was used for all three programs. The mean pharmacokinetic values estimated by each program for the 10 data sets were essentially identical. Slightly larger rate constants and smaller volume terms were derived by BOOMER. BOOMER yielded the lowest weighted sum of squares and the highest correlation coefficient. A mean concentration-time plot showed that the programs all produced values that described the data very well. The three computer programs used in this analysis derived essentially the same pharmacokinetic values to describe sets of serum drug concentration data. The BOOMER program provides an acceptable alternative to MS-DOS-based programs for pharmacokinetic analysis.
