PKRD: a pharmacokinetic program for least-squares and bayesian analysis of repeated-dose pharmacokinetic curves

A microcomputer program is presented which analyses multiple-dose pharmacokinetic curves using either a least-squares nonlinear analysis or a bayesian fit. The least-squares subroutine is designed to fit retrospective pharmacokinetic curves and can generate the so-called population pharmacokinetic parameters using the Standard Two-Stage method. The bayesian subroutine can instead be used prospectively to individualise the dosage regimen of a patient based on the concentrations measured in the initial phases of the drug treatment. All the pharmacokinetic subroutines of the program use the one-compartment model with either constant-rate intravenous infusion or oral route with zero-order absorption. The program was tested in a series of ten bone marrow transplantation patients treated with cyclosporine. The least-squares estimates of the one-compartment parameters, calculated by the program, were compared with those generated by the SIMKIN program.