CAN BIOAVAILABILITY OF LOW-VARIANCE DRUGS BE ESTIMATED WITH AN UNPAIRED, SPARSE SAMPLING DESIGN

Objective: Bioavailability (F) with nonintravenous administration is t raditionally estimated by comparison of the area under the plasma conc entration versus time curve (AUC) after drug administration by each of the nonintravenous and intravenous routes in the same individual, Thi s paired approach may not always be possible, We simulated whether F a nd absorption rate constant (k(a)) could be estimated accurately for a drug with low variance using different patients for nonintravenous an d intravenous routes and whether sparse sampling permitted accurate es timates. Methods: Using pharmacokinetic parameters for cisatracurium b esylate (INN, cisatracurium besilate), we simulated data sets represen ting 20 administrations (10 intravenous and 10 nonintravenous) with ei ther three (sparse) or 16 (extensive) samples per administration. Simu lations were performed twice, with k(a) values of 0.1 (slow absorption ) or 0.3 (rapid absorption) min(-1). With use of NONMEM, we estimated F and k(a) for each data set using both two-stage and mixed-effects mo deling approaches and paired and unpaired designs to determine the per centage of estimates that deviated >25% from the simulated value, Resu lts: Estimates of F with extensive data were satisfactory for all appr oaches, With sparse sampling, two-stage analysis of unpaired data was not possible, two-stage analysis of paired data yielded erroneous esti mates, and mixed-effects modeling gave satisfactory estimates. Estimat es of k(a) were sometimes erroneous with all approaches except for pai red analysis of extensive data with slow absorption; sparse data and t wo-stage analysis increased the likelihood of errors compared with ext ensive data and mixed-effects modeling. Conclusions: Mixed-effects mod eling facilitates estimation of F and k(a) for low-variance drugs in s ituations in which traditional paired extensive data designs are not p ossible.