Effect of misspecification of the absorption process on subsequent parameter estimation in population analysis.

A prospective simulation study has been carried out to evaluate the effect of potential misspecification of the absorption rate constant (ka) in population pharmacokinetic analysis when few to no concentration-time data were available in the absorption phase and estimation of ka was not possible. Data were simulated for 100 subjects using a one-compartment model at steady state with first-order input. Data were generated over a range of ka values: ka was misspecified in the NONMEM analysis by factors of 0.25, 0.5, 1, 2, 3, and 4. In general, clearance (CL) was typically estimated with a small, constant underprediction, regardless of the range of misspecification of ka or whether data were present in the absorption phase. The same was not true for volume of distribution (V), values were biased and sensitive to the degree of misspecification, but only when the data contained even a little information about absorption. If studies are to be designed in which information absorption is either not required or is of no therapeutic use, then blood samples could be concentrated in the postabsorption phase and the absorption input fixed according to the best a priori information available.