Modelling interoccasion variability after repeated oral drug administration.

The concentration-time course could be described by a one-compartment body model. Drug input was modelled as a bolus dose, since no absorption phase could be distinguished in the data. Use of a more complex structural model, i.e. inclusion of further kinetic compartments, did not improve the fit to the data. No influence of gender could be detected. Drug accumulation did not take place under the dosing scheme used.

Three different error models were combined with the structural model. The base model employed an exponential error term to describe interindividual variability in the volume of distribution (V). A second model (model V) introduced an additional occasion dependent error term for the volume parameter. The third model (model F) attributed the occasion related variability to changes in bioavailibility (F). All models used a proportional error model for the residual variability.

V [l/kg] ke [1/h] ETA(V) ETA(OCC) EPS

base model 19.99 0.192 0.047 0.466

model V 25.25 0.198 0.019 0.582 0.009

model F 16.19 0.201 0.375 0.657 0.002

Model V provided the best overall fit as judged by the objective function minimum. On the other hand, interoccasion variability and interindividual variability of a pharmacokinetic parameter are expected to be of the same order of magnitude. This is readily appreciated, if repeated studies in one individual on different occasions are viewed as a special case of a population study. Moreover, after oral drug administration V is always V/F, i.e. volume and systemic availability are confounded. The true volume (as obtainable after parenteral drug administration) will probably be smaller. Model F appears to give a more realistic description of the underlying pharmacokinetics.