Pool Model versus Agonist-Antagonist Interaction Model for the Remoxipride Effect on Prolactin
Guangli Ma, Lena E. Friberg, Mats O. Karlsson
Division of Pharmacokinetics and Drug Therapy, Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden
Objectives: The pool model  has previously been shown to be inferior to an agonist-antagonist interaction (AAI) model [2, 3] to describe prolactin response following administration of two antipsychotic drugs . This study aimed to compare the pool model and the agonist-antagonist interaction (AAI) model to describe prolactin concentrations after administration of the antipsychotic drug remoxipride, i.e. the data the pool model was originally developed from.
Methods: The remoxipride and prolactin concentration data were from 8 healthy male volunteers . There were 5 study occasions and on each occasion two 0.5 h infusions of remoxipride were administered. The intervals between the first dose and the second dose on the 5 occasions were 2, 8, 12, 24 and 48 hours.
Five models were compared in NONMEM and by visual predictive checks; (1) the original pool model , (2) a pool model with enforced mass balance, (3) a pool model with enforced mass balance and a circadian rhythm function for prolactin release, (4) the AAI model , and (5) the AAI model with circadian rhythm .
Results: The AAI model had 85 units lower OFV than the pool model with mass balance, while the pool model had 1 less THETA and 1 less ETA than the AAI model. Addition of a circadian sub-model improved both the pool model and the AAI model. A VPC revealed that the circadian pool model failed to adequately predict the prolactin profile after remoxipride administration.
The pharmacodynamic parameters estimated by the circadian AAI model were in line with previous studies and current understanding about prolactin. The prolactin rhythm predicted by the circadian AAI model was close to reports in the literature.
Conclusions: As previously observed for other antipsychotic drugs , the circadian AAI model was superior to the other investigated models in describing the prolactin response after remoxipride administration. The AAI model appears to work well across drugs and for a range of different types of administration schedules.
 Movin Osswald, G. and M. Hammarlund Udenaes. J Pharmacol Exp Ther; 1995; 274: 921-7.
 Bagli, M., et al. J Pharmacol Exp Ther; 1999; 291: 547-54.
 Friberg, L.E., et al. PAGE 15 (2006) Abstr 940 [www.page-meeting.org/?abstract=940].