A mechanism-based pharmacokinetic-pharmacodynamic model of the time-course of prolactin following antipsychotic drug treatment
Friberg, Lena E.(1), An M. Vermeulen(2), Mats O. Karlsson(1)
(1)Division of Pharmacokinetics and Drug Therapy, Uppsala University, Sweden; (2)Advanced PK/PD Modeling & Simulation, Johnson & Johnson Pharmaceutical Research and Development, Beerse, Belgium
Objective: Antipsychotic drugs interact with dopamine receptors and can cause release of prolactin into the circulation which may lead to unwanted side-effects. The aim was to develop a pharmacokinetic-pharmacodynamic (PKPD) model characterising the time-course of prolactin in healthy and schizophrenic subjects following administration of various doses and formulations of antipsychotic drug treatment.
Methods: A total of 1462 subjects with 9022 prolactin observations from five Phase 1 studies (rich data) and four Phase 3 studies (sparse data) were available for the analysis. The antipsychotic drugs were administered as oral immediate release, oral extended release, i.m. injection or i.v. infusion as a single dose or as repeated once-daily dosing. Two thirds of the data (index data set) was used for building the model in NONMEM. Because of extensive run times, simulated drug concentrations were added to the data set, which were allowed to drive the PD model and the FO method had to be used. The remaining one third of the data (evaluation data set) was used for model evaluation.
Results: A spare-receptor model described the interaction of hypothetical dopamine concentrations with dopamine receptors for prolactin release, mimicking the tonic inhibition by dopamine. An agonist-antagonist interaction model illustrated the competition between the drug and dopamine for the receptors . Increased prolactin stimulated dopamine production and explained the tolerance development. In contrast, the pool model  resulted in poor population predictions. The diurnal rhythm of prolactin was described by two cosine functions, where patients had a phase-shift of –1.45 h. The typical baseline value of prolactin was significantly higher in schizophrenic women (29.5 ng/mL) than in schizophrenic men (13.7 ng/mL) and in male healthy volunteers (6.8 ng/mL). The estimated half-life of circulating prolactin was 32 minutes, in line with literature values . The negative correlation (-0.86) between baseline variability and variability in the potency parameter for drug relative to dopamine indicated a smaller response in subjects with a high baseline than in those with a low baseline. The evaluation data set was well predicted by the model.
Conclusion: The developed model accurately described the prolactin concentrations after different forms of administration and was based on known mechanisms of dopamine release with parameter estimates of prolactin kinetics close to literature values.
 Bagli M et al. J Pharmacol Exp Ther; 1999; 291:547-54.
 Movin Osswald G, Hammarlund Udenaes M. J Pharmacol Exp Ther; 1995; 274:921-7.
 Cooper DS et al. J Clin Invest; 1979; 64:1669-1680.