Could prolactin levels be a more informative predictor for clinical effect of D2-receptor antagonists than drug concentrations in the treatment of schizophrenia?
Klas J. Petersson(1), An M. Vermeulen(2), Lena E. Friberg(1)
(1) Department of Pharmaceutical Biosciences, Uppsala University, Sweden (2) Advanced PK/PD Modeling and Simulation, Johnson & Johnson Pharmaceutical Research and Development, a Division of Janssen Pharmaceutica N.V., Beerse, Belgium
Objectives: In a setting with many simultaneously ongoing processes and high variability like in schizophrenia a clear drug concentration-effect relationship can be hard to establish. Disease progression, placebo response, compliance and dropouts are issues important to consider but difficult to characterize. The clinical effects of antipsychotics are due to their antagonistic effects at the D2-receptor, which also gives rise to side effects such as elevated prolactin levels. The objective of the present study was to use a model-based approach to investigate if increases in prolactin concentrations are better predictors of clinical response, measured using PANSS, than drug concentrations.
Methods: Data were from 1187 patients with acute schizophrenia, pooled across 3 phase III trials, who were treated with placebo or paliperidone for 6 weeks. The prolactin data was previously modeled with an agonist-antagonist interaction model in which a potency parameter for individual prolactin release after treatment was estimated [1]. PANSS data from the same patients were subjected to modeling in NONMEM VI using a bilinear placebo model where the shift between the two slopes was estimated by a sigmoid Emax function:
shift=tγ/(tγ+breakγ),
Panss=(1-shift)*(base1+slope1*t) + shift*(base2+slope2*t)
Individual potency parameters, predicted change in prolactin concentrations from baseline, and predicted drug concentrations were tested as predictors for clinical effect of paliperidone.
Results: In the placebo group, the typical early decrease in PANSS was 0.7 units/day and in the later phase PANSS increased by 0.17 units/day. The shift between the first and second phase occurred around day 10. γ was estimated to 1.8 which gives a smooth transition between the slopes. For paliperidone treated patients slope1 was correlated to the relative prolactin increase from the baseline model. For slope2, drug concentrations, treatment-no treatment effect, and the prolactin potency parameter gave similar fits to the data. The best predictors were prolactin elevation at the time of PANSS measurement or AUC of elevated prolactin during the preceding 24 hours.
Conclusions:The prolactin elevation of antipsychotic treatment was here shown to be a better predictor of clinical efficacy than drug concentration. This could be expected since prolactin elevations contain information both about the individual sensitivity to D2-antagonists and compliance.
References:
[1] Friberg et al. An Agonist-Antagonist Interaction Model for Prolactin Release Following Risperidone and Paliperidone Treatment. Clin Pharmacol Ther. 2008 Dec 24. [Epub ahead of print]
