A Mixed-Effects Markov Model for Characterizing the Time Course of the Transitions between Different Extrapyramidal Side Effects Severity Levels
Ahmed Abbas Suleiman(1), Klas J. Petersson(1), Venkatesh Pilla Reddy(2), Johannes H. Proost(2), An Vermeulen(3), Lena E. Friberg(1)
(1) Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden; (2) Department of Pharmacokinetics, Toxicology and Targeting, University Centre for Pharmacy, University of Groningen, The Netherlands ,(3) Advanced PK-PD Modeling and Simulation, Janssen Research & Development, Beerse, Belgium.
Objectives: Extrapyramidal side effects (EPS) associated with antipsychotic drug administration are often spontaneously reported by patients and graded by clinicians into different severity levels in clinical trials. Our aim was to build a mixed-effects model capable of characterizing the time course of transition probabilities between the different states of EPS after administration of placebo and the antipsychotic drug paliperidone.
Methods: Data were obtained from three phase-III, 7-weeks long, randomized, placebo controlled studies testing paliperidone taken at doses of 3-15 mg once daily, in patients suffering from schizophrenia (placebo: n=320; paliperidone: n=867). An approach using a model with a Markov property where the probability of the different grades of EPS was modeled as compartments was applied [1,2]. Three compartments were used in this model after lumping both moderate and severe events into one compartment owing to the sparse data available for both (1=no EPS, 2=mild EPS, 3=moderate or severe EPS). The rate constants of movement between the compartments which determine the transition probabilities between the EPS states were estimated. This allows predicting all possible transitions at any time instead of estimating the probability of a transition over a fixed period of time. The analysis was performed by estimating the likelihood using the Laplacian estimation method in NONMEM 7. Various functions of increasing complexity (linear, exponential, Weibull, asymptotic, polynomial, Emax) were tested to characterize the different relationships for both the placebo and drug effects.
Results: The rate of transitioning between different probability states of EPS was shown to decrease exponentially with respect to time in the placebo group. The effect of the administration of paliperidone was added proportionally on top of the placebo effect. It was found that the rate of worsening of an EPS manifestation while taking paliperidone increases linearly with the model predicted area under the concentration-time curve. Simulations of EPS events indicated that the predicted incidence rates were similar to the observed ones.
Conclusions: The Markov property was successfully implemented in a mixed-effects compartmental model in NONMEM and was capable of characterizing the transition of the patients between different severity levels of EPS. This approach can also be used for analyzing other categorical side effect data.
References:
[1]. Lacroix, B.D.,et al., Simultaneous modeling of the three ACR improvement thresholds - 20, 50 and 70% - in rheumatoid arthritis patients treated with certolizumab pegol. PAGE 19 (2010) Abstr 1811 [www.page-meeting.org/?abstract=1811]
[2]. Bergstrand, M., et al., Mechanistic modeling of a magnetic marker monitoring study linking gastrointestinal tablet transit, in vivo drug release, and pharmacokinetics. Clinical pharmacology and therapeutics, 2009. 86(1): p. 77-83
