Piyanan Assawasuwannakit (1), Xiaomei Chen (1), Satish Sharan (2), Lanyan Fang (2), Liang Zhao (2), Andrew C. Hooker (1), Mats O. Karlsson (1)
(1) Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden, (2) Division of Quantitative Methods and Modeling, Office of Research and Standards, Office of Generic Drugs, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, MD 20993, USA
Introduction: Long-acting injections (LAIs) are extended-release formulations which reduce dosing frequency. Therefore, LAIs could be regarded as an alternative to improve adherence. It is of particular importance where a disease could impede adherence, as seen in patients with psychiatric disorders including schizophrenia [1]. Bioequivalence (BE) studies for these antipsychotic LAIs are challenging due to their long half-lives of up to several weeks. In addition, these BE studies are required to be performed in patients due to long-term side effects, where steady state conditions are necessary to maintain effective exposure. A conventional BE study is a two-formulation, two-period, two-sequence randomized crossover study with washout between doses, which is both impractical (steady state for LAIs can take many months) and impossible (washout is not possible in patients). Therefore, a standard crossover study for LAIs involves patients having been stabilized on a reference product before a test product is administered. An alternative to crossover and parallel steady state studies could be a crossover switch study design where a switch of test product administration occurs after steady state attainment of a reference product in each subject. Area under the concentration time curve (AUC) and maximum concentration (Cmax) are measured at two intervals, the steady state interval of the reference product before the switch, and the first interval for the test product.
Objectives: In this work, we considered once-monthly paliperidone palmitate long-acting intramuscular injection for schizophrenia treatment [2]. The aim was to use a model-informed BE analysis in conjunction with the developed model [2] to determine a switch study design for LAI BE studies considering a trade-off between study duration and sample size. The specific aims were:
- (1) To establish the sensitivity of pharmacokinetic (PK) metric ratios to changes in absorption parameters due to formulation effects in crossover switch studies.
- (2) To develop surrogate BE criteria for the switch study setting.
- (3) To explore four study designs, namely 1. crossover steady state studies, 2. crossover switch studies, 3. parallel steady state studies and 4. parallel switch studies to yield a design which results in relatively shortened study duration and reasonable number of recruited subjects based on power.
Methods: (1) Simulations were performed for crossover steady state and switch designs on a typical subject where absorption parameters of a test drug were adjusted. The test-to-reference ratios of AUC and Cmax were computed and compared between these two designs. (2) Surrogate BE criteria, the ratios of geometric mean of AUC and Cmax of a test to reference drug corresponding to the lower (80%) and upper (125%) bound of the standard BE limit at steady state, were computed for the switch studies. The criteria were obtained from simulations with 500,000 subjects and formulations at the edges of the standard BE limit (80% and 125%). (3) To assess study power, a test drug was assumed to be identical to a reference drug formulation. Simulations were performed with various number of subjects for 1000 replicates in each study design. Power was calculated as the number of replicates passing BE limit divided by 1000 for AUC and Cmax.
Results: (1) The metric ratios of AUC and Cmax from the first interval after the switch compared to steady state were more sensitive to the change of all absorption parameters except relative change in bioavailability (F) between formulations. Therefore, we mainly focused on F in the switch study setting. (2) Much tighter BE limits for both crossover and parallel switch studies were determined (The crossover design: AUC 94.15%-107.04%; Cmax 95.59%-107.25%). (3) To obtain around 80% power for both AUC and Cmax, conducting crossover steady state studies required around 20 subjects, followed by parallel steady state and crossover switch studies around 160 and 240, respectively. Around 2000 subjects would be needed if parallel switch studies were conducted. Crossover switch studies greatly surpassed parallel steady state studies in study duration.
Conclusions: The crossover switch study design with surrogate BE criteria provides an alternative to overcome challenges of LAI BE studies. It requires a reasonable number of sample size, in a similar order to that of the parallel steady state study design, while substantially shortens study duration.
References:
[1] Greene et al. J Med Econ. 2018;21(2):127–34.
[2] Samtani et al. Clin Pharmacokinet. 2009;48(9):585–600.
Reference: PAGE () Abstr 9436 [www.page-meeting.org/?abstract=9436]
Poster: Study Design