Investigation of the impact of population parameters describing gastric emptying on bioequivalence metrics
Karatza Eleni , Karalis Vangelis
Department of Pharmacy, School of Health Sciences, National and Kapodistrian University of Athens, Greece
Introduction: Gastric emptying was shown to significantly affect the disposition of losartan (LOS), which is a BCS class I compound and its active metabolite (EXP-3174) [1]. This phenomenon was also noted to affect the bioequivalence outcome of BCS class I and III compounds [2].
Objectives: In the first place this study aimed to the validate the two models developed in order to describe gastric emptying using delay differential equations (DDE) or a sinusoidal equation. In the second place through this work, we intended to investigate which among their parameters may significantly affect bioequivalence (BE) metrics, namely Cmax and AUC.
Methods: LOS and EXP-3174 plasma concentration profiles were obtained from a single dose, 2x2 bioequivalence study in 31 men and women, receiving 100mg (2x50mg) losartan potassium in the form of immediate release tablets. Following a population pharmacokinetic analysis using Monolix ® 2018R1 the disposition of LOS and metabolite EXP-3174 were described by a joint two compartment-one compartment model with delayed first order metabolite formation. Plasma oscillations noted in certain losartan C-t profiles were attributed to gastric emptying that was best modeled either by first order gastric emptying followed by delayed first order absorption constant (Delay-model) or by a sinusoidal equation describing gastric emptying followed by first order absorption (Sinus-model) [3]. Matlab® DDE solver (dde23) was used to solve the equations obtained using parameters estimated by population modeling. The C-t profiles obtained were compared to the empirical C-t profiles using Wilcoxon paired rank test. Principal component analysis (PCA) was performed with the individual parameter estimates obtained from each model separately and the respective bioequivalence metrics (Cmax, AUC and Tmax) of each volunteer using R.
Results: Predicted concentrations obtained were similar to the empirical concentrations in both cases examined. Indeed, there was no statistically significant difference of the mean empirical C-t profile and solution of the Delay-model (p=0.2744) or solution of the Sinus-model (p=0.5966) using the population estimates. This finding was also evident through the superposition of the three C-t profiles obtained. Using PCA in both cases five principal components (PC) were identified explaining 82% of cumulative variability using individual parameter estimates of the Delay-model and 80% of cumulative variability using individual parameter estimates of the Sinus-model. Results from PCA using individual model parameters derived from the Delay-model and bioequivalence metrics showed through loadings of PC1 that constant lag time (0.4543) was significantly correlated with Cmax (-0.8959) and AUC (-0.8151).Additionally, through loadings of PC3 gastric emptying rate constant (0.7961) and delayed first order absorption constant (0.8235) were significantly correlated with Tmax (-0.8213). Results from PCA using individual model parameters derived from the Sinus-model and bioequivalence metrics showed through loadings of PC1 that amplitude of gastric emptying (-0.3630) and 2π/period of gastric emptying (-0.4286) were significantly correlated with AUC (-0.9018) and Cmax(-0.8279). In addition, through loadings of PC3 amplitude of gastric emptying (-0.8365), 2π/period of gastric emptying (-0.3004) and first order absorption rate constant (0.8507) were significantly correlated with Cmax (0.3999) and Tmax(-0.6963).
Conclusions: These findings indicate that gastric emptying may be modeled efficiently using both approaches proposed herein. A significant effect of parameters describing gastric emptying on bioequivalence metrics was noted suggesting that this phenomenon may affect the outcome of a bioequivalence study using BCS class I compounds
References:
[1] Karatza E and Karalis E.A joint model describing gastric emptying driven pharmacokinetics of losartan and its active metabolite EXP-3174 based on data from two bioequivalence studies,AAPS annual meeting and exposition 04-07 November 2018, Washington DC (USA)
[2] Talattof A, Price JC, Amidon GL. Gastrointestinal Motility Variation and Implications for Plasma Level Variation: Oral Drug Products. Mol Pharm. 2016 Feb 1;13(2):557-67.
[3] Karatza E and Karalis E.Mathematical modeling of gastric emptying: a joint model for losartan and its active metabolite EXP-3174 (II-68). Page 2018 Montreux Switzerland
