Satyawan B. Jadhav (1), Ryan L. Crass (1), Benny M. Amore (2), Sunny Chapel (1), William J. Sasiela (2), Maurice G. Emery (2)
(1) Ann Arbor Pharmacometrics Group, Ann Arbor, MI, USA; (2) Esperion Therapeutics Inc., Ann Arbor, MI, USA
Objectives: Bempedoic acid is an oral small molecule approved in February 2020 by the United States Food and Drug Administration as an adjunct to diet and maximally tolerated statin therapy for adults with heterozygous familial hypercholesterolemia or established atherosclerotic cardiovascular disease. Bempedoic acid lowers low-density lipoprotein cholesterol (LDL-C) through inhibition of adenosine triphosphate citrate lyase (ACL), an enzyme upstream of 3-hydroxyl-3-methylglutaryl Coenzyme A (HMG-CoA) reductase in the cholesterol biosynthesis pathway. Small mean elevations in serum creatinine were observed in Phase 2 studies and were further evaluated in Phase 3 trials. The objective of this work was to develop a population exposure-response model to quantitatively describe the serum creatinine change from baseline with bempedoic acid exposure using pooled data across the clinical development program.
Methods: Data were pooled from 15 studies (three Phase 1, eight Phase 2, four Phase 3) including placebo and 10 unique bempedoic acid doses ranging from 20 to 240 mg administered once daily. Model development was performed using NONMEM®. A population pharmacokinetic model consisting of two-compartment disposition, transit compartment absorption, and first-order elimination was previously developed. Subject-level pharmacokinetic information was propagated into exposure-response models via individual post-hoc parameter estimates for apparent drug clearance (CL/F), apparent central volume of distribution (V2/F), and the first-order absorption rate (Ka). Following identification of an adequate base exposure-response model, including relevant inter-individual and residual random effects, a full covariate model was evaluated by including all pre-specified covariate effects simultaneously. Covariate reduction was performed using the Wald’s Approximation Method to identify a parsimonious final model [1]. Evaluation of the final model was subsequently performed using a visual predictive check.
Results: A total of 28806 serum creatinine measurements from 4483 subjects were included in model development. A direct effect, sigmoidal Emax model adequately described the change in serum creatinine from baseline with bempedoic acid concentration and time. The maximum effect (Emax) was estimated as a 6.0% (95% CI: 4.9%, 7.0%) increase in serum creatinine from baseline with 50% of the maximal response (EC50) at a bempedoic acid concentration of 8.37 μg/mL (95% CI: 7.37, 9.37 μg/mL) in the final model. Age, weight, sex, Black race, and estimated renal function were all significant covariates on estimated baseline serum creatinine concentration in the final model; however, none were predicted to have clinically meaningful effects on response (0.8 to 1.25-fold change relative to reference).
Conclusions: The effect of bempedoic acid exposure on serum creatinine was adequately described using a sigmoidal Emax model. The magnitude of maximum change in serum creatinine from baseline with bempedoic acid is unlikely to be clinically meaningful.
[1] Kowalski KG, Hutmacher MM. Efficient screening of covariates in population models using Wald’s Approximation to the likelihood ratio test. J. Pharmacokin. Pharmacodyn. 2001;28:253-276.
Reference: PAGE () Abstr 9293 [www.page-meeting.org/?abstract=9293]
Poster: Drug/Disease Modelling - Safety