Elisa A.M. CALVIER (1), Jean-Marie Martinez (1), David Fabre (1), Aurélie Brunet (1), Yan-Yan Zhang (2)
(1) Sanofi, 371 rue du Pr. J. Blayac, 34184 Montpellier Cedex 04, France, (2) Sanofi, 19f Tower III - Kerry Center - 1228 Middle Yan'an Rd - Jing An District, Shanghai, 200040 Shanghai, China
Objectives:
Alirocumab is a fully human monoclonal antibody directed against Proprotein Convertase Subtilisin Kexin type 9 (PCSK9), which is mainly expressed in the liver. PCSK9 directly binds to LDL-R (LDL-Receptor) and promotes LDL-R internalization and degradation. By blocking PCSK9 mediated regulation of LDL-R, alirocumab lowers serum LDL-C levels in a dose-dependent manner. Clinical studies did not evidence any impact of ethnicity on the pharmacokinetics of alirocumab. PRALUENT® (alirocumab) is currently approved in more than 60 countries worldwide, including Japan, Hong Kong and Taiwan, but not yet in China. Therefore, the objectives were to:
- Qualify a population pharmacokinetic (popPK) model for alirocumab in the Chinese population.
- Compare alirocumab exposure parameters between Chinese and non-Chinese patients
Methods:
A popPK model for total alirocumab (i.e., free and bound to PCSK9) previously developed on pooled clinical data of phase I, II and III studies (n patients ≈ 3000, mostly Caucasians) [1] was used for the analysis of a pooled dataset from phase I and III studies in Chinese patients (n = 68) using NONMEM prior’s subroutine [2,3]. Chinese healthy volunteers (n=28) received a 75, 150 or 300mg single subcutaneous (SC) dose and Chinese high cardiovascular risk patients with hypercholesterolemia on background statin therapy (n=40) received an SC 75mg dose every 2 weeks (Q2W) over 24 weeks, with a possible up-titration to 150mg Q2W at week 12. The popPK model was a 2 compartment-model with a first order absorption rate and bioavailability (i.e., SC administration), a linear as well as a non-linear (i.e., Michaelis-Menten equation) clearance. Four covariates were included: weight and statin co-administration on clearance, age on peripheral volume of distribution and free-PCSK9 on the Michaelis–Menten constant. Informative priors were first only used on bioavailability (structurally unidentifiable parameter). No priors were used for the remainder of the parameters. Informative priors were then iteratively added on parameters that were not precisely estimated or to stabilize minimization. Model validation included goodness of fit plots, bootstrap (n samples =1000), VPC (n simulations = 100) with stratification on dose and clinical study and derivation of quality criteria using in house-tools [4]. The quality criteria were namely the average fold error (AFE), the root mean squared error (rmse%) and the mean prediction error (mpe%) expressed as percentage of mean concentrations. Finally, the qualified model was used to derive alirocumab individual Cmax and AUC0-336h at steady state in Chinese patients with statins as add-on therapy receiving a 75mg dose Q2W (n=36). These parameters were compared with those previously found in non-Chinese patients under the same dosing regimen: Caucasians/Blacks (n= 317) and non-Chinese Asians (n=14).
Results:
Informative priors on all model parameters (population PK, covariate, inter-individual variability and residual error parameters) were required for model stability and precision of parameter estimates. Parameter estimates in the Chinese population were close to those previously found in non-Chinese, with a maximal difference of -10 % on the variance of the Michaelis-Menten constant. Individual and population predictions were in good agreement with observations and did not reveal any trends with observed concentrations or time. The AFE, rmse% and mpe% were, respectively, 1.56, 60.7% and 9.58% for population and 1.17, 25.9% and 3.49% for individual predictions. For all model parameters, mean estimates obtained in the final model were very close to the bootstrap values estimated from the 987 successful runs and were always included in the [2.5th – 97.5th] range. The VPCs showed that, whatever the study and the dosing regimen, a large majority of the observed concentrations were included in the range [5th-95th percentiles]. A median Cmax value of 6.27 mg/L and a median AUC0-336h value of 1650 mg.h/L at steady state for the dose of 75 mg as add-on therapy to statins were predicted. These values are close to those reported in non-Chinese patients (Cmax of 7.20 and 7.20 mg/L and AUC0-336h of 1870 and 1880 mg.h/L in Caucasians/Blacks and non-Chinese Asians respectively).
Conclusions: A popPK model for alirocumab in the Chinese population was qualified and allowed for the estimation of individual Cmax and AUC0-336h at steady state. These exposure parameters were close to those reported in the non-Chinese population.
References:
[1] J.-M. Martinez, A. Brunet, F. Hurbin, A. T. DiCioccio, C. Rauch, and D. Fabre, “Population Pharmacokinetic Analysis of Alirocumab in Healthy Volunteers or Hypercholesterolemic Subjects Using a Michaelis–Menten Approximation of a Target-Mediated Drug Disposition Model—Support for a Biologics License Application Submission: Part I,” Clin. Pharmacokinet., May 2018.
[2] S. L. Beal, L.B. Sheiner, A. Boeckmann, & R.J. Bauer, “NONMEM User’s Guides.” Icon Development Solutions, Ellicott City, MD, USA, 2009.
[3] P. O. Gisleskog, M. O. Karlsson, and S. L. Beal, “Use of Prior Information to Stabilize a Population Data Analysis,” J. Pharmacokinet. Pharmacodyn., vol. 29, no. 5/6, pp. 473–505, 2002.
[4] P. Nolain, R. Combet, D. Marchionni, H. Speth, J.-M. Martinez, “PopkinR: a suite of Shiny applications focused on the pharmacometrics workflow.,” PAGE Meeting 2018. [Online]. Available: https://www.page-meeting.org/default.asp?abstract=8684.
Reference: PAGE 28 (2019) Abstr 8838 [www.page-meeting.org/?abstract=8838]
Poster: Drug/Disease Modelling - Other Topics