Ying Zhang (1), Michael A Tortorici (1), Dipti Pawaskar (1), Ingo Pragst (2), Ly Minh Nguyen (3), Jagdev Sidhu (4)
(1) Clinical Pharmacology and Early Development, CSL Behring, King of Prussia, Pennsylvania, USA; (2) Global Clinical Research, CSL Behring GmbH, Marburg, Germany; (3) Department of Pharmaceutical Sciences, SUNY at Buffalo, Buffalo, New York, USA; (4) Clinical Pharmacology and Early Development, CSL Limited, Parkville, Australia.
Objectives: Hereditary angioedema (HAE) is a rare, debilitating, and potentially life-threatening, autosomal dominant genetic disease caused by a deficiency in functional C1 esterase inhibitor (C1-INH). Long-term prophylactic subcutaneous administration of C1-INH is an established treatment option for patients with HAE. An analysis was conducted to assess the relationship between C1-INH functional activity levels (C1-INH(f)) and the risk of a HAE attack in patients with HAE. A Shiny application (app) was developed to generate and demonstrate the simulation results with various dosing scenarios.
Methods: A population pharmacokinetics (POPPK) analysis was performed using data from the Phase 1 to Phase 3 clinical studies of treatment with C1-INH (IV) or C1-INH (SC) treatment (total 2103 samples). A repeated time-to-event model was used to characterize the timing and frequency of attacks as a function of C1-INH(f) that enabled the C1-INH(f) to be directly related to the HAE attack event using 90 subjects from the Phase 3 COMPACT study experienced 1191 attacks and 234 censored events. Parametric model development assessed three main components; a background effect, a non-drug effect (e.g. time effect), and a C1-INH (SC) effect, which allowed for informative use of the changes in C1-INH(f). C1-INH(f) covariate effects were evaluated using the Wald’s approximation method procedure. The final model was used to simulate the absolute hazard of attacks over a wide range of C1-INH(f) values (20–120%). The hazard ratio was computed using the geometric mean of observed baseline C1-INH(f) as the reference (25.4%), compared to C1-INH(f) ranging from 25.4% to 120%. An interactive Shiny app was created using R packages to simulate and demonstrate the simulation results for C1-INH(f) and the hazard risk at various dosing scenarios.
Results: The C1-INH(f) following administration of C1-INH (SC) was adequately described by a linear one-compartment model with first-order absorption and first-order elimination, with inter-individual variability on all the parameters. The population pharmacokinetic model found body weight to be a significant covariate on clearance. The PK/PD model demonstrated a strong exposure-response (E-R) relationship, with increasing C1-INH(f) decreasing the risk of experiencing an HAE attack. The final model included 2 components, a baseline hazard and a non-linear drug effect. Age had a significant effect on the E-R relationship in terms of a higher risk of HAE attack for a given baseline plasma C1-INH(f) level. The response to treatment C1-INH on the risk of HAE was independent of age. The mean trough C1-INH(f) after subcutaneous administration was predicted to yield a 70% reduction in the relative risk of an HAE attack after 40 IU/kg dosing and an 81% reduction after 60 IU/kg dosing. Simulations based on the E-R model predicted that higher C1-INH(f) significantly lowers the risk for HAE attacks in a greater proportion of patients with maximal effect occurring near normal C1-INH(f). The Shiny app can perform the simulation efficiently and demonstrate the simulation results with the interactive and dynamic display.
Conclusions: Simulations based on data from the COMPACT program suggest that the prevention of HAE attacks is maximized when C1-INH(f) are restored to the normal range (> 70%).
References:
[1] Zuraw, BL. Clinical practice. Hereditary angioedema. New Engl. J. Med. 359(10), 1027–1036 (2008).
[2] Longhurst HJ et al. Prevention of Hereditary Angioedema Attacks with Subcutaneous C1-Inhibitor. New Engl. J. Med. 376(12), 1131–1140 (2017).
[3] Karlsson, K.E. et al. Performance of three estimation methods in repeated time-to-event modeling. AAPS J. 13(1), 83–91 (2011).
[4] Kowalski, K.G., Hutmacher, M.M. Efficient screening of covariates in population models using Wald’s approximation to the likelihood ratio test. J. Pharmacokinet. Pharmacodyn. 28(3), 253–275 (2001).
Reference: PAGE 27 (2018) Abstr 8564 [www.page-meeting.org/?abstract=8564]
Poster: Drug/Disease Modelling - Other Topics