Denis Menshykau1, Jagdev Sidhu2, Laura Shaughnessy3, Rocio Lledo-Garcia4, Pinky Dua4, Marie Teil5, Akash Khandelwal1
1 UCB Biosciences GmbH, Monheim am Rhein, Germany. 2.UCB Australia Pty. Ltd, Malvern Victoria, Australia. 3.UCB BioSciences Inc, Raleigh, NC, USA 4 UCB Pharma, Slough, UK. 5 UCB Biopharma SRL, Bulle, Switzerland.
Introduction/Objectives: For women living with chronic inflammatory diseases, adequate disease control during pregnancy is crucial to ensure the best fetal and maternal health, and reduce adverse pregnancy outcomes; however, there is a paucity of knowledge regarding the impact of pregnancy on TNF inhibitor (TNFi) pharmacokinetics (PK). Certolizumab pegol (CZP; CIMZIA™) is the only Fc-free TNFi with no to minimal placental transfer and data on systemic exposure during pregnancy and postpartum.
A Pop PK analysis was conducted to assess CZP PK over the course of pregnancy and postpartum, in women with chronic inflammatory diseases.
Methods: CHERISH was a longitudinal, prospective, open-label PK Phase 1B study in pregnant women with chronic inflammatory diseases. CZP was administered subcutaneously at dosing regimens of 200 mg every 2 weeks (Q2W) and 400 mg Q2W and every 4 weeks (Q4W). PK samples were collected predose (before a dose) and postdose (1 week after a dose) throughout the pregnancy. Additionally, predose and postdose samples were collected 12 weeks postpartum. Data from 20 participants was included into Pop PK model development.
Model development was performed using a frequentist prior approach [2, 3]. Prior information was set based on a Pop PK model for CZP in adults with rheumatoid arthritis (Reference model (RM)) [5].
Cimzia population PK model development was performed in a stepwise manner in NONMEM 7.4.3 [4]:
- Prediction of individual PK parameters in CHERISH study based on the RM.
- Re-estimation ($PRIOR subroutine) of RM parameters based on CHERISH study [1].
- Inclusion of covariates related to the pregnancy stage (pregnancy trimester or postpartum);
- Evaluation of other baseline and time-varying covariates (body weight (BW), albumin (ALB), anti-drug antibody (ADA) titer etc.).
- Simulations: illustrate the effect of pregnancy and chosen covariates on Cimzia PK.
Covariate inclusion was additionally investigated from a causal inference standpoint [6]. A Directed Acyclic Graph (DAG) was constructed to delineate potential relationships between CZP concentration and covariates.
Results: A one compartment model with first order absorption (Ka=0.236 1/day) and linear elimination (CL/F=0.416 L/day) from the central compartment (V/F=7.86 L) best described the PK of CZP in the CHERISH study. IIV was included on CL/F only, and the RUV was described by a mixed model. The structural model parameters were estimated with good precision (RSE < 25%). Baseline BW was included as a covariate for on CL/F and V/F. Pregnancy trimester and log-transformed ADA titer were identified as the only significant covariates for CL/F.
The DAG analysis indicated both a direct impact of time-varying covariates (e.g., ALB) and a backdoor path from pregnancy stage to CZP concentration. Therefore, assessing the causal influence of time-varying covariates on CZP concentration necessitated accounting for the stage of pregnancy. Following the adjustment for pregnancy stage, only time varying ADA titer was included as CL/F covariate in the model. Overall, causal inference provided a theoretical foundation for inclusion of pregnancy stage into the model.
Progression through pregnancy and high ADA titer were associated with an increase in CL/F. Individuals with higher ADA titer (75th percentile) during pregnancy exhibited CL/F up to 1.43 fold higher relative to individuals postpartum with median levels of ADA titer. However, the confidence interval for the combined effect of pregnancy stage and ADA titer effects on CL/F overlapped with the CL/F range of the typical individual postpartum.
The developed model was used to perform simulations to assess the effect of pregnancy trimester on CZP exposure for 200 mg Q2W and 400 mg Q4W dosing regimens. Deviation in median CZP concentrations between pregnant and nonpregnant states increased during the third trimester of pregnancy. However, a large overlap was observed in simulated concentrations between pregnant and non-pregnant individuals.
Conclusions: This population PK model provides the first robust characterization of CZP PK during pregnancy. Progression through the pregnancy and high ADA titer were associated with an increase in CL/F, however overall, substantial overlap was observed between simulated exposure in pregnant and non-pregnant population. CZP PK changes during pregnancy do not warrant dose adjustment
[1] NCT04163016 A Study in Pregnant Women With Chronic Inflammatory Diseases Treated With Cimzia (Certolizumab Pegol) (CHERISH)
[2] Gisleskog, P.O., Karlsson, M.O. & Beal, S.L. Use of Prior Information to Stabilize a Population Data Analysis. J Pharmacokinet Pharmacodyn 29, 473–505 (2002). https://doi.org/10.1023/A:1022972420004
[3] Chan Kwong, A.HX.P., Calvier, E.A.M., Fabre, D. et al. Prior information for population pharmacokinetic and pharmacokinetic/pharmacodynamic analysis: overview and guidance with a focus on the NONMEM PRIOR subroutine. J Pharmacokinet Pharmacodyn 47, 431–446 (2020). https://doi.org/10.1007/s10928-020-09695-z.
[4] Beal SL, Sheiner LB, Boeckmann A and Bauer RJ. NONMEM user’s guide. (1989-2009), Icon Development Solutions, Ellicott City, MD, USA, 2009.
[5] NCT04740814 A Study to Assess the Pharmacokinetics of Certolizumab Pegol in Adults With Active Rheumatoid Arthritis
Reference: PAGE 32 (2024) Abstr 10944 [www.page-meeting.org/?abstract=10944]
Poster: Drug/Disease Modelling - Other Topics