Satoshi Shoji (1), Akiyuki Suzuki (1), Puneet Gaitonde (2), Peter Lockwood (2), Chun-Hua Cai (2) & Scott Marshall (3).
(1) Pharmacometrics Pfizer R&D Japan (2) Clinical Pharmacology Pfizer Inc. (3) Pharmacometrics Pfizer R&D Ltd.
Introduction: Tanezumab is a monoclonal antibody that binds to and inhibits the actions of nerve growth factor (NGF). The NGF inhibitor class may offer an important breakthrough in the treatment of chronic pain and is under clinical investigation for the treatment of pain associated with osteoarthritis (OA) or other chronic pain conditions. Model-informed drug development (MIDD) approach including population PK, PK/PD, target-mediated drug disposition modeling analyses have been incorporated across the drug development program [1-5]. This analysis describes an integrated population PK analysis of intravenous (IV) and subcutaneous (SC) data from patients with OA and chronic low back pain (CLBP) in pivotal Phase 2b (P2b) and Phase 3 (P3) studies.
Objectives: The main objectives were 1) to create IV/SC joint PK model across the P2b and P3 studies, conditioning on prior characterization of factors impacting drug disposition, 2) to evaluate the factors impacting the absorption characteristics of tanezumab following SC administration across emerging P3 SC OA and CLBP studies, and 3) to investigate the impact of baseline NGF and Treatment-Emergent anti-drug antibodies (TE ADA) patient status on tanezumab PK.
Methods: The model building was conducted based on the following two-step strategy, 1) analysis following both IV and SC administration based on a previous analysis of IV OA studies [5], 2) inclusion of the more recent P3 SC OA and CLBP studies into the model from 1). The potential impact of NGF levels, TE ADA, and ADA titer was assessed by incorporating the covariate effect into the final model. These analyses used the non-linear mixed-effects modeling approach. The software packages NONMEM and Perl-speaks-NONMEM (e.g., VPC, SCM, and SIR) and R software were also used.
Results: A two-compartment model with parallel linear and non-linear eliminations and first-order absorption adequately described the plasma tanezumab concentration versus time data following tanezumab IV and SC administration across 4646 patients over 18,000 concentrations measurements. The estimated disposition parameters CL, Vc, Vp, and KM (RSE%) were 0.133 (0.961) L/day, 2.6 (0.733) L, 1.77 (1.26) L and 31.2 (24.1) μg/L. The bioavailability (F1) following SC administration into abdomen for the more intensively sampled study was estimated to be 73% and 76% compared to 62% and 65% in the more sparsely sampled studies for females and males, respectively. Steady-state simulations for the 8 weekly regimen indicated that despite the lower bioavailability and Cmax, the tanezumab concentrations measured from Week 4 post-dose were similar for the IV and SC routes of administration.
The final model included effects of WT, dose, CLcr and SEX on CL, WT and SEX on Vc, and WT on Vp. In addition, injection site, AGE, SEX and WT were estimated to impact the extent of absorption (F1), while AGE, SEX and WT were also estimated to impact the rate of absorption (ka) following SC administration. WT was the only factor to influence CL and Vc with a greater than +/-20% change across the WT distribution. A small 14% decrease in CL was observed with CLcr down to the 30 mL/min. In terms of ka, WT and SEX led to a greater than +/-20% change, while WT, AGE, SEX and site of injection (abdomen or thigh) also have a limited impact on F1. Baseline NGF had a minimal but statistically significant effect on VM (maximum elimination capacity). The TE ADA status was associated with a 7% increase in CL. A model with ADA titer predicted 6% increase in CL at the median of positive titer levels.
Conclusions: This IV/SC joint PK model adequately described the plasma tanezumab concentration versus time data following tanezumab IV and SC administration in patients with OA or CLBP. WT was the most important covariate. This is consistent with the original IV OA population PK estimation where WT was also the only covariate to significantly impact inter-individual variability (IIV) reducing it from 35% to 27% with other covariates only reducing IIV by another 1% [5]. This analysis supports the conclusions from prior analyses that there was no need to adjust the dose of tanezumab across the range of demographics (WT, Gender, Age, Race, Renal function) or choice of injection site. Baseline NGF had a minimal effect on VM. The TE ADA status was associated with a slightly higher clearance of tanezumab, but there was no evidence of a clinically meaningful difference.
References:
[1] Xie R, Arends R, Olson S, & Marshall S. Population PK/PD Analysis of the Effect of Tanezumab on Overall Daily Pain Score Data in Adults with Moderate-to-Severe Pain due to Osteoarthritis of the Knee. PAGE 18 (2009) Abstr 1561.
[2] Cleton A, Lalovic B, Xie R & Arends R. Population Pharmacokinetics of a Monoclonal Antibody Tanezumab in Chronic Pelvic Pain Conditions. PAGE 20 (2011) Abstr 2197.
[3] Jonsson N, Arends R, Xie R & Marshall S. Population PKPD analysis of weekly pain scores after intravenously administered tanezumab, based on pooled Phase 3 data in patients with osteoarthritis of the knee or hip. PAGE 22 (2013) Abstr 2898.
[4] Shoji S, Marshall S, Xie R, & Arends R. Slow drug-target complex kinetics and first dose overestimation of free target suppression in target-mediated drug disposition (TMDD) approximation models: An evaluation for tanezumab a NGF antibody for treatment of pain. PAGE 26 (2017) Abstr 7121.
[5] Jonsson N, Xie R, Marshall S & Arends R, 2016, Population pharmacokinetics of tanezumab in phase 3 clinical trials for osteoarthritis pain. Br J Clin Pharmacol vol. 81:688–699.
Reference: PAGE 29 (2021) Abstr 9595 [www.page-meeting.org/?abstract=9595]
Poster: Drug/Disease Modelling - CNS