Extended model-based meta-analysis of joint longitudinal FEV1 and exacerbation rate in randomized COPD trials
Carolina Llanos-Paez(1), Claire Ambery(2), Shuying Yang(2), Misba Beerahee(2), Elodie Plan(1), Mats O. Karlsson(1)
(1) Department of Pharmacy, Uppsala University, Uppsala, Sweden. (2) Clinical Pharmacology Modelling and Simulation, GSK, London, UK.
Introduction/Objectives: Parametric Model-Based Meta-Analysis (MBMA) is a quantitative approach that integrates relevant aggregate data (AD) from published randomized clinical trials (RCTs) to describe the effect of treatment, time and patient population characteristics on the clinical outcome of interest. A published MBMA of joint longitudinal forced expiratory volume in one second (FEV1) and exacerbation rate (ER) in patients with chronic obstructive pulmonary disease (COPD) included RCTs published up until 2013 with bronchodilators and anti-inflammatories given either as mono- (82%), dual- (17%), or triple-therapy (1%) combinations [1]. More recent studies have compared the efficacy of the inhaled triple-therapy [2,3] as well as new bronchodilators [4,5]. This study aims to evaluate the predictability of the published MBMA COPD model for FEV1 [1] and ER [6] and update this legacy MBMA with published new clinical data for COPD over a 7-year period.
Methods: AD were collected from published RCTs (2013 - 2020). The literature search, study selection, data extraction, processing and analysis were performed following the described Methods in the legacy MBMA [1]. The endpoint of interest for analysis included absolute morning trough FEV1 values and the mean annual rate of moderate or severe exacerbations per patient. Key structural components of the model as well as missing covariate values were handled as described in the legacy MBMA [1]. In this study the augmented data defined as the combination of both the new data (post-2013) and the legacy data (pre-2013) were used for analysis. Data management and exploration was performed in R software v.3.5.2 whereas data analysis and modelling were performed in NONMEM software v.7.5 together with an Intel FORTRAN compiler and Perl-speaks-NONMEM (PsN) v.5.2.6. The predictive performance of both published models for FEV1 and ER was assessed visually using goodness-of-fit plots. With respect to including covariates, decision was based on parameter plausibility and uncertainty, goodness-of-fit plots, and objective function value. Furthermore, a comparative effectiveness for FEV1 across bronchodilators and anti-inflammatory compounds was assessed with the drug effect parameters being re-parameterized as relative effects of two drugs for all comparison of interest [7].
Results: The post-2013 data include a total of 132 references comprising 156 studies. In the augmented data there is a total of 298 studies with 4,137 mean FEV1 observations for analysis. Fifty-two (17%) studies reported annual ER for each study arm contributing to a total of 135 observations. A total of 23 compounds were given across 914 study arms as mono-(71%), dual-(25%) or triple-therapy (4%). The legacy MBMA model predicted the post-2013 FEV1 data well. Therefore, no changes in the structural, statistical and covariate models were deemed necessary. Four new compounds (olodaterol, revefenacin, batefenterol and fluticasone furoate) were included in the analysis, and the typical estimated efficacy (95%CI) for the reference dose for olodaterol, revefenacin, batefenterol and fluticasone furoate are 89 mL (82.5 – 96.0), 144 mL (128.4 – 159.3), 190 mL (165.3 – 214.8) and 43.6 mL (33.4 – 53.7), respectively. A typical value (RSE%) of FEV1 baseline is 1.17 (1.0) with a coefficient of variation (CV) for inter-study variability (ISV) and inter-arm variability (IAV) of 10% and 2.1%, respectively. The typical rate of disease progression is 32 mL/per year (RSE: 9.6%), for a baseline FEV1 of 1.2 L, with an ISV CV of 54%. For most anti-inflammatories, superiority or inferiority cannot be established, with roflumilast being superior to most of the drugs. Differences among bronchodilators are more evident, with batefenterol being superior to most of the drugs.
The exacerbation model overpredicted the post-2013 mean annual ER data. Inclusion of year when the study started on the pre-treatment placebo rate improved the model predictive performance. This covariate could explain potential improvements in the disease management over time.
Conclusions: The addition of 7 years’ worth of new clinical data to the legacy COPD MBMA enabled a more robust model. This can be applied as a useful benchmark for decision-making in new drug development for COPD. Future studies should focus on combining AD and individual patient data to have a suitable model for individual outcomes predictions, and a better description of covariate effects.
References:
[1] Korell J et al. Clin. Pharmacol. Ther. 2016; 99(3): 315–324
[2] Lipson A. D et al. Am. J. Respir. Crit. Care Med. 2017; 196(4):438–446
[3] Lipson A. D et al. N. Engl. J. Med. 2018; 378(18): 1671–1680
[4] Crim C et al. BMC Pulm. Med. 2020; 20(1): 119
[5] Ferguson T. G et al. Chronic Obstr. Pulm. Dis. 2019; 6(2): 154–165
[6] Ribbing J et al. in ACoP. 2015; Abstract T-47.
[7] Korell J et al. in PAGE 23. 2014; Abstract 3095 www.page-meeting.org/?abstract=3095
Funding Information: GSK funded this research in the form of a Research payment to Uppsala University.
Acknowledgement: The computations were enabled by resources in project NAISS 2021/22-772 provided by the National Academic Infrastructure for Supercomputing in Sweden (NAISS) at UPPMAX, funded by the Swedish Research Council through grant agreement no. 2022-06725.
Conflict of Interest: CL-P, EP and MOK declare that they have no conflict of interest. CA, SY and MB are GSK employees and hold GSK shares.