Francesco Bellanti (1), Micha Levi (2), Doug Marsteller (2), Eugène Cox (1)
(1) Certara Strategic Consulting, The Netherlands, (2) Teva Pharmaceutical Industries, USA
Introduction: A significant unmet medical need remains for severe asthmatics, a patient population defined as not well controlled on inhaled corticosteroids and/or other current standard of care and experiences clinically significant asthma exacerbations and associated with those ER visits and hospitalizations. The key challenge in the development of drug with improved clinical benefit in patients with severe asthma is to establish a robust clinical development plan from proof of concept studies through phase II/III. Such a development may benefit from the ability to translate of phase I/IIa clinical signal such as Forced Expiratory Volume (FEV1), to late stage (phase III/IV) clinical outcomes such as reductions in clinical asthma exacerbation rates. Model-based meta-analysis (MBMA) of relevant clinical trial data allows evaluating such a translational opportunity.
Objectives:
- To perform an MBMA of comparative effectiveness of biologics for the treatment of moderate to severe asthma
- To develop a joint MBMA model that relates early clinical well-established endpoints (FEV1) to late stage clinical outcomes (exacerbation rate)
Methods: The MBMA was based on a comprehensive database of summary clinical outcome efficacy data from randomized controlled trials in moderate to severe asthma [1]. FEV1 and exacerbation rate data were modelled simultaneously; differences in variances of mean endpoint values between studies were accounted for using appropriate weighing schemes. Different variance structures were used for FEV1 (Gaussian) and exacerbation rate (Poisson) endpoints. The treatment effect on both endpoints was modelled as a function of drug and dose, where endpoints effects were related using a scaling factor. The mean FEV1 and (log) exacerbation rate responses were described by the sum of the placebo response and drug effect. Given presumed non-Gaussian distribution of the placebo effect, this was modelled non-parametrically by estimating study, time and endpoint specific fixed effects. The effect of continuous covariates on endpoints was tested using power functions. All analyses were performed using R (version 3.4.2) typically applying functions gnls() and nlme().
Results: A total of 24 and 21 trials (all placebo controlled) and 660 and 108 observations were included for FEV1 and exacerbation rate, respectively. Exploratory data analysis indicated that an increased FEV1 (change from baseline) effect was related to a lower exacerbation rate. A joint MBMA model of FEV1 and exacerbation rate was developed based on the observed correlation. Drug-specific effects were shared between endpoints, and the scaling factor correlating the drug effect of FEV1 and exacerbation rate was estimated to be -0.872 (RSE of 2%). Baseline exacerbation rate and baseline eosinophil levels were found to significantly affect drug effect, with estimated exponents of 0.902 (RSE of 43.6%) and 0.419 (RSE of 26.9%), respectively. The joint MBMA model was externally validated using additional biologic data that were not part of the analysis data set. The results confirmed robustness and predictive ability of the developed framework. Model-based simulations were subsequently performed to evaluate comparative treatment effects of biologics for both FEV1 and exacerbation rate endpoints.
Conclusions: A joint MBMA model of drug effect on FEV1 and exacerbation rate was successfully developed using summary clinical outcome data from randomized controlled trials in moderate to severe asthma. This joint model may be used to provide early insights on expected late stage clinical response (exacerbation rate) for new drugs based on FEV1 response in early Phase I/IIa studies. As such, this predictive framework can support timely decisions (e.g. Go/noGo criteria, dose/regimen selection, and study design optimization) in the clinical development programs of biologics for the treatment of moderate to severe asthma.
References:
[1] https://www.certara.com/wp-content/uploads/Resources/Brochures/BR_ClinicalTrialOutcomeDatabases.pdf
Reference: PAGE 28 (2019) Abstr 8820 [www.page-meeting.org/?abstract=8820]
Poster: Drug/Disease Modelling - Other Topics