Longitudinal Model-Based Meta-Analysis of Lung Function Response to Support Phase III Study Design in Chinese Patients With Asthma.

Abstract

Asthma is a chronic disease of the lungs characterized by airway inflammation, bronchoconstriction, and increased airway responsiveness. Forced expiratory volume in the first second (FEV1) is used as a measure of lung function and to help diagnose and monitor lung diseases, including asthma. An exponential longitudinal model has been previously developed to adequately describe the FEV1 response in asthma patients with placebo. This model was the basis of a longitudinal model-based meta-analysis which was undertaken to describe the trough FEV1 responses ranging up to 1year from nine clinical studies in a population with asthma (N=3,896), following placebo, dual combination (fluticasone furoate/vilanterol), and triple combination (fluticasone furoate/umeclidinium/vilanterol) given via inhalation. Numerical, graphical and simulation-based diagnostics showed that a Weibull model adequately characterized the longitudinal trough FEV1 response with time. Automatic covariate selection supported by statistically based regression models identified a range of patient characteristics influencing the model parameters. Race was a significant covariate on baseline but not on the parameters that impact the FEV1 trajectory. Based on the trough FEV1, all active treatments were found to be significantly different when compared with placebo and showed clinically meaningful improvement in FEV1. The model was able to predict the longitudinal FEV1 response in Chinese patients with inadequately controlled asthma and was used to provide additional support with respect to the design for a shorter-duration phase III study to the China National Medical Products Administration (NMPA).