PGx7692_GSK2834425, GSK2285997, GSK2592356, GSK573719, GW685698, GW642444_Exploratory evaluation of the role of genetic variation in efficacy response to Trelegy and its component parts in COPD subjects

Trial description: The purpose of this exploratory pharmacogenetics (PGx) analysis, 209135, is to evaluate the association between genetic variation and efficacy in subjects with chronic obstructive pulmonary disease (COPD) treated with fluticasone furoate (FF), umeclidinium bromide (UMEC) or vilanterol (VI), as monotherapy or in combination, using data from 34 COPD clinical studies. The studies included in this analysis are: 200109, 200110, 200812, 200820, 201211, 201315, 201316, AC4113589, AC4115408, B2C111045, CTT116853, CTT116855, DB2113359, DB2113360, DB2113361,DB2113373,DB2113374, DB2114634, DB2114930, DB2114951, DB2116134, DB2116960, DB2116961, HZC102871, HZC102970, HZC112206, HZC112207, HZC112352, HZC113107, HZC113108, HZC113109, HZC113684, HZC115151, RLV116974. Genetic effects on efficacy will be investigated within molecule (FF alone or in combination with UMEC and/or VI, UMEC alone or in combination with FF and/or VI, or VI alone or in combination with FF and/or UMEC) by pooling subjects across studies using dosages of FF 100 or 200 mcg, UMEC 62.5 or 125 mcg and VI 25 mcg. Genetic effects will also be investigated in study CTT116855 within each treatment combination (FF/UMEC/VI 100 mcg /62.5 mcg/25 mcg, FF/VI 100mcg /25 mcg or UMEC/VI 62.5 mcg/25 mcg). HZC115151 will not be included in the pooled analyses as it is effectiveness rather than an efficacy study. It will be examined for trends for those significant results arising from the pooled study molecule analyses or CTT116855 combination analyses.Primary endpoints include annual rate of on-treatment moderate/severe exacerbations, annual rate of on-treatment moderate/severe exacerbations in the subset of subjects with a baseline blood eosinophil count ≥150 cells/µL, proportion of subjects with Composite Clinically Important Deterioration (CID) including St George’s Respiratory Questionnaire (SGRQ) score at week 24, change from baseline in trough forced expiratory volume in 1 second (FEV1) at week 4, and change from baseline in SGRQ total score at week 12. Secondary endpoints include annual rate of on-treatment severe exacerbations, proportion of subjects with Composite Clinically Important Deterioration (CID) including the COPD Assessment Test (CAT) score at week 24, change from baseline in trough forced expiratory volume in 1 second (FEV1) at week 52, change from baseline in SGRQ total score at week 52, proportion of responders according to SGRQ total score at week 12, proportion of responders according to SGRQ total score at week 52.The genetics data from these studies were generated at different points in time and three different arrays were used in data generation. These include the Affymetrix Axiom Precision Medicine Research, the Affymetrix Axiom Biobank array with GSK custom content v2 and the Affymetrix Axiom Biobank array with GSK custom content v1. DNA was also derived from either blood or saliva. Each array and sample type (blood or saliva) will constitute an imputation set. Each imputation set will be used separately, in conjunction with a reference set of haplotypes, to impute approximately 30 million variants across the genome including the HLA region. Following imputation and quality control, data will be merged across imputation sets. Analysis models will include an adjustment for imputation set, genetic ancestry principal components, treatment, smoking status at baseline and endpoint specific covariates. Analysis will be conducted on genome-wide variants with a minor allele frequency ≥0.01 and an imputation quality score (r^2) ≥0.30. Approximately 15 candidate variants will be selected from ADME and target genes.Analyses will be conducted using a generalized linear model assuming a negative binomial distribution for annual rate endpoints, a linear regression model for change from baseline endpoints, or a logistic regression model for binary endpoints. Genetic main effects, testing for additive effects, will be generated within each molecule (pooled studies) or treatment combination (CTT116855). Estimation of genotype by treatment interactions using appropriate contrasts of linear combinations will be conducted using genetic effect size estimates obtained within the molecule or combination arms. Main and interaction effects will be generated for all genetic variants. A type 1 error rate of 0.05, adjusting for multiple genetic variants, will be maintained by splitting alpha between candidate and genome-wide variants. The significance threshold for genome-wide variants is ≤ 2.5x10^-8. The significant threshold for candidate variants will be 0.025/effective number of independent variants. No adjustment will be made for multiple endpoints or multiple analysis populations.TRELEGY® is a trademark of the GlaxoSmithKline group of companies.