P142 The distribution of blood eosinophil count in a copd clinical trials database: comparing the uk with the rest of the world

IntroductionThere is accumulating evidence that blood eosinophil count may have predictive value for those individuals with COPD who are more likely to respond to an inhaled corticosteroid in terms of exacerbation reduction and there is evidence that higher blood eosinophil count can also have some predictive value for those at risk of exacerbations. Blood eosinophil counts are known to be raised in a number of conditions including allergies and parasitic or fungal infections. It is therefore possible that the blood eosinophil count would vary between countries and thus influence their predictive value. We have investigated the distribution of blood eosinophil counts in the UK in comparison with blood eosinophil counts worldwide from data in the GSK clinical trials database.MethodsIn this post-hoc analysis, the following criteria were used to select studies for consistency with analyses conducted to examine the effects of inhaled corticosteroids on outcomes: global, randomised, double-blind, parallel-group clinical trials in COPD of at least 24 weeks’ duration that included any of fluticasone propionate (FP), fluticasone furoate (FF), salmeterol/FP or FF/vilanterol (VI) as a randomised study drug and a non-steroid-containing arm and for which subjects had a pre-randomisation blood sample taken for eosinophils.1,2 Individual subjects’ pre-randomisation eosinophil counts from countries that recruited at least 100 subjects across all trials were pooled to form the global sample (Argentina, Australia, Canada, Chile, Czech Republic, Denmark, Estonia, France, Germany, Greece, Italy, Korea, Lithuania, Mexico, Netherlands, Norway, Peru, Philippines, Poland, Romania, Russia, Slovakia, South Africa, Spain, Sweden, United Kingdom, United States). Individual subjects’ pre-randomisation eosinophil counts for subjects in the UK were pooled to form the UK sample. An empirical cumulative distribution function (CDF) for the UK sample was overlaid on an empirical CDF plot for the global sample.ResultsThe blood eosinophil count in COPD patients included in these trials in the UK is very similar to that worldwide (Figure).ConclusionsThis suggests that blood eosinophil count could be used in the UK to help predict response to inhaled corticosteroids in COPD.Abstract P142 Figure 1Blood eosinophil count in COPD patients in the UK and WorldwideReferencesPavord. Lancet Respir Med 2016, in press.Pascoe. Lancet Respir Med 2015.