Late Breaking Abstract - Blood eosinophil count (EOS) can accurately predict responsiveness to inhaled corticosteroids (ICS) in COPD, but only if measured while patients are not receiving steroids.

Abstract

It is unclear to what extent EOS can predict responsiveness to ICS in COPD. As steroid administration can suppress EOS, we hypothesised that EOS measured while patients are not receiving steroids (EOS off steroids), can better predict responsiveness to ICS.

We compared EOS off steroids (for at least 8 weeks) versus EOS on ICS in a post-hoc analysis of ISOLDE, a 36-months, double-blind trial comparing inhaled fluticasone propionate (500mcg twice daily,MDI) versus placebo in 751 patients with moderate or severe COPD. We used Wilcoxon signed-rank test to assess whether ICS administration suppresses EOS and mixed methods repeated measures methodology (MMRM) to assess whether EOS can predict responsiveness to ICS with regards to pulmonary function.

ICS significantly suppressed EOS count (p=0.0012, 77% of those with EOS off steroids ≥ 0.2, had EOS <0.2 while receiving ICS). MMRM modelling demonstrated that EOS off steroids can accurately predict responsiveness to ICS (p<0.0001), whereas EOS on ICS was not predictive (p=0.95). Our model suggests that increasing EOS (while off steroids) is associated with better ICS response; ICS limited the mean annual rate of post-bronchodilation FEV1 decline by 0, 30, 65 and 80 mLs for patients with EOS off steroids of 0, 0.2, 0.4 and 0.5 eosinophils/cubic millimetre, respectively.

In conclusion, EOS can accurately predict responsiveness to ICS in COPD, provided that it is measured while patients are not receiving steroids. Further data is required to identify a clinically relevant cut-point.

We thank GlaxoSmithKline and ClinicalStudyDataRequest.com for providing access to ISOLDE data.