Abstract Introduction Inhaled corticosteroids (ICS) can reduce exacerbations and improve outcomes in patients with chronic obstructive pulmonary disease (COPD) who exhibit eosinophilic airway inflammation. However, not all patients benefit equally, and unnecessary ICS use increases the risk of pneumonia and other adverse effects. The blood eosinophil count (BEC) has been proposed as a simple and accessible biomarker to guide personalised ICS therapy, yet the strength and consistency of the supporting evidence remain uncertain. In the ICS-RECODE individual participant data (IPD) meta-analysis, we evaluated the utility of BEC as a biomarker for optimising and targeting ICS use in COPD. Methods ICS-RECODE is a prospectively designed, two-stage IPD meta-analysis (PROSPERO ID: CRD42024508286). Individual participant data from randomised controlled trials (RCTs) were re-analysed using harmonised statistical models to assess interactions between BEC and ICS treatment effects. Models were adjusted for established prognostic factors, including age, sex, smoking status, prior exacerbation history, baseline symptoms, lung function and BEC, concomitant COPD medications and comorbidities. Interaction coefficients from individual trials were combined using random-effects meta-analysis to quantify the overall BEC-treatment interaction. The credibility of observed effect modification was evaluated using the ICEMAN instrument Results Thirteen RCTs were included, contributing data from 21,519 participants. BEC was a strong and consistent predictor of ICS treatment response. For every 100 cells/μL increase in BEC, the reduction in exacerbation rates with ICS increased by approximately 4% (95% CI 2-7%) for any exacerbation (mild, moderate, or severe), 5% (95% CI 3-7%) for moderate or severe exacerbations, and 9% (95% CI 3-14%) for severe exacerbations (figure). No significant effect of ICS on severe exacerbations was observed in patients with BEC ≤ 100 cells/μL. ICS treatment significantly reduced all-cause mortality among participants with BEC ≥ 300 cells/μL. In contrast, there was no significant interaction between BEC and ICS effects on lung function or health status, as assessed by the Saint George’s Respiratory Questionnaire (SGRQ). The BEC-ICS interactions observed for exacerbation and mortality outcomes were rated as highly credible according to the ICEMAN assessment. Conclusion BEC is a robust and credible biomarker of ICS responsiveness in COPD. Higher BEC levels identify patients who derive greater benefit from ICS in reducing exacerbation risk and improving survival, while those with low BEC gain little benefit and remain exposed to potential harms. These findings provide strong evidence to support the use of BEC to guide personalised corticosteroid therapy in COPD. This abstract is funded by: National Institute of Health and Care Research (NIHR HTA, NIHR152516)
Mathioudakis et al. (Fri,) studied this question.