Comparative Real-World Effectiveness of Dupilumab, Mepolizumab, and Benralizumab in Patients with Chronic Obstructive Pulmonary Disease: A Propensity Score-Matched Cohort Study.
Liao Kuang-Ming KM, Huang Sheng-Chi SC, Wu Jheng-Yan JY, Kuo Chia-Yu CY et al.
Dupilumab, mepolizumab, and benralizumab have each demonstrated efficacy in placebo-controlled trials for chronic obstructive pulmonary disease (COPD) with type 2 inflammation. However, no head-to-head trials or real-world comparative effectiveness studies have been reported. We aimed to compare the effectiveness of dupilumab, mepolizumab, and benralizumab in patients with COPD using propensity score-matched cohort analyses. Retrospective cohort study using the TriNetX US Collaborative Network. Adults with COPD initiating dupilumab, mepolizumab, or benralizumab between 2022 and 2025 were identified. Biologic-naive new users were included in two pairwise comparisons: dupilumab versus mepolizumab and dupilumab versus benralizumab. Propensity score matching (1:1) balanced baseline characteristics. The primary outcome was COPD exacerbation; secondary outcomes included acute respiratory failure, pneumonia, and bronchitis over 365 days. Hazard ratios (HRs) with 95% confidence intervals (CIs) were estimated using Cox proportional hazards models. Propensity score matching identified 3078 dupilumab-mepolizumab pairs and 2109 dupilumab-benralizumab pairs. Compared with mepolizumab, dupilumab was associated with a numerically lower risk of COPD exacerbation (HR, 0.76; 95% CI, 0.57-1.01) and a significantly lower risk of bronchitis (HR, 0.69; 95% CI, 0.49-0.98). No significant differences were observed between dupilumab and benralizumab for COPD exacerbation (HR, 0.98; 95% CI, 0.70-1.40), acute respiratory failure, pneumonia, or bronchitis. In this large real-world cohort study, no statistically significant difference in COPD exacerbation risk was observed between dupilumab and mepolizumab or between dupilumab and benralizumab. However, blood eosinophil counts and baseline lung function measurements were not available in the dataset, so residual confounding cannot be excluded.