Comparative real-world effectiveness and safety of benralizumab and two mepolizumab dosing regimens in eosinophilic granulomatosis with polyangiitis: a 24-month prospective single-center cohort study

Background – Long-term prospective real-world data on anti–IL-5 and anti–IL-5 receptor biologics in eosinophilic granulomatosis with polyangiitis (EGPA) are limited. Methods – In this 24-month prospective single-center observational study, 66 adults with EGPA received benralizumab 30 mg (every 4 weeks for the first three doses, then every 8 weeks), mepolizumab 300 mg every 4 weeks or mepolizumab 100 mg every 4 weeks. Remission was defined as BVASv3 = 0 with prednisone ≤5 mg/day. Glucocorticoid (GC)-free status was defined as no oral GCs. The primary analysis was conducted in patients starting biologics as first-line. To address confounding by indication, we performed multivariable logistic regression and multinomial propensity score inverse probability weighting (IPTW). Analyses were repeated in an all-lines dataset (81 treatment lines) using generalized estimating equations (GEE) clustered by patient. Results – Clinical outcomes improved over the 24-month follow-up across all regimens. By 24 months, remission was 73.7% (14/19) with benralizumab, 81.0% (17/21) with mepolizumab 300 mg, and 50.0% (6/12) with mepolizumab 100 mg, without statistically significant differences between regimens. GC-free status increased over time, reaching 63.2% (12/19) with benralizumab, 85.7% (18/21) with mepolizumab 300 mg, and 58.3% (7/12) with mepolizumab 100 mg at 24 months, with no significant between-group differences. Adjusted and propensity-based sensitivity analyses yielded consistent conclusions. Eosinophils decreased markedly, with near-complete suppression with benralizumab. Pulmonary function indices showed improvement during follow-up, with the clearest within-group improvement observed in the benralizumab group. Treatment persistence at 24 months varied across regimens, although no statistically significant difference was observed (log-rank p=0.222). Discontinuations were mainly driven by inadequate control of ENT or respiratory manifestations. In all-lines GEE analyses, regimen effects on remission and GC-free status were broadly consistent with the primary analysis. Adverse events were uncommon and no serious events were reported. Conclusions – Anti–IL-5/IL-5R biologics were effective and well tolerated over 24 months in EGPA, with substantial GC sparing effect. The observational design and potential confounding by indication limit comparative effectiveness inference. These findings support individualized biologic selection in EGPA.