Differential association of fluticasone furoate and budesonide with clinically detected COVID-19: a retrospective cohort study

Objectives To assess whether fluticasone furoate (FF) use, compared with budesonide (BUD), is associated with fewer clinically detected COVID-19 events among inhaled corticosteroids (ICS) users, and to explore virus-specificity using influenza as a comparator outcome. We hypothesised that FF may provide strong local anti-inflammatory effects with limited systemic immunosuppression. Design Retrospective cohort with outpatient follow-up over 4 years. Setting Single Japanese medical centre. Participants 334 adults (102 ICS users; 232 non-ICS) followed from July 2020 to July 2024. Main outcome measures Clinically detected COVID-19 (primary) and influenza (secondary). The primary exposure comparison was FF versus BUD among ICS users; ICS versus non-ICS was analysed secondarily (exploratory). Cox proportional hazards and logistic regression are adjusted for demographics, comorbidities, vaccination and systemic corticosteroids. Results Seventy-nine COVID-19 and 14 influenza events occurred. Among ICS users, FF was associated with fewer clinically detected COVID-19 events than BUD (adjusted HR 0.12, 95% CI 0.02 to 0.73; crude 6.5% vs 32.3%; Fisher’s exact p=0.0047). Under symptom-based testing, ICS users also had fewer clinically detected COVID-19 events than non-ICS users (adjusted HR 0.46, 95% CI 0.22 to 0.94), although this comparison is limited by baseline imbalance and should be interpreted as exploratory and non-causal. Conclusions Under symptom-triggered testing, FF was associated with fewer clinically detected COVID-19 events than BUD in the head-to-head comparison among ICS users. The ICS versus non-ICS comparison is exploratory due to confounding by indication and structural imbalance. These findings are hypothesis-generating and warrant prospective studies with systematic testing and stronger designs.