LAUSR

Corticosteroids reduce mortality in hospitalized patients with coronavirus disease 2019 (COVID‐19), but the response seems to vary according to the level of respiratory support needed. This retrospective cohort study included COVID‐19 patients with oxygen saturation (SatO2) in room air <92% admitted between March 3 and April 30, 2020. Following the interim protocol, patients could receive dexamethasone or methylprednisolone, and were classified according to oxygen requirements. The primary endpoint was admission to the intensive care unit (ICU) or mortality. Kaplan–Meier and Cox hazards analyses were used. Of the 115 patients included, 38 received corticosteroids. Among requiring high‐flow, noninvasive ventilation (NIV) or fraction of inspired oxygen (FiO2) > 0.40, the hazard ratio (HR) for death or ICU admission, between the corticosteroids and non‐corticosteroids group, was 0.07 (95% CI 0.01–0.4), p = .002, and for patients requiring low‐flow oxygen, the HR was 0.70 (95% CI 0.13–3.8), p = .68. Significant differences were also observed when all patients were analyzed together. A significant reduction in mortality and ICU admission frequency was observed among patients requiring high‐flow oxygen or NIV, but not among those requiring low‐flow oxygen. Better targeting of COVID‐19 patients is needed for the beneficial use of corticosteroids.