ABSTRACT
Background To estimate the association between the administration of corticosteroids and all-cause mortality of hospitalized patients with severe/critical Corona Virus Disease 2019 (COVID-19).Methods We conducted this study at six tertiary hospitals in the area of Chongqing, China. Clinical records from all consecutive adult subjects admitted with SARS-CoV-2 infection from 1 November 2022 to 20 January 2023 were retrospectively reviewed. In-hospital and 28-day mortality were analyzed before and after propensity score matching (PSM). Logistic regression model and cox regression model was used to further examine the relationship between corticosteroid treatment and the risk of mortality.Results A total of 406 severe and critically ill COVID-19 patients were included in this study. They were divided into the corticosteroids group (231, 56.9%) and non-corticosteroids group (175, 43.1%) according to corticosteroids use (0.5-1mg/kg/d methylprednisolone or any corticosteroid at equivalent dose, no more than 10 days). Corticosteroid treatment did not reduce in-hospital mortality in overall cohorts (36.4% vs.28.0%) and matched cohorts (32.5% vs.34.5%). Univariate analysis showed that the 28-day mortality in the corticosteroids group was significantly higher than that in the non-corticosteroids group [hazard ratio (HR), 0.706; 95% CI, 0.507–0.983; p = 0.039] in the overall population. When the multivariate analysis was further used to control confounders, the results showed that corticosteroids were not associated with improved 28-day mortality (HR, 1.121; 95% CI, 0.641–1.959; p = 0.689). With PSM, similar results were obtained with univariate and multivariate analysis.Conclusions Corticosteroids in hospitalized patients with severe/critical COVID-19 did not reduce mortality in the overall population.