Real-world effectiveness of steroids in severe COVID-19: longer courses associated with lower risk of death or ICU admission (preprint)

ABSTRACT

Purpose We aim to investigate the associations of steroid and length of steroid use with outcomes in severe COVID-19.Methods Severe cases of COVID-19, defined by hypoxia at presentation, and admitted to a multi-site healthcare institution in London were analysed between 02-Sep-2020 and 27-May-2021. The associations between duration of steroid treatment (prescription-days) and outcomes were explored using Cox proportional-hazards models adjusting for confounders. Length of steroid treatment was analysed as both a continuous variable and categorised into < 3, 3–10, and > 10 days. The primary outcome was in-hospital mortality and secondary outcome was in-hospital mortality or intensive care unit (ICU) level-3 admission.Results 734 severe COVID-19 cases were included, with 137/734 (18.7%) treated with steroids for < 3 days, 497/734 (67.7%) for 3–10 days, and 100/734 (13.6%) for > 10 days. Cox modelling with continuous days showed increasing length of steroids decreased the hazard of in-hospital mortality by a factor of 0.98 [95% CI 0.96-1.0] per additional day and in-hospital mortality or ICU admission by a factor of 0.91 [95% CI 0.87–0.95] per additional day. Further, when taking 3–10 days steroid treatment group as the reference group, > 10 days steroid showed trends towards decreased hazards for death (HR 0.59 [95%CI 0.30–1.14]) and was significantly protective for death/ICU outcome (HR 0.28 [95%CI 0.11–0.68]).Conclusion The protective effect of steroid for severe COVID-19 reported in randomised clinical trials was replicated in this large real-world cohort. We found an association between longer steroid courses and lower risk of death or ICU admission that warrants further investigation.