Corticosteroid therapy in critically ill patients with COVID-19: a multicenter, retrospective study

Li, Yiming; Meng, Qinghe; Rao, Xin; Wang, Binbin; Zhang, Xingguo; Dong, Fang; Yu, Tao; Li, Zhongyi; Feng, Huibin; Zhang, Jinpeng; Chen, Xiangyang; Li, Hunian; Cheng, Yi; Hong, Xiaoyang; Wang, Xiang; Yin, Yimei; Zhang, Zhongheng; Wang, Dawei

BACKGROUND: Corticoid therapy has been recommended in the treatment of critically ill patients with COVID-19, yet its efficacy is currently still under evaluation. We investigated the effect of corticosteroid treatment on 90-day mortality and SARS-CoV-2 RNA clearance in severe patients with COVID-19. METHODS: 294 critically ill patients with COVID-19 were recruited between December 30, 2019 and February 19, 2020. Logistic regression, Cox proportional-hazards model and marginal structural modeling (MSM) were applied to evaluate the associations between corticosteroid use and corresponding outcome variables. RESULTS: Out of the 294 critically ill patients affected by COVID-19, 183 (62.2%) received corticosteroids, with methylprednisolone as the most frequently administered corticosteroid (175 accounting for 96%). Of those treated with corticosteroids, 69.4% received corticosteroid prior to ICU admission. When adjustments and subgroup analysis were not performed, no significant associations between corticosteroids use and 90-day mortality or SARS-CoV-2 RNA clearance were found. However, when stratified analysis based on corticosteroid initiation time was performed, there was a significant correlation between corticosteroid use (≤ 3 day after ICU admission) and 90-day mortality (logistic regression adjusted for baseline: OR 4.49, 95% CI 1.17–17.25, p = 0.025; Cox adjusted for baseline and time varying variables: HR 3.89, 95% CI 1.94–7.82, p < 0.001; MSM adjusted for baseline and time-dependent variants: OR 2.32, 95% CI 1.16–4.65, p = 0.017). No association was found between corticosteroid use and SARS-CoV-2 RNA clearance even after stratification by initiation time of corticosteroids and adjustments for confounding factors (corticosteroids use ≤ 3 days initiation vs no corticosteroids use) using MSM were performed. CONCLUSIONS: Early initiation of corticosteroid use (≤ 3 days after ICU admission) was associated with an increased 90-day mortality. Early use of methylprednisolone in the ICU is therefore not recommended in patients with severe COVID-19.