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

DOI: https://doi.org/10.1186/s13054-020-03429-w

IF: 15.1

2020-01-01

Critical Care

Abstract:Abstract 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.