2195 
Although corticosteroids have been hailed as life-saving drugs among those hospitalised with COVID-19, the benefits only apply seriously ill patients.
From an early stage of the COVID-19 pandemic it became clear that infection with the virus was associated with an increased expression of inflammatory markers including C-reactive protein and both interleukin-1 and 6. Consequently, clinicians sought to examine the efficacy of systemic corticosteroids which serve to diminish this inflammatory response. As a result, several randomised trials have been undertaken among those hospitalised with COVID-19 to provide an estimate of the efficacy of these drugs. For the present meta-analysis, a team from the department of internal medicine, South Illinois university school of medicine, US, sought to establish the association of systemic corticosteroid therapy compared to usual care or placebo on all-cause mortality in people hospitalised with COVID-19 up to January 2021. The team calculated odds ratios to compare for each of the trials, how corticosteroids impacted on mortality among those who were critically ill, i.e., admitted to intensive care units or receiving invasive mechanical ventilation, compared to non-critically ill patients. The primary outcome of interest was all-cause mortality up to 30 days after randomisation and included studies with a shorter timeframe where 30-day mortality was not available.
Findings
A total of 5 randomised trials were identified conducted in several countries and which included 7645 patients. Three different corticosteroids, dexamethasone, hydrocortisone and methylprednisolone were used in the trials and compared against placebo (two trials) and usual care in the remaining three studies. Interestingly, three of these trials were prematurely ended once the results from one of the studies (RECOVERY) became available. In terms of overall mortality there were a total of 728 deaths among those receiving systemic corticosteroids compared to 1330 deaths in the 4842 patients in the control arm. This led to a non-significant summary odds ratio of 0.82 (odds ratio, OR = 0.82 95% CL 0.64 – 1.05, p = 0.09). In contrast, where patients were critically ill, there was a significant impact from corticosteroids (OR = 0.67 95% CL 0.51 – 0.87, p = 0.01). However, in an analysis of the subgroup of patients who were not critically ill at randomisation, the odds ratio was non-significant (OR = 0.95 95% CL 0.75 – 1.19, p = 0.21). In discussing their findings, the authors note how systemic corticosteroids are really only of benefit to those deemed critically ill. However, it was also clear that the early termination of studies resulted in a potential underestimation of the effect of corticosteroids on mortality.
They concluded by calling for future studies to clarify the role of systemic corticosteroids in the management of COVID-19.
Citation
Robinson R et al. Impact of systemic corticosteroids on hospitalised patients with COVID-19: January 2021 meta-analysis of randomised controlled trials. MedRxiv 2021 https://doi.org/10.1101/2021.02.03.21251065
