Meta-Analysis of Glucocorticoids for Covid-19 Patients Not Receiving Oxygen

BACKGROUND: Glucocorticoids reduce mortality in hospitalized patients with severe and critical coronavirus disease 2019 (Covid-19), although a possible harm was documented in patients with Covid-19 not requiring oxygen. METHODS: We searched Embase, BioMed Central, medRxiv, bioRxiv, PubMed, and the Cochrane Central Register of Controlled Trials for any randomized trial or matched study ever performed on adult patients with Covid-19 not receiving oxygen therapy treated with intravenous or oral glucocorticoids versus any comparator (standard therapy or placebo); there were no restrictions on dose or time of administration. The primary end point was all-cause mortality at the longest available follow-up. RESULTS: Five randomized trials and one propensity-matched study involving 6634 hospitalized patients not on oxygen were finally included (3704 received glucocorticoids and 2930 received standard treatment). The overall mortality of patients treated with glucocorticoids was significantly higher than the mortality of patients in the control group (509 of 3704 [14%] in the glucocorticoid group vs. 294 of 2930 [10%] in the control group; odds ratio, 1.56 [95% confidence interval, 1.27 to 1.92], with three articles reporting mortality events and contributing to the combined odds ratio; P<0.001; number needed to harm=27). CONCLUSIONS: Glucocorticoid use likely increases mortality in hospitalized patients with Covid-19 not receiving oxygen, with a number needed to harm of 27. (PROSPERO number CRD42022342996.)