Beneficial and harmful outcomes of tocilizumab in severe COVID-19: A systematic review and meta-analysis

Introduction The results of studies of tocilizumab (TCZ) in COVID-19 are contradictory. Our study aims to update medical evidence from controlled observational studies and randomized clinical trials (RCTs) on the use of TCZ in hospitalized patients with COVID-19. Methods We searched the following databases from January 1, 2020 to April 13, 2021 (date of the last search): MEDLINE database through the PubMed search engine and Scopus, using the terms ("COVID-19" [Supplementary Concept]) AND "tocilizumab" [Supplementary Concept]). Results Sixty four studies were included in the present study: 54 were controlled observational studies (50 retrospective and 4 prospective) and 10 were RCTs. The overall results provided data from 20,616 hospitalized patients with COVID-19: 7668 patients received TCZ in addition to standard of care (SOC) (including 1915 patients admitted to intensive care units (ICU) with reported mortality) and 12,948 patients only receiving SOC (including 4410 patients admitted to the ICU with reported mortality). After applying the random-effects model, the hospital-wide (including ICU) pooled mortality odds ratio (OR) of patients with COVID-19 treated with TCZ was 0.73 (95% confidence interval (CI) = 0.56-0.93). The pooled hospital-wide mortality OR was 1.25 (95% CI = 0.74-2.18) in patients admitted at conventional wards versus 0.66 (95% CI = 0.59-0.76) in patients admitted to the ICU. The pooled OR of hospital-wide mortality (including ICU) of COVID-19 patients treated with TCZ plus corticosteroids (CS) was 0.67 (95% CI = 0.54-0.84). The pooled in-hospital mortality OR was 0.71 (95% CI = 0.35-1.42) when TCZ was early administered (<= 10 days from symptom onset) versus 0.83 (95% CI 0.48-1.45) for late administration (>10 days from symptom onset). The meta-analysis did not find significantly higher risk for secondary infections in COVID-19 patients treated with TCZ. Conclusions TCZ prevented mortality in patients hospitalized for COVID-19. This benefit was seen to a greater extent in patients receiving concomitant CS and when TCZ administration occurred within the first 10 days after symptom onset.