Tocilizumab in the treatment of COVID-19 – a meta-analysis

Link to article at PubMed

QJM. 2021 May 19:hcab142. doi: 10.1093/qjmed/hcab142. Online ahead of print.


OBJECTIVE: To evaluate whether IL-6 inhibitor tocilizumab (TCZ) reduces mortality among hospitalized COVID-19 patients.

METHODS: Systematic review and meta-analysis of randomized controlled trials (RCTs) comparing TCZ versus placebo/control, for treatment of adults with COVID-19. Primary outcome was 28-30 days all-cause mortality. Search was conducted up to April 1st 2021. Two independent reviewers screened citations, extracted data, and assessed risk of bias. Relative risk (RR) with 95% confidence intervals (CI) were pooled. We performed subgroup analysis for patients with critical illness and sensitivity analyses.

RESULTS: Eight RCTs were included, assessing 6,481 patients with mostly severe non-critical COVID-19 infection. TCZ was associated with a reduction in all-cause 28-30-day mortality compared to placebo/control (RR = 0.89, 95%CI 0.82-0.96). Among the subgroup of critically ill patients no reduced mortality was demonstrated (RR = 0.94, 95%CI 0.74-1.19). No mortality benefit with TCZ was demonstrated in trials that used steroids for >80% of patients. TCZ was associated with significantly reduced risk for mechanical ventilation (MV); for combined endpoint of death or MV; and for intensive care unit (ICU) admission. No significant difference in adverse events was demonstrated. Risk of serious superinfection was significantly lower with TCZ (RR = 0.57, 95%CI 0.35-0.93).

CONCLUSION: The treatment with TCZ reduces 28-30 days all-cause mortality, ICU admission, superinfections, MV and the combined endpoint of death or MV. Among critically ill patients, and when steroids were used for most patients, no mortality benefit was demonstrated. Additional research should further define sub-groups that would benefit most and preferred timing of administration of TCZ in severe COVID-19.

PMID:34010403 | DOI:10.1093/qjmed/hcab142

Leave a Reply

Your email address will not be published.