Background Anakinra may represent an important therapy to improve the prognosis of COVID-19 patients . This meta-analysis using individual patient data was designed to assess the efficacy and safety of anakinra treatment in patients with COVID-19 . Methods Based on a pre-specified protocol (PROSPERO: CRD42020221491), a systematic literature search was performed in MEDLINE (PubMed), Cochrane, medRxiv.org, bioRxiv.org and clinicaltrials.gov databases for trials in COVID-19 comparing administration of anakinra with standard-of-care and/or placebo . Individual patient data from eligible trials were requested . The primary endpoint was the mortality rate and the secondary endpoint was safety . Findings Literature search yielded 209 articles, of which 178 articles fulfilled screening criteria and were full-text assessed . Aggregate data on 1185 patients from 9 studies were analyzed and individual patient data on 895 patients from 6 studies were collected . Most studies used historical controls . Mortality was significantly lower in anakinra-treated patients (38/342 [11.1 %] ) as compared with 137/553 (24.8 %) observed in patients receiving standard-of-care and/or placebo on top of standard-of-care (137/553 [24.8 %] ); adjusted odds ratio (OR), 0.32; 95% CI , 0.20 to 0.51; p <0.001 . The mortality benefit was similar across subgroups regardless of diabetes mellitus, ferritin concentrations, or baseline P/F ratio . The effect was more profound in patients exhibiting CRP levels> 100 mg/L (OR 0.28,95% CI 0.27-1.47). Safety issues, such as increase of secondary infections, did not emerge . Interpretation Anakinra may be a safe anti-inflammatory treatment option in patients hospitalized with moderate-to-severe COVID-19 pneumonia to reduce mortality, especially in the presence of hyperinflammation signs such as CRP> 100mg /L . Funding Sobi.