We systematically assessed benefits and harms of tocilizumab (TCZ), which is an antibody blocking IL-6 receptors, in hospitalized COVID-19 patients. Five electronic databases and two preprint webpages were searched until March 4, 2021. Randomized controlled trials (RCTs) and inverse probability treatment weighting (IPTW) cohorts assessing TCZ effects in hospitalized, COVID-19 adult patients were included. Primary outcomes were all-cause mortality, clinical worsening, clinical improvement, need for mechanical ventilation, and adverse events (AE). Inverse variance random-effects meta-analyses were performed with quality of evidence (QoE) evaluated using GRADE methodology. Nine RCTs (n = 7,021) and nine IPTW cohorts (n = 7,796) were included. TCZ significantly reduced all-cause mortality in RCTs (RR 0.89, 95%CI 0.81-0.98, p = 0.03; moderate QoE) and non-significantly in cohorts (RR 0.67, 95%CI 0.44-1.02, p = 0.08; very low QoE) vs. control (standard of care SOC or placebo). TCZ significantly reduced the need for mechanical ventilation (RR 0.80, 95%CI 0.71-0.90, p = 0.001; moderate QoE) and length of stay (MD -1.92 days, 95%CI -3.46 to -0.38, p = 0.01; low QoE) vs. control in RCTs. There was no significant difference in clinical improvement or worsening between treatments. AEs, severe AEs, bleeding and thrombotic events were similar between arms in RCTs, but there was higher neutropenia risk with TCZ (very low QoE). Subgroup analyses by disease severity or risk of bias (RoB) were consistent with main analyses. Quality of evidence was moderate to very low in both RCTs and cohorts. In comparison to SOC or placebo, TCZ reduced all-cause mortality in all studies and reduced mechanical ventilation and length of stay in RCTs in hospitalized COVID-19 patients. Other clinical outcomes were not significantly impacted. TCZ did not have effect on AEs, except a significant increased neutropenia risk in RCTs. TCZ has a potential role in the treatment of hospitalized COVID-19 patients.