Objectives: To assess the effectiveness and safety of tocilizumab, a humanised anti-interleukin-6 receptor antibody, in the treatment of critical or severe coronavirus disease 2019 (COVID-19) patients. Methods: This was a retrospective cohort study of severe or critical COVID-19 patients (≥18 years) admitted to one hospital in Kuwait. Fifty-one patients received intravenous tocilizumab, while 78 patients received the standard of care at the same hospital. Both groups were compared for clinical improvement and in-hospital mortality. Results: The tocilizumab (TCZ) group had a significantly lower 28-day in-hospital mortality rate than the standard-of care-group (21.6% vs. 42.3% respectively; p = 0.015). Fifty-five per cent of patients in the TCZ group clinically improved vs. 11.5% in the standard-of-care group (p < 0.001). Using Cox-proportional regression analysis, TCZ treatment was associated with a reduced risk of mortality (adjusted hazard ratio 0.25; 95% CI: 0.11–0.61) and increased likelihood of clinical improvement (adjusted hazard ratio 4.94; 95% CI: 2.03–12.0), compared to the standard of care. The median C-reactive protein, D-dimer, procalcitonin, lactate dehydrogenase and ferritin levels in the tocilizumab group decreased significantly over the 14 days of follow-up. Secondary infections occurred in 19.6% of the TCZ group, and in 20.5% of the standard-of-care group, with no statistical significance (p = 0.900). Conclusion: Tocilizumab was significantly associated with better survival and greater clinical improvement in severe or critical COVID-19 patients. © 2021 John Wiley & Sons Ltd.