Effectiveness of tocilizumab in hospitalized moderate-to-severe COVID-19 patients: a real-life study

BACKGROUND: To assess the clinical effectiveness of Tocilizumab (TCZ) in moderate-to-severe hospitalized COVID-19 patients and factors associated with clinical response. METHODS: Five hundred eight inpatients with moderate-to-severe SARS-CoV-2 infection were enrolled. TCZ effect in addition to standard medical therapy was evaluated in terms of death during hospital stay. Unadjusted and adjusted risk of mortality for TCZ treated patients versus TCZ untreated ones was estimated using robust Cox regression model. We considered the combination of TCZ and ICU as time-dependent exposure and created a model using duplication method to assess the TCZ effect in very severe COVID-19 patients. RESULTS: TCZ reduced death during hospital stay in the unadjusted model (HR 0.54, 95%CI 0.33-0.88) and also in the adjusted model, although with loss of statistical significance (HR 0.72, 0.43-1.20). Better effectiveness was observed in patients with low SpO2/FiO2 ratio (HR 0.35, 0.21-0.61 vs. 1.61, 0.54-4.82, P<0.05), and, without statistical significance, in patients with high CRP (HR 0.51, 0.30-0.87 vs. 0.41, 0.12-1.37, P=NS) and high IL-6 (HR 0.49, 0.29-0.82 vs. 1.00, 0.28-3.55, P=NS). TCZ was effective in patients not admitted to ICU, both in the unadjusted (HR 0.33, 0.14-0.74) and in the adjusted (HR 0.39, 0.17-0.91) model but no benefit was observed in critical ICU-admitted patients both in the unadjusted (HR 0.66, 0.37-1.15) and in the adjusted model (HR 0.95, 0.54-1.68). CONCLUSIONS: Our real-life study suggests clinical efficacy of TCZ in moderate-to-severe COVID-19 patients but not in end-stage disease. Thus, to enhance TCZ effectiveness, patients should be selected before grave compromise of clinical conditions. Predictive model study