Background In December 2019 a novel coronavirus designated SARS-CoV-2 was identified, and the disease COVID-19 has caused many deaths. SARS-CoV-2 infection has been associated with the development of cytokine storm (including interleukin 6 (IL-6)), which can cause lung damage and lack of oxygen. Tocilizumab (TCZ) inhibits ligand binding to the IL-6 receptor and may be a potential treatment for the hyperinflammation symptoms of COVID-19. However, data regarding the efficacy of TCZ in COVID-19 are lacking. The rapid spread of the pandemic in France, especially in the Paris region, constrained us to the off-label use of TCZ in patients with severe clinical conditions.

Methods A single-centre observational cohort study of 44 patients infected with COVID-19 was carried out between 6 April and 21 April 2020 in Groupe Hospitalier Intercommunal Le Raincy-Montfermeil (GHILRM). Twenty-two patients diagnosed with COVID-19 were treated with TCZ and were compared with 22 patients not treated with TCZ matched for age, gender and length of hospital stay for COVID-19. Respiratory rate and oxygen supplementation as well as laboratory parameters (such as C-reactive protein (CRP), aspartate aminotransferase and alanine aminotransferase) were collected at baseline and during 14 days of follow-up. Our primary objective was to assess the efficacy of TCZ on respiratory clinical conditions.

Findings The average respiratory rate was lower in the TCZ group than in the control group (21.5 vs 25.5 breaths/min at day 14, 95% CI −7.5 to −0.4; p=0.03). Treated patients tended to be intubated less during the course of the disease (2/22 vs 6/22, 95% CI −0.4 to 0.1; p=0.12). In each group, 10 patients no longer required oxygen therapy. We found a significant decrease in CRP in treated patients on day 7 (p=0.04). TCZ caused cytolysis in more than half (14/22) of the patients but without clinical impact.

Interpretation There was a significant difference in the respiratory rate on day 14 of follow-up, with a greater decrease observed in the treated group. Fewer patients required mechanical ventilation in the TCZ group, especially among patients with more extensive CT lung damage, than in the control group. The same number of patients were weaned off oxygen on day 14 in the two groups, while the patients in the TCZ group had more severe impairment at inclusion. We consider that TCZ showed significant control of the biological inflammatory syndrome, suggesting that it may limit the effect of the cytokine storm. Our study seems to indicate the efficacy of TCZ, particularly in patients with severe initial pulmonary impairment. Selecting the best candidates and the best timing for TCZ therapy needs to be determined in randomised clinical trials.