It Is Time to Determine Tocilizumab’s Place in Coronavirus Disease 2019 (COVID-19)

To the Editor—We read with great interest the meta-analysis by Huang et al [1]. Huang et al conducted a meta-analysis of the first 5 randomized clinical trials (RCTs) with tocilizumab in COVID-19 and concluded that tocilizumab (TCZ) does not provide mortality benefit for patients with severe coronavirus disease 2019 (COVID-19). Since the meta-analysis by Huang et al, 3 new RCTs [2–4] have been published or are at a pre-publication stage. Veiga et al’s [2] trial was stopped early in July 2020 after an increase in deaths in opposition to REMAP-CAP and RECOVERY results [3, 4] which showed a positive benefit of tocilizumab on mortality. We think that an updated meta-analysis and discussion are necessary. In this context, we aimed to perform a meta-analysis on these 8 RCTs [2–9] on the impact of TCZ administration on mortality. We selected through a systematic search on PubMed and the preprint server medRxiv (until 3 March 2021) all RCTs that compared the clinical outcome of patients with COVID-19 treated with TCZ versus standard of care or placebo. Our primary endpoint was 28-day mortality. Secondary endpoints were mechanical ventilation incidence and safety endpoints (adverse events and serious infections). We included 8 RCTs. A total of 6303 patients were included: 3266 randomized to TCZ and 3037 to placebo (Figure 1). Overall, there were 810 (24.8%) deaths at day 28 in the TCZ group and 893 (29.4%) deaths in the placebo group (pooled odds ratio [OR], .86; 95% confidence interval [CI], .76–.96; P = .008). Mechanical ventilation incidence had a pooled OR (.72; 95% CI, .62–.84; P < .001) in favor of TCZ. There were 88 of 705 (12.5%) serious infections in the TCZ group and 60 of 353 (17%) serious infections in the placebo group (pooled OR, .67; 95% CI, .47–.97; P = .03) and no significant differences in adverse events (Figure 1). A, Forest plot for the effect of tocilizumab on mortality at days 28–30 in randomized trials. B, Forest plot for the effect of tocilizumab on mechanical ventilation incidence in randomized trials. Abbreviations: CI, confidence interval; M-H, Mantel-Haenszel. A few assumptions can be discussed to explain this contradiction on mortality effect in these RCT results. First, a lack of statistical power seems manifest in some RCTs. For example, Stone et al [6] and Salvarini et al [7] showed a mortality rate of less than 5% in their population and the required number of patients to conclude for the lack of impact on mortality in COVID-19 with tocilizumab used was not reached; Veiga et al’s [2] results must be interpreted with caution due to the sample size of the trial and considering that there were no significant differences in mortality at day 28. Second, we only have the results of short-term mortality. For example, in COVACTA [5], at day 28, 72% (83/115) of the patients who were still hospitalized required a high level of oxygen support: 17% (50/294) in the TCZ arm versus 23% (33/144) in the placebo arm. We can possibly expect a lower number of deaths in the TCZ arm in long-term mortality. Third, these RCTs included heterogenous populations, which may explain the heterogeneity of results [10]. We have recently defined the optimal group who is susceptible to have the greatest benefit from TCZ as severe/critical COVID-19 (except for patients after some time of mechanical ventilation) [10]. For example, in COVACTA [5], if we choose this population, the category 4 and 5 of the 7-category ordinal scale (high-flow oxygen or noninvasive ventilation or mechanical ventilation at an early stage), the number of deaths is clearly lower for TCZ than placebo (17% [24/139] vs 28% [15/54]). A mortality rate of 17% is extremely low in this intensive care unit population and is in contrast to the 17% global in-hospital mortality rate in the United States [11]. To conclude, TCZ reduces mortality in severe/critical COVID-19 pneumonia. Due to the heterogenous population in RCTs, secondary analyses on subgroups are needed and would be helpful to define the optimal group and timing for TCZ benefit. Nevertheless, a thorough analysis of RCT results is needed in order to not misinform the medical community. Finding the optimal group of patients who are susceptible to have the greatest benefit has now become the main challenge. Potential conflicts of interest. The authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.