Effectiveness and safety of tocilizumab for COVID-19: a systematic review and meta-analysis of randomized clinical trials

ABSTRACT BACKGROUND: Tocilizumab is an anti-human interleukin 6 receptor monoclonal antibody that has been used to treat coronavirus disease 2019 (COVID-19). However, there is no consensus on its efficacy for the treatment of COVID-19. OBJECTIVE: To evaluate the effectiveness and safety of tocilizumab for treating COVID-19. DESIGN AND SETTING: Systematic Review of randomized controlled trials (RCTs), Universidade Federal de São Paulo (UNIFESP), São Paulo (SP), Brazil. METHODS: We searched MEDLINE via PubMed, EMBASE, CENTRAL, and IBECS for RCTs published up to March 2021. Two authors selected studies and assessed the risk of bias and the certainty of the evidence following Cochrane Recommendations. RESULTS: Eight RCTs with 6,139 participants were included. We were not able to find differences between using tocilizumab compared to standard care on mortality in hospitalized patients with COVID-19 (risk ratio (RR) 0.97, 95% confidence interval (CI) 0.84 to 1.13; 8 trials; 5,950 participants; low-certainty evidence). However, hospitalized patients under tocilizumab plus standard care treatment seemed to present a significantly lower risk of needing mechanical ventilation (risk ratio = 0.78; 95% CI 0.64−0.94 moderate-certainty of evidence). CONCLUSIONS: To date, the best evidence available shows no difference between using tocilizumab plus standard care compared to standard care alone for reducing mortality in patients with COVID-19. However, as a finding with a practical implication, the use of tocilizumab in association to standard care probably reduces the risk of progressing to mechanical ventilation in those patients. REGISTRATION: osf.io/qe4fs.


How the intervention might work
COVID-19 creates a hyperinflammatory condition, activated by a cytokine cascade. Of all cytokines identified so far, IL-6 is most closely connected to disease severity. 7 TCZ inhibits IL-6 action and might be a way to reduce COVID-19 severe cases.

Why it is important to do this review
Several observational studies have been conducted on treating COVID-19 and they suggest that TCZ is beneficial for moderate, severe, or critical cases of COVID-19. [8][9][10] However, non-randomized studies may report spurious associations mainly arising from the introduction of confounding factors into the comparative groups, and relying on such results may lead to the introduction of potentially hazardous interventions into clinical practice. Randomized clinical trials (RCTs) became available only by the end of 2020 and they have, so far, shown mixed results for mortality. Therefore, systematic reviews evaluating the effects of tocilizumab considering only RCTs are urgently needed.

OBJECTIVES
The aim of this review was to evaluate the effectiveness and safety of tocilizumab for treating COVID-19.

Criteria for considering studies for this review
We undertook a systematic review including only RCTs.
Participants must have been diagnosed with COVID-19 by one of the following methods: real time reverse-transcriptase polymerase chain reaction, serum immunoglobulin M antibody assay, or clinical evaluation (typical computed tomographic scan with signs of pneumonia). We included trials evaluating the effect of tocilizumab used alone or in combination with standard care or other interventions.

Outcomes
Our primary outcome was mortality. Secondary outcomes included the need for mechanical ventilation, days until discharge from hospital, and adverse events.

Search methods for identification of studies
The search was for all relevant published and unpublished trials without restrictions on language, year, or publication status. Search strategies for each database are provided in Appendix 1.
References of included trials were checked to identify additional, relevant trials. When necessary, authors were contacted.

Study selection and data extraction
All abstracts and reports identified by the search were retrieved and independently evaluated by two authors. If the reference appeared relevant to the review topic, the full text was obtained.
The same two authors assessed and selected any relevant trials according to the review's eligibility criteria. In the presence of any disagreements, a third author was consulted.

Assessment of risk of bias and certainty of evidence
The risk of bias in each trial was assessed by two independent authors. We assessed the methodological quality of each included study using the risk of bias (RoB 2.0) tool as per the Cochrane recommendations. We evaluated the following domains: risk of bias arising from the randomization process, risk of bias due to deviations from the intended interventions (effect of assignment to intervention), missing outcome data, risk of bias in measurement of the outcome, risk of bias in selection of the reported result, and overall risk of bias. Each study was evaluated on all six domains and for each domain the evaluations were scored by assigning the classifications "low risk of bias", "some concerns of risk of bias", or "high risk of bias". 11 We used the GRADE (Grading of Recommendations Assessment, Development and Evaluations) approach to classify the strength of evidence as very low, low, moderate, or high. 12 We evaluated the following criteria: risk of bias, inconsistency, imprecision, and indirectness.
We summarized the findings, considering the primary outcomes from comparisons, using the GRADE pro platform.

Measures of treatment effect
We estimated the effects of tocilizumab treatments for our predefined outcomes. Relative risks with their 95% confidence intervals (CI) were estimated using Review Manager 5.4.1 software (London, United Kingdom). We pooled data from the included studies using the generic inverse variance method with a random-effects model. We assessed heterogeneity using the I 2 statistic. 13 The interpretation of I 2 depends on the magnitude and direction of the effect as well as the strength of evidence for heterogeneity. We used the following thresholds to assess I 2 : 0% to 40%: likely not important; 30% to 60%: moderate heterogeneity; 50% to 90%: substantial heterogeneity; 75% to 100%: considerable heterogeneity.

Results of the search
Our database search strategies yielded 413 records. After excluding duplicated reports and reports that were clearly irrelevant or not directly related to the review question, we assessed eleven full-text studies for further scrutiny. Eight multi-center RCTs 7,14-20 with 6,139 participants were finally included in our systematic review (Figure 1). Details of each trial are described in Table 1.

Participants over 18 years from Europe and South and North
America were randomized in each included trial into two groups: standard care alone or associated with tocilizumab 8 mg/kg (maximum dose of 800 mg/day). Tocilizumab was administered to participants as soon as they were randomized.
Standard care was not specified in the majority of the trials. Important baseline characteristics of the participants and interventions are described in Table 1. All trials included hospitalized patients with moderate to severe COVID-19.

Risk of bias in included studies
The RCTs were assessed by RoB 2.0 (Figure 2). Three of them were judged as being of some concern regarding the risk of bias, four of them were judged as having low risk of bias and only one was graded as having high risk of bias. The most penalized domain was deviation from intended interventions, which occurred mainly because of lack of blinding and/or inappropriate analyses (intention-to-treat).

Certainty of evidence
We rated the certainty of the evidence using the GRADE approach. We found low certainty of evidence for the all-cause mortality outcome ( Table 2). For that outcome, we downgraded one level due to methodological limitation (risk of bias) and one level due to imprecision (the 95% CI included both a benefit and harm, showing imprecision of the estimated effect). We found    (Figure 3).

Need for mechanical ventilation
Patients with COVID-19 treated with tocilizumab plus standard care presented significantly lower risk of progressing to mechanical ventilation when compared to those receiving standard care alone (RR 0.78, 95% CI 0.64 to 0.94; 6 trials; 4,705 participants; moderate certainty of evidence) (Figure 4).
A few trials, including EMPACTA, 18 COVACTA, 19 TOCIBRAS, 15 and RECOVERY 16    Explanations a. We downgraded one level because three studies (n = 1,075) had some concerns on the risk of bias and one study (n = 377) had a high risk of bias. b. We downgraded one level because the 95% CI includes both no effect and a possible benefit. c. We downgraded one level because two studies (n = 515) had some concerns on the risk of bias and one study (n = 377) had a high risk of bias.

DISCUSSION
In this systematic review, including only RCTs assessing the effects of tocilizumab in patients with COVID-19, we found moderate-certainty evidence from six RCTs demonstrating that the use of tocilizumab in combination with standard care was effective for the reduction of need for mechanical ventilation in hospitalized patients with COVID-19. Additionally, we were not able to find any difference between using tocilizumab in association with standard care or standard care alone on mortality in hospitalized patients with COVID-19. A previous review found no positive effect of using tocilizumab in COVID-19. However, this review included non-randomized trials. 21 Of note, non-randomized trials may have confounding factors in the comparative groups which often leads to spurious associations. 22 Relying on such results may lead to the introduction of potentially hazardous interventions into clinical practice.
Tocilizumab, a drug capable of controlling massive inflammation caused by IL-6, has begun to be studied globally.   It should also be emphasized that most of the included studies recruited moderate to severe COVID-19 patients. Therefore, these results should not be generalized to mild COVID-19 patients.
Furthermore, even among patients with moderate to severe COVID-19, more trials are needed to determine the best dosage and timing for initiating tocilizumab. Of note, we did not find a significant effect of tocilizumab on the risk of adverse events.
Although no safety concerns associated with tocilizumab were observed in our analysis, it should be noted that the best dosage and timing for initiating tocilizumab still need to be further investigated. All included studies used the tocilizumab standard dose: 8 mg per kilogram of body weight (one or two doses, up to 800 mg). Another problem that we saw was the heterogeneity of the basic treatment in the comparative groups. There were variations in medications and doses that did not allow us to rule out interference in the final results found for the treatment.
Some observational studies of tocilizumab treatment have described reduction in the need for invasive mechanical ventilation, or death. Many trials claimed that using tocilizumab in early stages may alter the results. In our subgroup analysis this evidence was not confirmed. Time from beginning of the disease ends just when the inflammatory stage begins and the latter is the bigger problem.
All included studies had limitations related to blinding and treatment allocation. This is another concern of ours and, combined with the degree of moderate certainty that we found, suggests the need for new RCTs.
We suggest carrying out new quality RCTs, with a balanced use of comedications in both groups, so that the question can be answered more robustly. These studies should be standardized as to the basic parameters for describing clinical trial results, such as using the CONSORT Statement (Consolidated Standards of Reporting Trials).

CONCLUSIONS
The best evidence available showed no difference between tocilizumab plus standard care compared to standard care alone for reducing mortality in patients with COVID-19. However, as a further result with a practical implication, the use of tocilizumab in association to standard care seemed to reduce the risk of progressing to mechanical ventilation in those patients. There is a need for further high-quality randomized double-blind studies using rigorous methodology.