key: cord-0852617-ymnvwwjw
authors: Rubio-Rivas, M.; Mora-Lujan, J. M.; Montero, A.; Homs, N. A.; Rello, J.; Corbella, X.
title: Beneficial and Harmful Outcomes of Tocilizumab in Severe COVID-19: A Systematic Review and Meta-Analysis
date: 2020-09-08
journal: nan
DOI: 10.1101/2020.09.05.20188912
sha: 374c4c03884b19481a03c1fc103239ecafdd3978
doc_id: 852617
cord_uid: ymnvwwjw

Background: Pending for randomized control trials, the use of tocilizumab (TCZ) in COVID-19 is based on observational studies and remains controversial. Purpose: To summarize evidence about the effect of TCZ to treat severe COVID-19. Data sources: PubMed (via MEDLINE), Scopus, and medRxiv repository databases from 1 January to 21 August 2020. Study Selection: Observational studies in any language reporting efficacy and safety outcomes of TCZ use in hospitalized adults with COVID-19. Data Extraction: Independent, dually performed data extraction and quality assessments. Data synthesis: Of 57 eligible studies, 27 were controlled and 30 were not. The overall included patients were 8,128: 4,021 treated with TCZ, in addition to standard of care (SOC), and 4,107 only receiving SOC. The pooled mortality was lower in the TCZ-group vs. the control group, with a relative risk (RR) of 0.73 (95%CI 0.57-0.93; p=0.010). The overall NNT to avoid one death was 20. In hospital wards, patients in the TCZ-group were transferred to intensive care unit (ICU) in a higher proportion than those in the control group; however, ICU mortality of the TCZ-group was lower than in the control group. Secondary infections occurred in a higher proportion in TCZ-treated patients. Among survivors, the length of stay was similar in both groups. Limitations: Conclusions should be considered as weak evidence since they are based on observational studies, most of them retrospective. A variety of factors influencing the indication and effect of TCZ could not be evaluated in-depth. Conclusions: TCZ to seem beneficial in preventing in-hospital mortality in severe, non-critically ill COVID-19 patients. Conversely, patients receiving TCZ appear to be at higher risk for secondary infections, especially those admitted to ICU.

Since early 2020, when the SARS-CoV-2 pandemic hit the world, a variety of treatments have been suggested as potentially useful for COVID-19 illness (1) (2) (3) (4) (5) . However, to date, only remdesivir (4) and dexamethasone (5) have demonstrated evidence-based efficacy on randomized, controlled clinical trials (RCTs). This double strategy combining antiviral and immunomodulatory therapy is in accordance with the two different pathological mechanisms that appear to coexist in the COVID-19 disease; the first triggered by the virus itself and the second by the cytokine storm and systemic dysregulated hostimmune hyperinflammatory response (6) .

While the pandemic continues to spread globally, a worrying 15% of patients will continue to transit into the most severe stage of the disease, requiring hospitalization or intensive care unit (ICU) admission. This advanced clinical stage presents as severe pulmonary injury and multi-organ failure, causing fatality in nearly half of cases, resembling complications from CAR T cell therapy (7) .

Among other pro-inflammatory cytokines, IL-6 plays a part in innate immunity, but excessive production by the host facing SARS-CoV-2 is detrimental (8) (9) (10) . Accordingly, the use of immunomodulatory agents such as tocilizumab (TCZ), a monoclonal antibody to the recombinant human IL-6 receptor, was initially reported as successful among 21 patients in China in February 2020 (11) . Since then, an emerging number of observational studies from America and Europe have been published or registered assessing the effect of TCZ on clinical practice in severe COVID-19 . In most of them, the authors report an association between earlier use of TCZ and reduced mortality; however, interpretation of these results is limited because several of them did not describe a comparison group or specify an a priori comparison.

Conversely, in view of preliminary results from the industry-sponsored Phase 3 COVACTA trial (ClinicalTrials.gov Identifier NCT04320615), the COVID-19 Treatment Guidelines Panel by the National Institute of Health, has taken a position against the use of TCZ in COVID-19 (68) . This randomized clinical trial (RCT) is the first global, randomised, double-blind, placebo-controlled phase III trial investigating TCZ in hospitalised patients with severe COVID-19, but it failed to improve clinical status, as primary endpoint, or several key secondary outcomes such as 4-week mortality (69) .

In the current emergency, while waiting for additional data from ongoing RCTs, most institutions and physicians worldwide are still tackling COVID-19 based only on realworld reported data. Our study aimed to summarize the updated results from available observational studies on the effect of TCZ on clinical outcomes in hospitalized patients with COVID-19. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. . https://doi.org/10.1101/2020.09.05.20188912 doi: medRxiv preprint

This report describes the results of a systematic review and meta-analysis following the guidance of the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement (70) . The protocol was published in the National Institute for Health Research international register of systematic reviews (PROSPERO); registration number CRD42020204934. A clinical question under the PICO framework format (Population-Intervention-Comparison-Outcome) was created (Appendix Table 1 ).

The search strategy was developed by three investigators (M.R-R., J.R., X.C.), which was revised and approved by the other investigators (J.M.M-L., A.M, N.A.H.). We searched the following databases from 1 January to 21 August 2020 (date of the last search): MEDLINE database through the PubMed search engine, Scopus, and the medRxiv repository (www.medrxiv.org), using the terms "COVID-19" [MesH]) AND "Tocilizumab" [MesH] .

Full-text observational studies in any language reporting beneficial or harmful outcomes from the use of TCZ in adults hospitalized with COVID-19 were included. Two investigators (M.R-R., J.M.M-L.) independently screened each record title and abstract for potential inclusion. Restriction of publication type was manually applied: secondary analyses of previously reported trials, protocols, abstracts-only and experimental studies were excluded. Potentially relevant articles were retrieved for full-text review. Two investigators (M.R-R., A.M.) then read the full text of the records the abstracts of which had been selected by at least 1 investigator. Discrepancies were resolved through discussion or by a third investigator (X.C.).

Publications were included if they met all the following criteria: 1) the study reported data on adults with COVID-19, diagnosed by polymerase chain reaction (PCR), admitted hospital-wide or in ICUs; 2) the study design was an observational investigation providing real-world original data on TCZ use in COVID-19, either intravenous or subcutaneous, at recommended doses of 400-800 mg or 162-324 mg, respectively; 3) the study data collection finished after 1 January 2020; and 4) the study provided data related to all-cause in-hospital mortality as pre-specified primary clinical outcome.

Those studies reported to be "case-control studies", in which subjects from the control group also presented COVID-19, just as those from the TCZ group, were also included in the present review. Studies focusing on a sole subgroup of patients (e.g. renal transplant recipients) were excluded. Furthermore, those studies with overlapping data (e.g. the same series reported in different studies) were rejected to avoid bias due to data overexpression. In such cases, the latest and largest study was selected. For this purpose, a careful revision was performed of patients' origins included in studies from the same country. The search was completed by the bibliography review of every paper selected for full-text examination. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. . https://doi.org/10.1101/2020.09.05.20188912 doi: medRxiv preprint

Two investigators (M.R-R., J.M.M-L.) independently abstracted the following details: study characteristics, including setting; intervention or exposure characteristics, including medication dose and duration; patient characteristics, including severity of disease; and outcomes, including mortality, admission to ICU, adverse events such as secondary infections, and length of hospital stay. Discrepancies were resolved by discussion in consultation with a third investigator (X.C.).

Quality assessment was performed using the Newcastle-Ottawa Scale (NOS) for observational studies (71) . Risk of bias was assessed for each included study independently by two investigators (M.R-R., N.A.H). This assessment was based on the Cochrane Handbook of SR of interventions (72) , and using the Cochrane Review Manager 5.3 risk of bias tool which takes account of allocation sequence generation, incomplete outcome reporting, concealment of allocation, masking of participants and investigators, selective outcome reporting or other sources of bias. Each potential source of bias was graded to determine whether studies were considered at high, low, or moderate risk of bias. In case of disagreement, a third author (J.R.) independently determined the quality assessments.

Categorical variables were described as absolute numbers and percentages. When the number of events and sample size was small (and followed a Poisson distribution), confidence intervals were estimated using Wilson's method (73) (74) (75) . We carried out meta-analysis of the pooled mortality ratio by including all included studies in the analysis. Those studies with a control group were also meta-analysed to assess the relative risk (RR) of mortality in TCZ-treated patients vs. those non-TCZ treated (RR of 1).

The inverse variance-weighted method was initially performed using a fixed-effects model. Heterogeneity between studies was assessed using the Q statistic. The percentage of variability between studies by the Higgins I 2 parameter and betweenstudy variability was measured by Tau 2 parameter and, when confirmed (p≤0.05), the analysis was completed by using the random-effects model. A random-effects model assumes that there is an underlying effect for each study which varies randomly across studies, with the resulting overall effect an average of these (76) . Studies with 0 events were not included in the meta-analysis. Forest plots were depicted accordingly. Publication bias was assessed using the Egger method (77) . Statistical analysis was performed by IBM SPSS Statistics for Windows, Version 26.0. Armonk, NY: IBM Corp.

No funding or sponsorship was received for this study or publication of this article.

The study received the ethics exemption by the Research Ethics Committee of Bellvitge University Hospital. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. . https://doi.org/10.1101/2020.09.05.20188912 doi: medRxiv preprint

A total of 781 articles were identified in our search. Of these, 81 qualified for full-text review following title and abstract screening, of which 57 were included in the SRMA. The PRISMA flow diagram is detailed in Figure 1 . The risk of bias of all the studies included is shown in Figure 2 .

The majority of included studies were carried out in different hospitals in high-income countries in America and Europe, such as the US ( (11, 12) . The distribution of the studies worldwide is shown in Appendix Figure 1 . 

Of the total of 57 studies included, 12 were prospective and 45, retrospective. In 30 of these cohort investigations, a control group was not described (10-38) and 27 added a comparison group (40-67). The overall results provided data from 8,128 hospitalized patients with COVID-19: 4,021 TCZ-treated, in addition to standard of care (SOC) (including 711 patients admitted to ICU), and 4,107 only receiving SOC (including 694 patients admitted to ICU). A comparison between the TCZ-group and the control group is detailed in Appendix Tables 2-7. Of the TCZ group, 2,645 (65.8%) were men with a mean age of 61.8 (SD 6.1) and median age 62.6 [range 59-65], according to the data provided. TCZ was given as a single dose in 2,030/2,952 patients (68.8%), and in 922 (31.2%), as two or more doses.

Concomitantly to TCZ use, additional treatment with steroids was given in 1,560/3,073 patients in the TCZ-group (50.8%) vs. 592/2,733 (21.7%) in the control group (p<0.001). Comparing both groups, remdesivir was used in 37/3,511 (1.05%) vs. 23/3,945 (0.58%) (p=0.023) patients. Finally, administration of TCZ was prescribed a median of 10 days [range [9] [10] [11] after the onset of COVID-19 symptoms, in those studies in which this data was provided. Median follow-up of the overall cohort was 10.3 days [range [12] [13] [14] [15] [16] [17] [18] [19] . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. . https://doi.org/10.1101/2020.09.05.20188912 doi: medRxiv preprint

Hospital-wide (including ICUs) pooled mortality of patients with COVID-19 treated with TCZ was 19.2% (95%CI 16.4-22.5) (I 2 =83.6 Q=305.3 tau 2 =0.23 p<0.001). Egger's method A=-3.354 p<0.001 ( Figure 2 ). In the control group, overall mortality was 27.4% (95%CI 21.1-35.6%) (I 2 =95.9 Q=569.1 tau 2 =0.38 p<0.001). These differences between the TCZgroup and the control group achieved statistical significance (p<0.001). The RR of mortality in the TCZ-group was 0.73 (95%CI 0.57-0.93; p=0.010) (I 2 =77.7 Q=107.4 tau 2 =0.24 p<0.001). Egger's method A=-0.712 p=0.380 ( Figure 3 ). The NNT to avoid one death was 20.

Overall Mortality in high-quality observational studies. More restrictive analysis excluding studies with <20 included patients, NOS <7 and studies with important risk of bias, showed pooled mortality in the TCZ-group to be 18.9% ( is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. . https://doi.org/10.1101/2020.09.05.20188912 doi: medRxiv preprint

Hospital-wide (including ICUs) pooled mortality in TCZ-treated patients in whom TCZ was early administered (<10 days from symptoms onset) was 15.9% (95%CI 10.9-23) (I 2 =74.9 Q=31.9 tau 2 =0.222 p<0.001). Egger's method A=-2.306 p=0.239. Hospital-wide (including ICUs) pooled mortality in patients with COVID-19 illness in whom TCZ was administered later (≥10 days) was 23.3% (95%CI 17.9-30.3) (I 2 =75.5 Q=36.8 tau 2 =0.115 p<0.001). Egger's method A=-2.983 p=0.021. Differences between groups did not achieve statistical significance (p=0.252).

In patients with COVID-19 initially admitted to hospital wards, pooled ICU admission rate after TCZ administration was 17.1% (95%CI 11.5-25.5) (I 2 =90 Q=149.5 tau 2 =0.51 p<0.001). Egger's method A=-3.272 p=0.015. Conversely, the risk for ICU admission in the control group, initially admitted outside the ICUs, was 9.5% (95%CI 2. 

Among survivors, the length of hospital stay in the TCZ-group was a median of 15.3 days [range 12.4-19.4] vs. 14 days [range 9-20] in the control group. These differences were not statistically significant (p=0.953).

. CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. . https://doi.org/10.1101/2020.09.05.20188912 doi: medRxiv preprint

Pending published evidence from RCTs, this systematic review and meta-analysis focused on available real-world observational studies, revealing a beneficial effect of TCZ use in preventing mortality in hospitalized adults with COVID-19. However, the present results also showed a higher relative risk for ICU admission and occurrence of secondary infections in such patients receiving TCZ.

To date, two existing SRMA have summarized current evidence on the beneficial and harmful effects of TCZ in COVID-19. The first by Lan SH Zhang et al., included 7 studies, with no conclusive evidence that TCZ would provide any additional benefit to patients with severe COVID-19 (78) . The second, registered in the medRxiv repository by Boregowda et al., included 16 studies, concluding that the addition of TCZ to SOC might reduce mortality in COVID-19 patients requiring hospitalization (79) .

The present SRMA updated and expanded the revision to 57 observational studies. As expected, most included studies emerged recently and were performed in hospitals from high-income European countries and the US. The finding that most of the included patients were men in their 6-8th decades of life was consistent with what has already been described in the general population requiring hospital admission due to COVID-19 (80) . It is precisely this subgroup of elderly patients who have been identified to be at higher risk of transition to the most serious stage of COVID-19. This manifests by moderate-severe lung injury and systemic hyperinflammatory state.

Although no standard doses have specifically been established for TCZ to be used in COVID-19, a vast majority of included studies used a single dose of 400-800 mg iv or 162-324 mg sc, and a few of them, allowed a second or even a third dose in case of worsening. Interestingly, the present SRMA showed that TCZ was mainly prescribed as second step to treat those patients at risk of transition to a more severe condition, after showing poor response to antiviral agents. Accordingly, in a notable proportion of included studies, TCZ was indicated in combination with other different drugs, mostly corticosteroids. In this respect, in view of the RECOVERY trial (5) which reported significant benefit from the use of steroids in severe COVID-19, it is difficult to distinguish in depth the contribution of TCZ on the final outcomes in such included patients receiving TCZ and steroids. On the other hand, in addition to TCZ, a very small proportion (<1%) of included patients from more recently studies also received remdesivir, as a proven antiviral against SARS-Cov-2.

In addition to concomitant drugs, other relevant factors should be taken into account when considering the impact of TCZ use on clinical outcomes in COVID-19. At has been proposed, the disease displays three stages of increasing severity: stage I (incubation and early infection), stage II (pulmonary involvement), and stage III (systemic hyperinflammation and multi-organ dysfunction), which correspond to distinct clinical findings and outcomes (6) . In this respect, in addition to the age and underlying comorbidities of included patients, one of the most important factors to be considered at clinical level is the severity of the clinical stage when indicating TCZ to treat the illness (80) . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. . https://doi.org/10.1101/2020.09.05.20188912 doi: medRxiv preprint Unfortunately, the severity of included patients could not be accurately inferred from clinical, laboratory, or radiological parameters documented in the included studies of the present SRMA. Therefore, mortality was compared between those patients admitted to hospital wards and ICUs, in those studies in which the hospital site from where TCZ was administered was specified. Logically, patients in whom TCZ was indicated during ICU admission showed higher mortality in comparison to those treated in hospital wards. However, since the COVID-19 pandemic induced an unprecedented influx of patients into the ICUs, the particular emergency situation and resource availability of each hospital involved in the included studies most likely conditioned either the criteria when transferring patients from hospital wards to ICUs, or the ethical decisions related to the withdrawal of life support decisions.

Outside of RCTs, observational data from the included studies only reflected the clinical practice of physicians when indicating TCZ use. Consequently, the present results show that TCZ was mostly prescribed in those more seriously-ill patients presenting at a more advanced stage of COVID-19, with severe lung injury and systemic hyperinflammatory multi-organ failure. In this regard, it would be unfair to infer that the higher the use of TCZ and average of concomitant corticosteroids, the higher the risk of ICU admission or death. Therefore, further RCTs on TCZ use in COVID-19 should clarify the interaction of confounding variables, it being crucial to know which patients are the best candidates to eventually receive TCZ as immunomodulatory agent, as well as the beneficial and harmful effects of its use in the absence of or in combination with steroids.

Moreover, in the particular case of the ICU setting, most of the included studies showed insufficient data to appropriately assess the effects when TCZ was prescribed in critically-ill patients. However, since a majority of such seriously-ill patients with COVID-19 admitted to ICUs are submitted to mechanical ventilation and other multiple invasive procedures, and receive concomitant wide-spectrum antibiotics or steroid treatment, there is a great concern regarding the well-known risk of TCZ favoring the occurrence of life-threatening secondary bacterial, viral or fungal opportunistic infections, as it has also been documented in the present SRMA in up to one-fifth of cases in the TCZ-group (53,81,82).

As mentioned, the first and major limitation of this SRMA on TCZ use in COVID-19 is the lack of data from RCTs. In their absence, the present revision is based on observational studies; therefore, conclusions should be considered as founded on weak evidence. Moreover, a second limitation is the fact that most of the included studies were retrospective in nature. A third limitation is the heterogeneity regarding the study population (I2 index) and the potential risk of detected bias. Fourth, variations in criteria for prescribing TCZ may not be ruled out in the included studies, although most of them indicated TCZ use to treat those patients with severe COVID-19 with systemic hyperinflammatory state. Fifth, important factors influencing the effect of TCZ on clinical outcomes such as the baseline characteristics of the patients included, the average time from symptoms onset to TCZ administration, the clinical severity of the disease at the time of TCZ administration, the doses and the form of administration used, the hospital site from where TCZ was indicated, or the use of concomitant drug regimens could not is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. . https://doi.org/10.1101/2020.09.05.20188912 doi: medRxiv preprint be evaluated in-depth, since they were not uniformly provided by the included studies. Sixth, in the vast majority of included studies, there is a lack of subgroup analyses according to age, sex or underlying conditions, concomitant treatments, requirement of mechanical ventilation or ICU admission, and comparisons between ventilated and nonventilated patients. Finally, there is a wide range in the median time of follow-up after TCZ administration, which hinders assessment of consistent improvement, late-onset adverse events and real in-hospital mortality in those patients with prolonged evolution. Thus, some patients considered "survivors" in some included studies may have ended up dying.

In conclusion, this systematic review and meta-analysis provides updated and extended data on the use of TCZ in COVID-19. Pending evidence from RCTs, this investigation was restricted to current available data from observational studies. The present results showed TCZ to be beneficial in reducing overall in-hospital mortality in adults with COVID-19, with a NNT to save one life of 20. These findings were more apparent in those non-critically-ill COVID-19 patients admitted to hospital wards, receiving TCZ at early stage of the hyperinflammatory response syndrome. By contrast, the TCZ-group appears to be at higher risk for secondary infections, especially in those critically-ill patients admitted to ICU. Notwithstanding these results, conclusions should be considered as weak evidence since they are based on observational studies, most of them retrospective. However, these findings may help physicians and researchers to optimize strategies towards precision medicine when designing further RCTs focused on the use of TCZ.

We dedicate this work to the memory of those patients worldwide who have not survived COVID-19. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. . https://doi.org/10.1101/2020.09.05.20188912 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. . https://doi.org/10.1101/2020.09.05.20188912 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted September 8, 2020. 

International variation in the management of severe COVID-19 patients

Hydroxychloroquine and azithromycin as a treatment of COVID-19: results of an open-label non-randomized clinical trial

A Trial of Lopinavir-Ritonavir in adults hospitalized with severe Covid-19

Compassionate Use of Remdesivir for Patients with Severe Covid-19

Dexamethasone in hospitalized patients with Covid-19. Preliminary report

COVID-19 illness in native and immunosuppressed states: A clinical-therapeutic staging proposal

Acute life-threatening toxicity from CAR T-cell therapy

Should we stimulate or suppress immune responses in COVID-19? Cytokine and anti-cytokine interventions

Elevated Interleukin-6 and Severe COVID-19: A Meta-Analysis

The cytokine release syndrome (CRS) of severe COVID-19 and Interleukin-6 receptor (IL-6R) antagonist Tocilizumab may be the key to reduce the mortality

Effective treatment of severe COVID-19 patients with tocilizumab

Tocilizumab treatment in COVID-19: a single center experience

Early use of tocilizumab in the prevention of adult respiratory failure in SARS -CoV -2 infections and the utilization of interleukin -6 levels in the management

Higher levels of IL-6 early after tocilizumab distinguish survivors from non-survivors in COVID-19 pneumonia: a possible indication for deeper targeting IL-6

Tocilizumab for the treatment of severe COVID-19 pneumonia with hyperinflammatory syndrome and acute respiratory failure: a single center study of 100 patients in Brescia, Italy

Hydroxychloroquine and Tocilizumab Therapy in COVID-19 Patients -An Observational Study

Effects of Tocilizumab in Critically Ill Patients With COVID-19: A Quasi-Experimental Study

COVID-19) Treatment Guidelines. National Institutes of Health

Roche provides an update on the Phase III COVACTA trial of Actemra/RoActemra in hospitalized patients with severe COVID-19 associated pneumonia

The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration

The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomized studies in meta-analysis

Cochrane Handbook for Systematic Reviews of Interventions version 6

Probable inference, the law of succession, and statistical inference

Fiducial Limits for the Poisson Distribution

Statistical Q14 methods in epidemiology

Measuring inconsistency in meta-analyses

Publication bias and dissemination of clinical research

Tocilizumab for severe COVID-19: a systematic review and metaanalysis

Addition of Tocilizumab to the standard of care reduces mortality in severe COVID-19: A systematic review and meta-analysis

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Hospitalized adults with severe COVID-19 -Observational studies Intervention or Exposure Tocilizumab (TCZ) administration ("TCZ-group")

Standard of care ("Control group")

• All-cause in-hospital mortality • Overall mortality restricted to high-quality observational studies is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprintThe copyright holder for this this version posted September 8, 2020. . https://doi.org/10.1101/2020.09.05.20188912 doi: medRxiv preprint Appendix Table 2 . Non-controlled studies. General data. Table 6 . Non-controlled studies. Secondary outcomes: Length of hospital stay, Mortality ICU vs. ward and secondary infections..

ICU admission TCZ n/N (%)

TCZ prescribed at the ward mortality n/N (%)

Xu et al . [11] 15.1 (5.8) 2/20 (10) NA 0/20 (0) Overall 0/20 (0) Luo et al. [12] NA NA 3/7 (42.9) 0/8 (0) NA Antony et al. [13] NA (range 5-10) 9/80 (