key: cord-0328614-4spnkuel
authors: Zeraatkar, D.; Cusano, E.; Diaz Martinez, J. P.; Qasim, A.; Mangala, S. O.; Kum, E.; Bartoszko, J. J.; Devji, T.; Agoritsas, T.; Lamontagne, F.; Rochwerg, B.; Vandvik, P. O.; Brignardello-Petersen, R.; Siemieniuk, R.
title: Tocilizumab and sarilumab alone or in combination with corticosteroids for COVID-19: A systematic review and network meta-analysis
date: 2021-07-07
journal: nan
DOI: 10.1101/2021.07.05.21259867
sha: e88c7a17ca4cb32c059cf87464b4f58a269b5234
doc_id: 328614
cord_uid: 4spnkuel

Objective: To compare the effects of interleukin-6 (IL-6) receptor blockers, with or without corticosteroids, on mortality in patients with COVID-19. Design: Systematic review and network meta-analysis Data sources: WHO COVID-19 database, a comprehensive multilingual source of global covid-19 literature, and two prospective meta-analyses Study selection: Trials in which people with suspected, probable, or confirmed COVID-19 were randomized to IL-6 receptor blockers (with or without corticosteroids), corticosteroids, placebo, or standard care. Results: We assessed the risk of bias of included trials using a modification of the Cochrane risk of bias 2.0 tool. We performed a Bayesian fixed effect network meta-analysis and assessed the certainty of evidence using the GRADE approach. We identified 45 eligible trials (20,650 patients), 36 (19,350 patients) of which could be included in the network meta-analysis. 27 of 36 trials were rated at high risk of bias, primarily due to lack of blinding. Tocilizumab (20 more per 1000, 15 fewer to 59 more; low certainty) and sarilumab (11 more per 1000, 38 fewer to 55 more; low certainty) alone may not reduce the risk of death. Tocilizumab, in combination with corticosteroids, probably reduces the risk of death compared to corticosteroids alone (35 fewer per 1000, 52 fewer to 18 more; moderate certainty) and sarilumab, in combination with corticosteroids, may reduce the risk of death compared to corticosteroids alone (43 fewer, 73 fewer to 12 more; low certainty). Tocilizumab and sarilumab, both in combination with corticosteroids, may have similar effects (8 more per 1000, 20 fewer to 35 more; low certainty). Conclusion: IL-6 receptor blockers, when added to standard care that includes corticosteroids, in patients with severe or critical COVID-19, probably reduce mortality. Tocilizumab and sarilumab may have similar effectiveness.

What is already known on this topic? 91  IL-6 receptor blockers have immunosuppressive effects that may be important in COVID-19 92 patients with immune system dysfunction and inflammation 93  Corticosteroids reduce the risk of death in patients with severe or critical COVID-19 94

What this study adds 95  Our systematic review and network meta-analysis provides a comprehensive review of the 96 evidence addressing the effects of IL-6 receptor blockers, alone or in combination with 97 corticosteroids, in COVID-19 98  IL-6 receptor blockers when added to a standard care that includes corticosteroids, in patients 99

with severe or critical COVID-19, probably reduce mortality. 100

 Tocilizumab and sarilumab in combination with corticosteroids may have similar effectiveness 101 for reducing mortality. 102 103 6 ncov/). Prior to its merge with the WHO COVID-19 database on 9 October 2020, we searched the US 134

The database includes, but is not limited to the following bibliographic and grey literature sources: 136

Medline (Ovid and PubMed), PubMed Central, Embase, CAB Abstracts, Global Health, PsycInfo, Cochrane 137 Library, Scopus, Academic Search Complete, Africa Wide Information, CINAHL, ProQuest Central, 138 SciFinder, the Virtual Health Library, LitCovid, WHO covid-19 website, CDC covid-19 website, 139

Eurosurveillance, China CDC Weekly, Homeland Security Digital Library, ClinicalTrials.gov, bioRxiv 140 (preprints), medRxiv (preprints), chemRxiv (preprints), and SSRN (preprints). Our search also includes six 141 In addition, we include data directly from trialists via two WHO-sponsored prospective meta-analyses (11, 152 15). We report our full strategy as a supplement to our drug therapy publication (2). 153

As part of the living SRNMA, pairs of reviewers, following calibration exercises, worked independently and 155 in duplicate to screen titles and abstracts of search records and subsequently the full texts of records 156 determined potentially eligible at the title and abstract screening stage and to link preprint reports with 157 their subsequent publications based on trial registration numbers, authors, and other trial characteristics. 158

Reviewers resolved discrepancies by discussion, and when necessary, by adjudication with a third-party 159 reviewer. 160

We included preprint and peer reviewed reports of trials that compared IL-6 receptor blockers with 161 standard care, placebo, or corticosteroids or that compared corticosteroids with standard care or placebo 162 in patients with suspected, probable, or confirmed COVID-19. There were no restrictions on severity of 163 illness, setting, or language of publication. 164 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted July 7, 2021. ; https://doi.org/10.1101/2021.07.05.21259867 doi: medRxiv preprint

As part of the living SRNMA, for each eligible trial, pairs of reviewers, following training and calibration 166 exercises, independently extracted trial characteristics (trial registration, publication status, study design), 167 patient characteristics (country, age, sex, type of care, severity of COVID-19 symptoms), and outcomes of 168 interest (means or medians and measures of variability for continuous outcomes and the number of 169 participants analyzed and the number of participants who experienced an event for dichotomous 170 outcomes) using a standardized, pilot tested data extraction form. Reviewers resolved discrepancies by 171 discussion and, when necessary, with adjudication by a third party. We updated our data when a study 172 preprint becomes available as a peer reviewed publication. For this review, we focus on all-cause mortality 173 closest to 90 days. 174

To assess risk of bias, reviewers, following training and calibration exercises, use a revision of the Cochrane 175 tool for assessing risk of bias in randomized trials (RoB 2.0) (16). We present our modified risk of bias tool 176 as a supplement to our drug therapy publication (2). Reviewers resolved discrepancies by discussion and, 177 when necessary, by adjudication with a third party. 178

We performed network meta-analysis using a Bayesian framework to compare tocilizumab with and 180 without corticosteroids, sarilumab with and without corticosteroids, corticosteroids without IL-6 receptor 181 blockers, and standard care or placebo. We used a plausible prior for the variance parameter and a 182 uniform prior for the effect parameter (17). We used three Markov chains with 100,000 iterations after 183 an initial burn-in of 10,000 and a thinning of 10. We used node splitting models to assess local incoherence 184 and to obtain indirect estimates. All network meta-analyses were performed using the gemtc package of 185 R version 3.6.3 (RStudio, Boston, MA) and all pairwise meta-analyses using the bayesmeta package. We 186 produced network plots using the network map command of Stata version 17.0 (StataCorp, College 187 Station, TX) with nodes weighted by the number of studies evaluating each treatment and edges weighted 188 by the inverse variance of the direct estimate (18). 189

We present fixed-effect meta-analyses as the primary analysis and random-effects meta-analyses as 190 sensitivity analysis because we found the estimates from random-effects to have credible intervals that 191 were implausibly wide due to the uncertainty around the heterogeneity estimate. 192

We summarized the effect of interventions on mortality using odds ratios and corresponding 95% credible 193 interval. 194 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted July 7, 2021. We assessed the certainty of evidence using GRADE approach for network meta-analysis, using a 200 minimally contextualized approach, with a null effect as the threshold of importance (22-25). The 201 minimally contextualized approach considers only whether credible intervals include the null effect and 202 thus does not consider whether plausible effects, captured by credible intervals, include both important 203 and trivial effects. To evaluate certainty of no benefit (or no effect), we used a 1% risk difference threshold 204 of the 95% credible interval for mortality. We decided on this preliminary threshold based on a survey of 205 the authors. 206

Two reviewers with experience in applying the GRADE approach rated each domain for each comparison 207 separately and resolved discrepancies by consensus. We rated the certainty for each comparison and 208 outcome as high, moderate, low, or very low, based on considerations of risk of bias, inconsistency, 209 indirectness, publication bias, intransitivity, incoherence (difference between direct and indirect effects), 

Study characteristics 220

We screened 45,854 titles and abstracts and 884 full-texts and identified 45 eligible trials, including 20,650 221 patients (6, 26-48). Figure 1 presents details about study selection. Twenty-one of these trials were 222 published, four were available as preprints, and 20 were unpublished and were retrieved from two 223 prospective meta-analyses (11, 49). 224 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted July 7, 2021. ; https://doi.org/10.1101/2021.07.05.21259867 doi: medRxiv preprint One trial, REMAP-CAP (7), randomized patients to tocilizumab or standard care (among centres with 229 access to tocilizumab) or to sarilumab or standard care (among centres with access to sarilumab). 230

Randomization to standard care was halted midway through the trial when an interim analysis showed 231 efficacy of tocilizumab and sarilumab after which patients were randomized to either tocilizumab or 232 sarilumab, with both groups receiving corticosteroids. We treated REMAP-CAP as three separate trials in 233 our analyses (i.e., tocilizumab versus standard care; sarilumab versus standard care; tocilizumab versus 234 sarilumab). We used 90-day mortality for the comparisons of tocilizumab and sarilumab with standard 235 care and obtained data on in-hospital mortality from the investigators of the trial on the comparison of 236 tocilizumab and sarilumab. 237

Another trial, Sarilumab-COVID-19, was conducted in two phases (35) . Phase three underwent two 238 protocol amendments involving patient eligibility and dose. The trial was thus treated as four separate 239

Patient characteristics 241 Table 1 were at high risk of bias-primarily due to lack of blinding. 250

We included 36 trials, with 19,350 patients and 5,269 deaths, in our network comparing tocilizumab and 252 sarilumab, with or without corticosteroids, and corticosteroids in comparison with standard care or 253 placebo (6, 26-28, 30, 31, 34, 35, 37-40, 42-46, 48). Nine trials could not be included in the analysis because 254 they did not report outcome data, or, for trials that compared IL-6 receptor blockers with standard care 255 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted July 7, 2021. ; https://doi.org/10.1101/2021.07.05.21259867 doi: medRxiv preprint or placebo, we could not retrieve subgroup data based on concomitant treatment with corticosteroids 256 (29, 32, 33, 36, 41, 47). Figure 2 presents the network plot. Supplementary 1 contains data for the network 257 meta-analysis. 258 Table 2 presents results from the network meta-analysis. Tocilizumab and sarilumab alone may not reduce 259 mortality compared to standard care and corticosteroids probably reduce the risk of death. Our findings are consistent with a prospective meta-analysis (11), and the largest trials on IL-6 receptor 278 blockers, RECOVERY and REMAP-CAP (6, 7). Though several smaller trials did not find a benefit with 279 tocilizumab, the totality of the evidence suggests that this is probably because smaller individual trials 280 were underpowered to detect such a modest reduction in mortality. 281

Our study adds to the evidence base by showing that IL-6 receptor blockers may reduce mortality when 282 added to a standard care regimen that includes corticosteroids and that sarilumab may have a similar 283 effect on mortality as tocilizumab. This result is largely driven by the component of the REMAP-CAP trial 284 that directly compared sarilumab to tocilizumab. 285 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted July 7, 2021.

The strengths of this review include the comprehensive search and screening strategy. In addition to trials 287 that we identified as part of our own search, we also included trials from two prospective meta-analyses 288 that included an inception cohort of registered trials thereby minimizing the effects of publication bias 289 (11, 15). 290

Our findings are limited by the risk of bias of the trials, the majority of which were at high risk of bias due 291 to lack of blinding. Lack of blinding may introduce bias through differences in co-interventions between 292 randomized groups. We took a conservative approach and rated down the certainty of evidence for risk 293 of bias. Some, including the linked WHO guideline panel, did not consider lack of blinding to be a serious 294 concern for mortality, because it is an objective outcome (13). 295

We only considered corticosteroid use at the time of randomization and some patients probably received 296 corticosteroids after randomization, but were considered in this review not to have received concomitant 297 corticosteroids. Whether or not IL-6 receptor blockers reduce mortality when not co-administered with 298 corticosteroids remains uncertain. 299

Four trials that we included in our systematic review were only available as preprint publications. Including 300 preprints in meta-analyses may increase the precision of estimates, allow timely dissemination, and may 301 minimize the effects of publication bias. It may, however, reduce the credibility of evidence syntheses and 302 risk serious errors if important differences appear in later published reports. As part of our living 303 systematic review and network meta-analysis, we have been maintaining a comprehensive comparison 304 of differences in key methods and results between preprints and publications. Differences between 305 preprints and peer reviewed publications have mostly been limited to baseline patient characteristics and 306 any changes we have observed so far would not have resulted in a meaningful change to the pooled effect 307 estimates or certainty of evidence (2). 308

Evidence from this systematic review and network meta-analysis shows that IL-6 receptor blockers when 310 added to a standard care that includes corticosteroids, in patients with severe or critical COVID-19, 311

probably reduce mortality. The available evidence suggests that tocilizumab and sarilumab may be 312 similarly effective. Our findings support linked WHO guidelines on IL-6 receptor blockers, which 313 recommends using either tocilizumab or sarilumab for patients with severe or critical COVID-19 (13). 314 315 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted July 7, 2021. ; https://doi.org/10.1101/2021.07.05.21259867 doi: medRxiv preprint 316 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted July 7, 2021. ; https://doi.org/10.1101/2021.07.05.21259867 doi: medRxiv preprint

Tables 317 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted July 7, 2021. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted July 7, 2021. ; https://doi.org/10.1101/2021.07.05.21259867 doi: medRxiv preprint 68 anti-SARS-CoV-2 antibody and cellular therapies 52 traditional Chinese/alternative medicine excluding specific molecules at specific doses 7 exercise/rehabilitation 9 personal protective equipment 5 psychological and educational 71 vaccine 11 oxygen delivery 3 diagnostic imaging 1 external organ support 1 nutrition and supplements 14 other 2 removed from preprint server by study authors 45 unique randomized trials addressing IL-6 receptor blockers and/or corticosteroids 136 records identified from external sources 128 Epistemonikos COVID-19 Evidence 4 reference lists of studies 2 data from authors 2 meta-analyses with 29 unpublished trials . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted July 7, 2021. ; https://doi.org/10.1101/2021.07.05.21259867 doi: medRxiv preprint 324 325 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted July 7, 2021. ; https://doi.org/10.1101/2021.07.05.21259867 doi: medRxiv preprint CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted July 7, 2021. ; https://doi.org/10.1101/2021.07.05.21259867 doi: medRxiv preprint Low risk of bias Probably low risk of bias Probably high risk of bias High risk of bias . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted July 7, 2021. ; https://doi.org/10.1101/2021.07.05.21259867 doi: medRxiv preprint

Drug treatments for covid-19: living systematic review and 334 network meta-analysis

Inflammatory biomarkers as independent prognosticators of 28-day mortality 336 for COVID-19 patients admitted to general medicine or ICU wards: a retrospective cohort study

Epidemiological and clinical characteristics of 99 cases of 2019 novel 339 coronavirus pneumonia in Wuhan, China: a descriptive study

An inflammatory cytokine signature predicts COVID-19 341 severity and survival

Tocilizumab in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, 343 open-label, platform trial

Interleukin-6 Receptor Antagonists in Critically Ill Patients 345 with Covid-19

Efficacy of Tocilizumab in Patients Hospitalized with Covid-347 19

Effect of Tocilizumab vs Usual Care in Adults Hospitalized With 349 COVID-19 and Moderate or Severe Pneumonia: A Randomized Clinical Trial

Effect of Tocilizumab vs Standard Care on Clinical Worsening in 352 Patients Hospitalized With COVID-19 Pneumonia: A Randomized Clinical Trial

Association of administration of interleukin-6 antagonists 355 with mortality and other outcomes among hospitalized patients with COVID-19: a prospective meta-356 analysis

Managing drug shortages during a pandemic: tocilizumab and COVID-358 19

A living WHO guideline on drugs for covid-19

Machine learning for identifying Randomized Controlled Trials: 362 An evaluation and practitioner's guide

Association Between Administration of Systemic Corticosteroids 364 and Mortality Among Critically Ill Patients With COVID-19: A Meta-analysis

RoB 2: a revised tool for assessing risk of bias in randomised trials

Predictive distributions for between-study heterogeneity and 368 simple methods for their application in Bayesian meta-analysis

Graphical tools for network meta-analysis in STATA

A weekly surveillance summary of U.S COVID-372 19 activity 2020

ISARIC (International Severe Acute Respiratory and Emerging Infections Consortium)

The GRADE Working Group clarifies the construct of certainty of 380 evidence

A GRADE Working Group approach for rating the quality 382 of treatment effect estimates from network meta-analysis

Advances in the GRADE approach to rate the 386 certainty in estimates from a network meta-analysis

Effect of Tocilizumab vs Usual Care in Adults Hospitalized 388 With COVID-19 and Moderate or Severe Pneumonia: A Randomized Clinical Trial

Tocilizumab in Hospitalized Patients with Severe Covid-19 391 Pneumonia

Timely Administration of Tocilizumab Improves Survival of 399 Hospitalized COVID-19 Patients

Tocilizumab plus standard care versus standard care in patients 402 in India with moderate to severe COVID-19-associated cytokine release syndrome (COVINTOC): an open-403 label, multicentre, randomised, controlled

Efficacy of Tocilizumab in Patients Hospitalized with Covid-405 19

Tocilizumab in Patients Hospitalized with Covid-19 Pneumonia

Effect of Tocilizumab vs Standard Care on Clinical Worsening in 409 Patients Hospitalized With COVID-19 Pneumonia: A Randomized Clinical Trial

Tocilizumab Failed to Reduce Mortality in Severe COVID-412 19 Patients: Results From a Randomized Controlled Clinical Trial. Respiratory Research. 2021. 413 34. Veiga VC

A Randomized Placebo-Controlled Trial of Sarilumab in 416 Hospitalized Patients with Covid-19

Sarilumab in patients admitted to hospital with severe or critical 418 COVID-19: a randomised, double-blind, placebo-controlled, phase 3 trial. The Lancet Respiratory 419 medicine

The Writing Committee for the REMAP-CAP Investigators. Effect of Hydrocortisone on Mortality and

Organ Support in Patients With Severe COVID-19: The REMAP-CAP COVID-19 Corticosteroid Domain 422 Randomized Clinical Trial

Methylprednisolone in adults hospitalized with 424 COVID-19 pneumonia : An open-label randomized trial (GLUCOCOVID)

Effect of Hydrocortisone on 21-Day Mortality or Respiratory 427

Support Among Critically Ill Patients With COVID-19: A Randomized Clinical Trial

Intravenous methylprednisolone pulse as a treatment for 429 hospitalised severe COVID-19 patients: results from a randomised controlled clinical trial. The European 430 respiratory journal

Evaluation of the Efficacy of Methylprednisolone Pulse Therapy 432 in Treatment of Covid-19 Adult Patients with Severe Respiratory Failure: Randomized, Clinical Trial

No clinical benefit of high dose corticosteroid 438 administration in patients with COVID-19: A preliminary report of a randomized clinical trial

Methylprednisolone as Adjunctive Therapy for Patients 441

Hospitalized With COVID-19 (Metcovid): A Randomised, Double-Blind, Phase IIb, Placebo-Controlled Trial

Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2020. 443 45. Tang X, Feng YM, Ni JX, et al. Early Use of Corticosteroid May Prolong SARS-CoV-2 Shedding in Non-444

Intensive Care Unit Patients with COVID-19 Pneumonia: A Multicenter, Single-Blind

Respiration; international review of thoracic diseases

Effect of Dexamethasone on Days Alive and Ventilator-Free 447 in Patients With Moderate or Severe Acute Respiratory Distress Syndrome and COVID-19: The CoDEX 448 Randomized Clinical Trial

Efficacy of corticosteroid therapy in the treatment of long-450 lasting olfactory disorders in COVID-19 patients

The WHO Rapid Evidence Appraisal for COVID-19 Therapies (REACT) Working Group. Association 454 Between Administration of Systemic Corticosteroids and Mortality Among Critically Ill Patients With 455 COVID-19: A Meta-analysis

Interleukin-6 Receptor Antagonists in Critically Ill Patients