key: cord-0686823-te0v91nr authors: Verdugo-Paiva, F.; Acuna, M. P.; Sola, I.; Rada, G. title: Remdesivir for the treatment of COVID-19: A living systematic review date: 2020-09-28 journal: nan DOI: 10.1101/2020.09.27.20202754 sha: fce585812545a51eb2b059111283d3b3c1bd20ec doc_id: 686823 cord_uid: te0v91nr Objective This living systematic review aims to provide a timely, rigorous and continuously updated summary of the evidence available on the role of remdesivir in the treatment of patients with COVID-19 Methods We adapted an already published common protocol for multiple parallel systematic reviews to the specificities of this question. Eligible studies were randomised trials evaluating the effect of remdesivir versus placebo or no treatment. We conducted searches in the LOVE (Living OVerview of Evidence) platform for COVID-19, a system that maps PICO questions to a repository maintained through regular searches in electronic databases, preprint servers, trial registries and other resources relevant to COVID-19. All the searches covered the period until 25 August 2020. No date or language restrictions were applied. Two reviewers independently evaluated potentially eligible studies according to predefined selection criteria, and extracted data on study characteristics, methods, outcomes, and risk of bias, using a predesigned, standardised form. We performed meta-analyses using random-effect models and assessed overall certainty in evidence using the GRADE approach. A living, web-based version of this review will be openly available during the COVID-19 pandemic. We will resubmit it every time the conclusions change or whenever there are substantial updates. Results Our search strategy yielded 574 references. Finally, we included 3 randomised trials evaluating remdesivir in addition to standard care versus standard care alone. The evidence is very uncertain about the effect of remdesivir on mortality (RR 0.7, 95% CI 0.46 to 1.05; very low certainty evidence) and the need for invasive mechanical ventilation (RR 0.69, 95% CI 0.39 to 1.24; very low certainty evidence). On the other hand, remdesivir likely results in a large reduction in the incidence of adverse effects in patients with COVID-19 (RR 1.29, 95% CI 0.58 to 2.84; moderate certainty evidence). Conclusions The evidence is insufficient for the outcomes critical for making decisions about the role of remdesivir in the treatment of patients with COVID-19, so it is not possible to balance the potential benefits, if any, with the adverse effects and costs. PROSPERO Registration number CRD42020183384 Keywords COVID-19, Coronavirus disease, Severe Acute Respiratory Syndrome Coronavirus 2, Coronavirus Infections, Systematic Review, Remdesivir, Antivirals We conducted searches in the L·OVE (Living OVerview of Evidence) platform for COVID-19, a system 23 that maps PICO questions to a repository maintained through regular searches in electronic 24 databases, preprint servers, trial registries and other resources relevant to COVID-19. All the 25 searches covered the period until 25 August 2020. No date or language restrictions were applied. 26 Two reviewers independently evaluated potentially eligible studies according to predefined selection 27 criteria, and extracted data on study characteristics, methods, outcomes, and risk of bias, using a 28 predesigned, standardised form. 29 We performed meta-analyses using random-effect models and assessed overall certainty in evidence 30 using the GRADE approach. 31 A living, web-based version of this review will be openly available during the COVID-19 pandemic. 32 We will resubmit it every time the conclusions change or whenever there are substantial updates. 33 per 1000 60 per 1000 26 less (34 less to 4 more) Very low (1) The evidence is very uncertain about the effect of remdesivir on mortality in patients with COVID-19 RR 0.69 (0.39 to 1.24) 1659 patients in 3 study [24, 25, 26] 116 per 1000 80 per 1000 36 less (44 less to 6 more) Very low (2) The evidence is very uncertain about the effect of remdesivir on the need for invasive mechanical ventilation in patients with Adverse effects leading to discontinuation RR (7) Remdesivir likely reduces the number of serious adverse effects CI: confidence interval; RR: Risk ratio; MD: Mean difference; GRADE: Grading of Recommendations Assessment, Development and Evaluation. *Other trial reported length of hospital stay, but data was not usable in meta-analysis 1 -The certainty of the evidence is based in the following judgments: Risk of bias: downgraded in one level since the overall risk of bias for studies was evaluated as 'high' and 'some concerns'; Inconsistency: downgraded in one level for inconsistency since the studies show contradictory results; Indirectness: no concerns; Imprecision: downgraded in one level for imprecision since each end of the confidence interval would lead to different conclusions; Publication bias: no concerns. 2-The certainty of the evidence is based in the following judgments: Risk of bias: downgraded in one level since the overall risk of bias for studies was evaluated as 'high' and 'some concerns'; Inconsistency: no concerns; Indirectness: no concerns; Imprecision: downgraded in two levels for imprecision since each end of the confidence interval would lead to widely different conclusions; Publication bias: no concerns. 3-The certainty of the evidence is based in the following judgments: Risk of bias: downgraded in one level since the overall risk of bias for studies was evaluated as 'high' and 'some concerns'; Inconsistency: no concerns; Indirectness: no concerns; Imprecision: no concerns; Publication bias: no concerns. 4-The certainty of the evidence cannot be estimated since the studies did not report this outcome. It is highly likely that the outcome was measured in the studies. 5-The certainty of the evidence is based in the following judgments: Risk of bias: downgraded in one level since the overall risk of bias for studies was evaluated as 'high' and 'some concerns'; Inconsistency: no concerns; Indirectness: no concerns; Imprecision: downgraded in one level for imprecision, since each end of the confidence interval would lead to different conclusions; Publication bias: no concerns. 6-The certainty of the evidence cannot be estimated since the studies did not report this outcome. It is highly likely that the outcome was measured in the studies. 7-The certainty of the evidence is based in the following judgments: Risk of bias: downgraded in one level since the overall risk of bias for studies was evaluated as 'high' and 'some concerns'; Inconsistency: no concerns; Indirectness: no concerns; Imprecision: no concerns; Publication bias: no concerns. patients, it has been reported to be higher than 10% in some centres [8] . 78 One of the strategies underway to identify effective interventions for COVID-19 is repurposing drugs 79 that have been used for the treatment of other diseases. 80 Remdesivir is among these investigational medications. It is a directly acting antiviral agent, initially 81 developed for the treatment of Ebola virus during the 2014 outbreak in Western Africa [9] . 82 Remdesivir displays antiviral activity against many RNA viruses including SARS-CoV-2, in both in vitro 83 [10] and animal studies [11] . 84 Following the publication of the ACTT-1, a trial conducted by the National Institute of Allergy and 85 Infectious Diseases (NIAID), the US Food and Drug Administration issued an emergency use 86 authorisation of remdesivir for the treatment of COVID-19 [12] . 87 However, the results of ACTT-1 were questioned immediately, particularly for the decision to stop it 88 early for benefit [13] . On the other hand, the decision of the government of the United States of 89 buying virtually all stocks of the drug, generated an urgent need of independent, transparent 90 information about the effects of remdesivir for COVID-19. 91 Using innovative and agile processes, taking advantage of technological tools, and resorting to the 92 collective effort of several research groups, this living systematic review aims to provide a timely, 93 rigorous and continuously updated summary of the evidence available on the effects of remdesivir in 94 patients with COVID-19 . 95 96 This manuscript complies with the 'Preferred Reporting Items for Systematic reviews and Meta-98 Analyses' (PRISMA) guidelines for reporting systematic reviews and meta-analyses [14] (see 99 Appendix 1 -PRISMA Checklist). 100 A protocol stating the shared objectives and methodology of multiple evidence syntheses 101 (systematic reviews and overviews of systematic reviews) to be conducted in parallel for different 102 . 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 28, 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 28, 2020. . https://doi.org/10.1101/2020.09.27.20202754 doi: medRxiv preprint report of the project is updated regularly on the website [18] . 138 The repository is continuously updated [18] and the information is transmitted in real time to the 139 L·OVE platform, however, it was last checked for this review the day before release on 25 August 140 2020. The searches covered the period from the inception date of each database, and no study 141 design, publication status or language restriction was applied. 142 The following strategy was used to retrieve from the repository the articles potentially eligible for 144 this review. 145 coronavir* OR coronovirus* OR betacoronavir* OR "beta-coronavirus" OR "beta-coronaviruses" OR 146 "corona virus" OR "virus corona" OR "corono virus" OR "virus corono" OR hcov* OR covid* OR 147 "2019-ncov" OR cv19* OR "cv-19" OR "cv 19" OR "n-cov" OR ncov* OR (wuhan* and (virus OR 148 viruses OR viral)) OR sars* OR sari OR "severe acute respiratory syndrome" OR mers* OR "middle 149 east respiratory syndrome" OR "middle-east respiratory syndrome" OR "2019-ncov-related" OR "cv-150 19-related" OR "n-cov-related" AND (remdesivir* OR "GS-5734" OR "GS 5734" OR GS5734*) 151 152 In order to identify articles that might have been missed in the electronic searches, we proceeded as 154 follows: 155 • Screened the reference lists of other systematic reviews. 156 • Scanned the reference lists of selected guidelines, narrative reviews and other documents. 157 158 We included randomised controlled trials evaluating patients infected with SARS-CoV-2 of any 160 The intervention of interest was remdesivir at any dosage, duration, timing or route of 162 administration. The comparison of interest was placebo (remdesivir plus standard of care versus 163 placebo plus standard of care) or no treatment (remdesivir plus standard of care versus standard of 164 care). 165 Our primary outcome of interest was all-cause mortality at longest follow-up. Secondary outcomes 166 were invasive mechanical ventilation and adverse effects leading to discontinuation. 167 We also extracted information on the following outcomes: time to viral clearance, length of hospital 168 stay and serious adverse effects. 169 . 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 28, 2020. The results of the searches in the individual sources were de-duplicated by an algorithm that 175 compares unique identifiers (database ID, DOI, trial registry ID), and citation details (i.e. author 176 names, journal, year of publication, volume, number, pages, article title, and article abstract). Then, 177 the information matching the search strategy was sent in real-time to the L·OVE platform where at 178 least two authors independently screened the titles and abstracts yielded against the inclusion 179 criteria. We obtained the full reports for all titles that appeared to meet the inclusion criteria or 180 required further analysis and then decided about their inclusion. 181 We recorded the reasons for excluding trials in any stage of the search and outlined the study 182 selection process in a PRISMA flow diagram which we adapted for the purpose of this project. 183 184 Using standardised forms, two reviewers independently extracted the following data from each 186 included trial: study design, setting, participant characteristics (including disease severity and age) 187 and study eligibility criteria; details about the administered intervention and comparison, including 188 dose, duration and timing (i.e. time after diagnosis); the outcomes assessed and the time they were 189 measured; the source of funding of the study and the conflicts of interest disclosed by the 190 investigators; the risk of bias assessment for each individual study. 191 We resolved disagreements by discussion, with one arbiter adjudicating unresolved disagreements. 192 193 The risk of bias for each randomised trial was assessed by using the 'risk of bias' tool (RoB 2.0: a 195 revised tool to assess risk of bias in randomised trials) [19] , considering the following domains of bias 196 for each outcome result of all reported outcomes and time points: bias due to (1) 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 28, 2020. . https://doi.org/10. 1101 For dichotomous outcomes, we expressed the estimate of treatment effect of an intervention as risk 205 ratios (RR) along with 95% confidence intervals (CI). 206 For continuous outcomes, we used the mean difference and standard deviation to summarise the 207 data along with 95% CI. For continuous outcomes reported using different scales, the treatment 208 effect was expressed as a standardised mean difference with 95% CI. For any outcomes where data was available from more than one trial, we conducted a formal 217 quantitative synthesis (meta-analysis) for studies clinically homogeneous using RevMan 5 [20], using 218 the inverse variance method with the random-effects model. We assessed inconsistency by visual 219 inspection of the forest plots and using the I² index. 220 221 As few trials were found, we did not perform sensitivity or subgroup analysis. 223 The certainty of the evidence for all outcomes was judged using the Grading of Recommendations 225 Assessment, Development and Evaluation working group methodology (GRADE Working Group) [21] , 226 across the domains of risk of bias, consistency, directness, precision and reporting bias. For the main 227 comparisons and outcomes, we prepared a Summary of Findings (SoF) tables [22] , [23] . 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 28, 2020. . https://doi.org/10.1101/2020.09.27.20202754 doi: medRxiv preprint We conducted searches using L·OVE (Living OVerview of Evidence) platform for COVID-19, a system 241 that maps PICO questions to a repository, maintained through regular searches in 27 databases, 242 preprint servers, trial registries and websites relevant to COVID-19. All the searches covered the 243 period until 25 August 2020. No date or language restrictions were applied. 244 The search in the L·OVE platform yielded 574 records after removal of duplicates. We considered 245 489 as potentially eligible and obtained and evaluated their full texts. We finally included 3 246 randomised trials (11 references) [24],[25], [26] . 247 The reasons for excluding studies at the time of full-text review were the following: not a primary 248 study in humans (396 records); wrong study design (51 records) and wrong comparison (3 records). 249 We also identified 16 ongoing randomised trials. 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 28, 2020. . https://doi.org/10.1101/2020.09.27.20202754 doi: medRxiv preprint The three trials identified were the Adaptive COVID-19 Treatment Trial (ACTT-1 [24]), the CAP-China 258 remdesivir 2 [25] and SIMPLE 2 [26]. All trials evaluated inpatient adults. ACTT-1 required for 259 inclusion that one of the following criteria were also fulfilled: SpO2 94% on room air Table 1 and 2 summarises inclusion criteria of the trials and 266 characteristics of the intervention. More details are presented in Appendix 2. 267 Table 1 presents the complete inclusion criteria of the trials. 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 28, 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 28, 2020. We judge that the overall risk of bias was "high" for all outcomes regarding the ACTT-1 trial [24] . The 293 study was judged to raise "some concerns'' in deviations from the intended intervention domain and 294 "high" in bias due to missing outcome data. CAP-China remdesivir 2 trial overall risk of bias was 295 . 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 28, 2020. . https://doi.org/10.1101/2020.09.27.20202754 doi: medRxiv preprint overall risk of bias was some concern for all outcomes due to deviations from intended interventions 297 [26]. Table 4 summarises the risk of bias assessments and details of each assessment are presented 298 in Appendix 2. 299 300 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 28, 2020 . . https://doi.org/10.1101 All studies reported this outcome [24] , [25],[26] and the evidence is very uncertain about the effect 320 of remdesivir on the need for invasive mechanical ventilation (RR 0.69, 95% CI 0.39 to 1.24; very low 321 certainty evidence). 322 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 28, 2020 . . https://doi.org/10.1101 the authors from the study data). It is unfortunate not knowing yet if one of the pharmaceutical interventions that has sparked more 361 interest is effective or not. One of the limitations comes from the lack of precision of the result for 362 the main outcomes. The early termination of the ACTT-1 trial can be seen as a missed opportunity in 363 this regard [25] . 364 In addition, all the trials concluded enrollment before the release of the RECOVERY trial which 365 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 28, 2020. . https://doi.org/10. 1101 There are at least 46 ongoing trials that we expect will provide data in the near future. Making sense 374 of this information is not going to be an easy task. Systematic reviews are considered the gold 375 standard to make sense of multiple trials addressing a similar scientific question, but the traditional 376 model for conducting reviews has several limitations, including a high demand for time and 377 resources [28] and a rapid obsolescence [29] . Amidst the COVID-19 crisis, researchers should make 378 their best effort to answer the urgent needs of health decision makers yet without giving up 379 scientific accuracy. Information is being produced at a vertiginous speed [30] , so alternative models 380 are needed. 381 One potential solution to these shortfalls are rapid reviews, a form of knowledge synthesis that 383 streamlines or omits specific methods of a traditional systematic review in order to move faster. 384 Unfortunately, in many cases, this rapidity comes at the cost of quality [31] . Furthermore, they do 385 not solve the issue of obsolescence. Living systematic reviews do address that issue [32] . They are 386 continually updated by incorporating relevant new evidence as it becomes available, at a substantial 387 effort. So, an approach combining these two models might prove more successful in providing the 388 scientific community and other interested parties with evidence that is actionable, rapidly and 389 efficiently produced, up to date, and of the highest quality [33] . 390 391 This review is part of a larger project set up to put such an approach into practice. The project aims 392 to produce multiple parallel living systematic reviews relevant to COVID-19 following the higher 393 standards of quality in evidence synthesis production [15] . We believe that our methods are well 394 suited to handle the abundance of evidence that is to come, including evidence on the role of 395 lopinavir/ritonavir for During the COVID-19 pandemic, we will maintain the search and selection of evidence for this review 397 continuously updated, and we will every time the conclusions change or whenever there are 398 substantial updates. Our systematic review aims to provide high-quality, up-to-date synthesis of the 399 evidence that is useful for clinicians and other decision-makers. All the review authors drafted and revised the protocol, conducted article screening and data 411 collection, and drafted and revised the review. 412 The COVID-19 L·OVE Working Group was created by Epistemonikos and a number of expert teams in 413 order to provide decision makers with the best evidence related to COVID-19. Up-to-date 414 information about the group and its member organisations is available here: 415 epistemonikos.cl/working-group 416 417 All authors declare no financial relationships with any organisation that might have a real or 419 perceived interest in this work. There are no other relationships or activities that might have 420 influenced the submitted work. As researchers will not access information that could lead to the identification of an individual 432 participant, obtaining ethical approval was waived. 433 434 All data related to the project will be available. Epistemonikos Foundation will grant access to data. 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 28, 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 28, 2020. . https://doi.org/10. 1101 -the-media-briefing-on-2019-ncov-on The continuing 2019-nCoV epidemic threat of novel 443 coronaviruses to global health -The latest 2019 novel coronavirus outbreak in Wuhan An interactive web-based dashboard to track COVID-19 in real 446 time Clinical Characteristics of Coronavirus Disease 2019 in China Emerging Infectious Disease in the 21st Century novel coronavirus patients' 453 clinical characteristics, discharge rate and fatality rate of meta-analysis Global Covid-19 Case Fatality Rates [Internet] UK: Centre for Evidence-Based Medicine 456 Clinical, laboratory and 459 imaging features of COVID-19: A systematic review and meta-analysis. Travel medicine and 460 infectious disease Adenine C-Nucleoside (GS-466 5734) for the Treatment of Ebola and Emerging Viruses Clinical benefit of remdesivir in rhesus macaques infected 478 with SARS-CoV-2 Food and Drug Administration. Coronavirus (COVID-19) Update: FDA Issues Emergency 480 Use Authorization for Potential COVID-19 Treatment Inside the NIH's controversial decision to stop its big remdesivir study Preferred reporting items for 483 systematic reviews and meta-analyses: the PRISMA statement Evidence synthesis relevant to COVID-19: a protocol for multiple 487 systematic reviews and overviews of systematic reviews Remdesivir for the treatment of COVID-19: A living 490 systematic review protocol Methods for the special L·OVE of Coronavirus infection RoB 2: a revised tool 496 for assessing risk of bias in randomised trials Review Manager (RevMan) [Software]. Version 5.3.5 Copenhagen: The Nordic Cochrane 499 The Cochrane Collaboration GRADE: an emerging consensus on rating quality of evidence and strength of 502 GRADE guidelines: 13. Preparing summary of 508 findings tables and evidence profiles-continuous outcomes ACTT-1 Study Group Members. Remdesivir for the Treatment of Covid The New England journal of medicine Remdesivir in adults with severe COVID-19: a 524 randomised, double-blind, placebo-controlled, multicentre trial. The Lancet Effect of Remdesivir vs Standard Care on Clinical Status 531 at 11 Days in Patients With Moderate COVID-19: A Randomized Clinical Trial