key: cord-337994-pkklt77i
authors: Brouqui, P.; Giraud-Gatineau, A.; Raoult, D.
title: Remdesivir investigational trials in COVID-19: a critical reappraisal
date: 2020-06-07
journal: New Microbes New Infect
DOI: 10.1016/j.nmni.2020.100707
sha: 
doc_id: 337994
cord_uid: pkklt77i

Abstract During outbreak of emerging disease, the most important aim is to discover an effective drug to save life. Consequently, a lot of effort are generally made by the industry to promote clinical trials with new drugs. Here we review evidence of the 8 most recent reports including 3 randomized controlled trials on the clinical efficacy of remdesivir in treating COVID-19 patient. We conclude that it is far too premature to identify remdesivir as a curative or life-saving intervention.

Since the first described infection with the severe acute respiratory syndrome coronavirus 2 (SARS-2 CoV2) in December 2019, the coronavirus disease 2019 (COVID-19) has developed into a pandemic, 3 the symptoms of which range from asymptomatic course to pneumonia, acute lung and multi-organ 4 failure and death. In order to develop a meaningful therapy strategy, different medications are used 5 "off label". One of these is remdesivir, a precursor of a nucleotide analogue that inhibits viral RNA 6 polymerases. As for Ebola, SARS-CoV and Middle East respiratory syndrome coronavirus (MERS-CoV), 7 remdesivir appears to be effective in vitro in SARS-Cov2 (1). Good outcomes have been reported in 8 cases report (2;3). Many studies are ongoing or already published to demonstrate the efficacy of 9 remdesivir on patient with COVID-19, some showing the lack of difference with control arms (4) , 10 some others reporting efficacy but discussed (5-7). Treating patients early in disease has always been 11 a crucial issue in treating potentially life-threatening infectious diseases. The aim of this review 12

presented below was to evaluate the quality of the published and not yet peer-reviewed trials on 13 remdesivir and to highlight pitfalls to inform readers that a careful analysis of reported data is 14 needed to offer a more accurate interpretation of the results. 15

We looks at all scientific paper available as peer and not yet peer reviewed paper in the major 17 literature from data base Pub Med, Web of Knowledge, scholar google and BioRxiv and MedRxiv. The 18 key words were [remdesivir alone or with COVID]. We recover 91 articles in MedRxiv, 81 in BioRxiv 19

and 112 in Pub Med. When we added COVID to remdesivir, PubMed recover 79 articles. On Web of 20

Knowledge remdesivir recover 25 articles. In Scholar Google remdesivir recovered 1480 articles in 21 2020. Of them we selected 17 papers responding to the aims of this article. When available we look 22 at the following endpoints: time to improvement at D14 and 28, death, and adverse events. 23

As today, 8 studies report the use of remdesivir in COVID and are summarized in Table 1. The first is a  25 single case, having received remdesivir on the day 11 of disease, and which on day 12 saw condition 26 improve (stopping oxygenation and oxygen saturation at 96%) (8). 27

The second is a non-yet peer review paper that reports the first 12 case of COVID in the united 28

states. It is a descriptive paper in which 3 of the 7 hospitalized patients received remdesivir for 29 compassionate use for a duration of 4-10 days (9) . All hospitalized patient had serial SARSCov2 RT 30 PCR testing. When reanalyzed, the mean delay in normalization of nasal RT PCR was 8.6 days in 31 remdesivir patient versus 6.75 days (p=0.85) in untreated patient. 32

The third reports a series of 5 cases, 3 of which received at least one dose of remdesivir. In two 33 patients, treatment occurred at the time of the disease's worsening. In one of them, the remdesivir 34 was discontinued after 5 days (ALT elevation and rash). In the third patient, the remdesivir was 35 stopped after a single dose due to renal dialysis to avoid the accumulation of cyclodextrin. Therefore, 36

the authors indicate that they cannot draw any conclusions based on their data as to the potential 37 efficacy of remdesivir in the treatment of COVID-19 (3). 38

The fourth study analyzes the remdesivir treatment of a single patient on the day 13 of his disease 39

(2). At the time of remdesivir administration, the patient was in intensive care, intubated and treated 40 with hydroxychloroquine 400mg/day and azithromycin since 7 days. Forty-eight hours after 41 remdesivir initiation or treatment, the patient's condition had improved. The patient was extubated 42 60 hours after treatment and was able to breathe in the ambient air 24 hours later. 43

The fifth study is an uncontrolled, prospective, open observational study of patients having received, 44

as compassionate used, a 10-day remdesivir therapy with a target follow-up period of 28 days. 45

Between 25.01.2020 and 07.03.2020, 61 patients were included in the study and received at least 46 one dose of remdesivir, some of which may have been part of previous studies. Of those patients, 8 47

were excluded of the study which, in an intention to treat analysis should have been considered as 48 failure. Finally, data from 53 patients were analyzed of whom one was already published in the study 49 N°3 (Lescure et al). Of them 40 received the complete 10-day remdesivir therapy, 10 received 5 to 9-50 day therapy and 3 patients received less than 5 days of remdesivir (7) . On average, COVID-19 51 symptoms lasted 12 days before remdesivir therapy was initiated. In the median follow-up period of 52 18 days, 36 of the 53 patients (68%) were able to improve under Remdesivir. An improvement was 53

shown in all patients who were mild receiving no or only low-dose oxygen supplementation (n = 12), 54 or in 5 of the 7 non-invasive ventilated patients. This also raised an ethical comment on the 55 compassionate used of remdesivir in some patients whom were not engaged in short term. Of the 53 56 patients followed, 10 were treated while they were on ambient air (2) or low flow oxygen ( 8) time to recovery and death at 28 days and 100 % of patient enrolled end the study and were 76 evaluated in both intention to treat (ITT) and per protocol (PP) analysis. Serious adverse event or 77 event leading to stop the drug were reported in 18 and 12 % in remdesivir versus 6 and 5% in 78 placebo demonstrating the poor safety of the drug. Although no significant difference was noted in 79 other treatment between the two groups, in almost all the RCT reporting evaluation of treatment for 80 COVID, patient are also treated with several other drugs such as antibiotics (9), among some have 81 demonstrated antiviral efficacy (11) , corticosteroid, antiviral , and anti-inflammatory among which 82 some anti IL6 seems promising (12) . This may bias the data such as shown in the Hillaker et al study 83

cited above. This questioned the multicentric nature of the randomized controlled studies which is 84 needed by the high number of patients to be enrolled. This is a bias which is difficult to control 85 because it is directly related to the "standard of care" of each center likely to be different in term of 86 equipment, protocols, surveillance, and staff skills. Consequently, the care of patient might not be 87 comparable in between centers and the outcome biased by the expertise of the team in charged. 88

In the preliminary announcement on efficacy of remdesivir on an RCT involving 1061 patients , the 89 NIH said that preliminary results indicate that patients who received remdesivir had a 31% faster 90 time to recovery than those who received placebo (11 days/15 days) but that the survival benefit on 91 1063 patients was insignificant compared to placebo (p=0.059) concluding that remdesivir has an 92 effect but not a wonder effect. In her commentary, Mahase said : ….in time of epidemics… "expedite 93 publication are fine but hinting that results are going to be positive, only benefits the drug companies 94 (6). Fast-flowing, conflicting information on remdesivir in the past few weeks has left people reeling. 95

Recently the paper was released with preliminary reports in the NEJM but with different results the 96 survival benefits becoming significant in the overall analyzed population (13). This conclusion is over 97

interpreted. In the table 2, as mentioned, the hazard ratio indicates that only mild form of infection 98 benefit from remdesivir but that there is no difference in severe form of COVID-19 with placebo. It is 99 noteworthy to notice that results are given in intention to treat but that one third of enrolled patient 100 in both arms only (33.8 / 35.7%) received the complete protocol, 180/531 and 185/518 for 101 remdesivir and placebo respectively. Of them 288/ 1049 (27.4%) were discharged because they were 102 cured before the end of treatment and were loss of follow up, the remaining still receiving the 103 treatment or having missing treatment data at time to analyses. While an analysis according to the 104 ITT principle aims to preserve the original randomization and to avoid potential bias due to exclusion 105

of patients, such a number of loss of follow up is unacceptable because it might modified the 106 benefits of randomization, those loss to follow-up often having a different prognosis than those who 107 complete the study (14). In this study 168 patient were discharged before the end of treatment in 108 the remdesivir arms versus 120 in the placebo, which is significantly different (p<0001). It is likely 109 that those patients had a baseline score of 4 or 5 as they discharge before the end of treatment 110 explaining in part the better outcome in the remdesivir arms. Some have suggested that <5% loss 111 leads to little bias, while >20% poses serious threats to validity (15) . Nevertheless, a per-protocol (PP) 112 analysis as recommend in the CONSORT guidelines should be reported for all planned outcomes to 113 allow readers to interpret the effect of an intervention (16). 114

The last released paper compares 5 days to 10 days treatment for remdesivir with no significant 115 mortality nor improvement of clinical status between the two arms. Altogether, any serious adverse 116 event is reported in 27.7% of treated patient among them 4.7% of acute kidney injury. In 7.3% of 117 patient adverse events lead to stop the treatment (17). 118

Still few studies have been reported on evaluation the new drug remdesivir. In many aspects, data 119 from a case report or series without controls mean little to nothing in the context of evaluating 120 efficacy of an experimental drug. On the other hand, RCTs takes time and rarely bring usable 121 information during time of outbreak. Three RCTs have data available, but two share the same aims 122 and give contradictory data. Only one is methodologically adequate with both IPP and PP analysis on 123 a cohort of patient having completed the study demonstrating the absence of difference between 124 drugs and standard of care. 125

As today no study convincingly supports the use of remdesivir in severe patients. It is likely that, such 126 as for influenza, the major key for COVID-19 outcome is the early treatment of patient at the time of 127 diagnosis. However serious adverse reactions, some leading to interruption of treatment, and the IV 128 route, would probably limit the use of remdesivir in this indication. 

Remdesivir in adults with severe COVID-19: 146 a randomised, double-blind, placebo-controlled, multicentre trial

Compassionate Use of 149 Remdesivir for Patients with Severe Covid-19

Covid-19: Remdesivir is helpful but not a wonder drug, say researchers

Novel Coronavirus in the United States

First 12 patients with coronavirus disease 2019 ( COVID-19) 157 in the United Staes

Scope and impact of financial conflicts of interest in biomedical 159 research: a systematic review

In vitro testing of 161 combined hydroxychloroquine and azithromycin on SARS-CoV-2 shows synergistic effect

Tocilizumab, an anti-164 IL6 receptor antibody, to treat Covid-19-related respiratory failure: a case report

Remdesivir for the 167 Treatment of Covid-19 -Preliminary Report

Loss to follow-up. Evid Based Spine Care

Evidence based medicine : how to practice 170 and teach EBM

statement: updated guidelines for reporting 172 parallel group randomised trials

Remdesivir for 5 or 10 174 Days in Patients with Severe Covid-19