Remdesivir : does the European Medicines Agency endanger our health ?

ANALYSIS: The approval of remdesivir or Veklury® (trade name) by the European Medicines Agency (EMA) is questionable on many levels. The study, which is pivotal to this approval, includes enough bias that a complete new analysis is required. The National Medicines Safety Agency (ANSM) cannot in any way be satisfied with the recommendations of the EMA without requesting a new complete analysis with the data, as well as the elements of toxicity of this drug.

The rest of this article is fairly technical in order to highlight the approximations that lead to questionable conclusions.

Situation

Remdesivir is the first drug approved by the EMA under the name Veklury® for the treatment of SARS-CoV-2. Here is the CHMP text dated June 25, 2020 giving a favourable opinion.

On June 25, 2020, the Committee for Medicinal Products for Human Use (CHMP) adopted a favourable opinion, recommending the granting of a conditional marketing authorization for the medicine Veklury®, intended for the treatment of SARS-CoV-2 , a new coronavirus responsible for a respiratory disease known as coronavirus disease 2019, or COVID-19.  The sponsor of this medication is Gilead Sciences Ireland UC.

Detailed recommendations for the use of this product are described in the product information (PI), which is published in English . The European public assessment report (EPAR) will be published after the European Commission has granted marketing authorization and will be available in all official languages of the European Union.

In its report on the drug, the CHMP refers to the study published in the New England Journal of Medicine (NEJM) on May 22, 2020 which concludes that remdesivir aka Veklury® would reduce hospital stay by 15 days in the placebo group at 11 days in the remdesivir group or for severe cases respectively from 18 days to 12 days with an improvement ratio (recovery rate ratio) of 1.37.

Table 3 - NEJM Remdesivir study

The appendix to the study report provides more detailed information about the clinical trial as well as some of the side effects.

ANALYSIS

The analysis of the NEJM study on which the EMA decision is based is important, as it reveals many inconsistencies.

* This study is carried out in multiple sites (60) and several countries (10) making delicate an equivalent follow-up of patients and information even if it gives the impression of being much more serious than the study criticized by The Lancet. Hospitals are mentioned in the study, but form such a long list that it would prove tedious. On the contrary here the proliferation of detail of information and names of the doctors involved suggests that the patients exist. We recall the comment by Martin Landray, co-investigator of the British study Recovery, on the cost of Gilead's studies 20 to 50 times higher than their study.

* The study protocol first interested us to see that it had been modified during the study and that the main objective had been changed after the start of the inclusions and had become the reduction of the time of hospitalization. * In terms of statistical modelling, the study results have not been corrected for the various possible sampling biases. We could have pointed to many other irregularities, but more important elements caught our attention. * Side effects are only taken into account after the 5th occurrence. In addition, it is specified that the patients did not receive the same treatments. * By reading the appendix and the main text of the study, we can recompose the following table:

 NEJM study sample on remdesivir (Veklury®)

The delta line of the table above shows information given in the table in figure 1 on unspecified adverse events or withdrawals of consent. These elements should have been taken into account in the analysis since this represents 10% of the sample of each arm.

In the detail of the totals, there are 49 patients in the remdesivir arm and 53 patients in the placebo arm which are not found in the analysis.

* The report shows a significant difference between the remdesivir group and the placebo group. The proportion of cured patients in the total sample is 62% in the remdesivir group and 52% in the placebo group, and this result is statistically significant. However when we recalculate the observed ratio of 1.32, it is very different from the calculated ratio, probably here for rounding errors. In addition, in patients with mild forms of the disease, there are no significant effects with more than 80% of patients recovering; the impact on healing is greatest in severely ill patients. However, this analysis will have to be adjusted in the light of the patients for whom we have no information. With nearly 60% of cured patients in the remesivir arm, 40% of patients are not cured, including those who have died. It is therefore fundamental to understand what happened to the 32.3% of patients, especially in terms of side effects. No information on this group is available. Likewise, the report describes this improvement in healing ratio which must take up the above elements before any conclusions can be drawn.

Analysis of cured NEJM study on remdesivir (Veklury®)

* Looking at the ratio of patients receiving less than 10 doses and cured, we see that this represents 31% of the remdesivir arm and 23% of the placebo arm. Conversely, for those who received the complete treatment of 10 doses, this represents 33% of the remdesivir arm and 36% of the placebo arm. Since some patients in the placebo arm did not receive the full treatment, it seems essential to correct this bias since it would be difficult to conclude that it was the treatment that had an effect since it did not not been followed to the end. * In addition, the trial specifies that patients treated for other illnesses continued to take their treatment. Some patients will therefore not have received the same treatment, which adds a significant bias in the cause and effect relationship of the drug and cross-effects with other molecules. This analytical bias is as important or more important than the analysis of side effects which are only measured when there are 5 occurrences.

NEJM study dosages on remdesivir (Veklury®)

* However, the most surprising analysis concerns the mortality rate in severely affected cases. By redoing the calculations, there is already a significant difference between the calculated mortality rate (5.95% remdesivir arm and 10.36% placebo arm) versus what is reported in Table 7.1% and 11.9% respectively. A difference of 16% and 13% of measurement. Rounding offs maybe? The fact remains that if this difference is due to a rounding error, we are entitled to ask questions about the overall quality of the analysis.  We tried to recalculate this figure without success from figure 1 of the study (Inclusion and randomization of patients). This is the basis of any study to be able to recalculate the exact figures at any time for reasons of transparency.

Of course, one point to note is that this is the 14th day mortality, which is probably put forward only to mask the results at 28 days. The study therefore tries to highlight mortality at 14 days, the time of treatment without taking into account the side effects in order to influence the reader. This difference in measurement is therefore not reported, because it lacks general information on mortality at term.

Mortality NEJM study on remdesivir (Veklury®)

But why not talk about mortality at 28 days (secondary criterion)?

Several hypotheses come to mind, but the main one is that by putting forward the 14-day mortality rate it will not be noted that it is not the overall mortality (subterfuge). We will have to analyse the causes of mortality and look in detail at the side effects. It is therefore also essential to look at the cases that have taken less than 10 doses to understand the causes of the discontinuation of treatment, and the duration of treatment.

However, it remains to calculate the term mortality which would take into account all the side effects. And it just isn't done and probably never will be.

We tried to reach the EMA to ask questions about it.

Such an important decision must be fully analysed. We contacted EMA President Christa Wirthumer-Hoche to answer these questions and she replied that she was not the right person for this, without referring us to the "right people". We also tried to reach the CHMP members without success.

Enlightening comments from one of our readers, who raises many points. Here is a summary (the whole is attached in the Appendix)

The conclusion of the support study is unambiguous:

* The difference in mortality is not statistically significant "and, understands,

"There is no statistically significant difference on a" deterioration "of the placebo group vs remdesivir"

* Gilead, announces "an improvement for patients with moderate symptoms", problem, in the study there are only 120 patients "moderate" (less than patients where data is missing) and the curves of the study are only on the entire sample and the most severe cases. On “moderate” cases: laugh, see by difference a harmful result of remdesivir. * Recall that the initial study focused on (400 "severe" patients with administration of remdesivir for 5 days or 10 days without placebo) and on (600 "moderate" patients with administration of remdesivir for 5 days and 10 days and placebo). * Finally, coincidence, this study appeared on May 22, as the study removed from The Lancet by Dr. Mehra. * And an additional coincidence, the commenter in the June 1 official Gilead study announcement, works at Brigham & Women Hospital and Harvard as Dr. Mehra.

Conclusion

Would the undesirable effects be the reason for the modification of the main objective of the protocol being tested. Recall that the main objective was amended on April 2, 2020 from "recovery of health points (return to a general condition score of 1 to 3 on a scale of 8)" to "duration of hospitalization". The official reason given for this amendment is that the Covid19 lasts longer than originally planned. In general such a modification is not acceptable. In addition, among the secondary criteria, is included the 28-day mortality which is nowhere reported in the study.

Could this be one of the reasons why unwanted events are only taken into account after the 5th occurrence?

There are therefore many important biases not taken into account by the CHMP and it is for this reason that we consider that the CHMP has not done a complete audit work, and that the emergency situation invoked by the pandemic does not justify the negligence which could endanger the health of patients and this all the more so as the preliminary French pharmacovigilance data alerted on renal insufficiencies with in certain cases, necessity of dialysis in patients on remdesivir.

With this analysis, we cannot therefore conclude with certainty that remdesivir has a positive effect on the number of days spent in hospital without looking in detail at the subsets and especially the causes of cures of patients who stopped treatment before the 10 doses.

This recommendation from the EMA is therefore biased to such an extent that it would be desirable for the French National Authority to at least request additional information on the analyses before concluding. The health of the French cannot be put into the hands of this recommendation without a thorough scientific and medical verification. Primum Non Nocere " Above All, Do No Harm!".

Two other questions arise about the patient monitoring system to assess the toxicity of this molecule, and how far will the lobbying of the company Gilead go?

APPENDIX: Analysis of the chronology of Gilead's press releases and of the study

* February 26, 2020 Gilead press release: From the start, the objective is to evaluate remdesivir as a 5-day vs 10-day treatment with placebo on “moderate” patients (60% of the panel), without placebo on patients "Severe" (40% of the panel).

* April 10, 2020 Gilead press release: So April 10 Gilead says: "the results for severe patients for the end of April and for moderate patients for May" knowing that on severe patients will not come from China. “Gilead is conducting two phase 3 clinical trials of remdesivir in countries with a high prevalence of COVID-19. Data from the SIMPLE study in patients with severe illness are expected this month, followed by data from the SIMPLE study in patients with moderate illness in May. In addition, Gilead supports several clinical trials conducted by other organizations, including two studies conducted in Hubei Province, China. Gilead was informed that the study in China in patients with severe illness was terminated prematurely due to low enrolment. The study in China on patients with mild to moderate illness is underway. A global study on remdesivir led by NIAID continues to recruit patients and data from this study is expected in May. "

* April 29, 2020 Gilead press release: On April 29 (see announcement of April 9) we have an announcement about the study on “severe” boxes on the effect of the treatment duration of 5 or 10 days where we can understand no placebo simply comparison between 2 dosages of the treatment !!!

We can read 2 versions: "the 5 day treatment is better than the 10 day treatment" where "the 5 day treatment is less worse than the 10 day treatment". It should be noted that in "severe" cases, a priori 400 patients out of 1000 patients from the start, there is no placebo, we just studied the treatments over 5 days versus 10 days. And on the 600 others, there are 3 branches: remdesivir 5 days or 10 days, or nothing.

“Gilead announces the results of the Phase 3 trial of the experimental antiviral remdesivir in patients with severe COVID-19. The study shows similar efficacy with dosing times of 5 and 10 days of Remdesivir –

Gilead Sciences announced today the preliminary results of the SIMPLE open phase 3 trial evaluating the 5 and 10 day dosing times of the antiviral remdesivir investigation in hospital patients with severe manifestations of COVID-19 . The study demonstrated that patients receiving 10 days of remdesivir treatment achieved a similar improvement in clinical status compared to those receiving 5 days of treatment (0.75 [95% CI 0.51 - 1.12] on the 14th day). No new safety signals were identified with remdesivir in the two treatment groups. Gilead plans to submit the complete data for publication in a peer-reviewed journal in the coming weeks. "

* May 22 - June 1: the results are published in the NEJM, and in a press release. However, according to the April 9, 2020 announcement, the results were due to be released in late May only. From Gilead's perspective, the results are favourable for "moderate" patients only.

"Gilead announces results of phase 3 trial of remdesivir in patients with moderate COVID-19 - Study shows 5-day treatment of remdesivir resulted in significantly greater clinical improvement compared to treatment with only standard of care - announced today the results of the Phase 3 SIMPLE trial in hospital patients with moderate COVID-19 pneumonia. This open study evaluated the 5 and 10 day courses of experimental antiviral remdesivir plus standard of care, compared to standard of care alone. The study demonstrated that patients in the 5-day remdesivir treatment group were 65% more likely to have clinical improvement on day 11 compared to those in the standard care group (OR 1.65 [95% CI 1 , 09-2.48]; p = 0.017). The odds of improvement in clinical status with 10-day treatment of remdesivir compared to standard of care were also favourable, tending towards, but not reaching statistical significance (OR 1.31 [95% CI 0 , 88-1.95]; p = 0.18). No new safety signals were identified with remdesivir in the two treatment groups. Gilead plans to submit the complete data for publication in a peer-reviewed journal in the coming weeks. "

The study was released the same day as that of Dr. Mehra's The Lancet. We discover that Dr. Fransisco Marty, who is involved in the official announcement of Gilead, works at Brigham & Women Hospital and at Harvard as Professor Mehra. They leave their 2 studies the same day. Who will believe in a coincidence

"Our understanding of the spectrum of severity of SARS-CoV-2 infection and presentations of COVID-19 continues to evolve," said Francisco Marty, MD, infectious disease physician at Brigham and Women's Hospital and associate professor of medicine at Harvard Medical School 

GILEAD specifies that its conclusion only concerns “moderate” patients on day 11.

However when we look at the study, we don't see the same thing:

* The majority of patients (943) are in the "severe disease" category while GILEAD announces that REMDESIVIR 5 days a little "moderate" patients (120 patients) on whom the study says nothing. * On the contrary, the difference would be more in cases 4 and 5, which could imply that in groups 1,2,3 (moderate patients), there is no difference. The patients were divided into 8 groups of 1 to 8 with moderate patients from 1 to 3.

The key point: the study clearly says that remdesivir has no effect on deaths on day 11 and nothing after day 11 and superb understatement, the placebo group on day 11 is not subject to a statistically degraded situation compared to to the two remdesivir groups, or with 500 patients in each group, if there is really a slight difference, it would appear with a significant probability. In addition we do not even talk about the branch of treatment over 10 days.