16 April

Comment on
Compassionate Use of Remdesivir for Patients with Severe Covid-19
N Engl J Med. 2020 Apr 10.

By Christian Hoffmann


Today, we will not discuss the Top Ten Papers. Instead, we will talk about remdesivir and give you 10 good reasons to be careful. Last Friday, on April 10, the New England Journal of Medicine published data on patients who were treated with 10 days of remdesivir on a compassionate use basis (Grein 2020). These results gained a lot of media attraction and Daniel O’Day, Gilead’s CEO, wrote the same day that “the majority” of patients “demonstrated clinical improvement”. There is no doubt that remdesivir is currently the biggest hope for COVID-19. Results of two large Phase III randomized clinical trials are expected by the end of this month. Remdesivir is also among the four drugs tested in WHO’s huge SOLIDARITY trial.

Grein et al. offer an optimistic view on remdesivir. Although viral data were not available, they concluded with a clinical “improvement in 68%” (36/53) and a “noteworthy” low mortality of 13%, seemingly lower than seen in a randomized clinical trial (RCT) on lopinavir/r (Cao 2020). The authors also emphasize repeatedly the severity of disease in their patients, as many required ventilation – more than in the lopinavir/r trial.

Is the author’s optimism justified? We don’t believe so. We think that, given the published data, remdesivir has only low or at best moderate clinical activity. We are concerned that remdesivir which was not effective against Ebola, will fail in COVID-19 as well.

  1. First, 8/61 patients were not included in the analysis, among them 7 with no post-treatment data. Any reader will wonder what happened to these patients – did they die while on treatment? If these patients had been included in the analysis, we would have observed “improvement in 59%” (36/61).
  2. One patient was excluded due to an “erroneous start date”. This is remarkable as at least 13/51 received less than the planned treatment of 10 days but remained in the analysis. The same is true for at least two patients with, let’s say, an “erroneous stop date” (treatment given for 13 days). The author’s statement that treatment duration was not uniform, “largely because clinical improvement enabled discharge”, is misleading. Only 3/13 subjects who were treated for less than 10 days were discharged early.
  3. In the mentioned randomized clinical trial comparing lopinavir/r and standard of care, day 14 clinical improvement rates on a predefined ordinal scale were 46% with lopinavir/r and 30% with standard care (Cao 2020). According to an NEJM editorial commenting on this trial, results were “disappointing” (Baden 2020). Applying the same scale to the “study” presented here, the day 14 rate would have been at best 49% with remdesivir.
  4. The authors emphasize the severity of their patients’ health, indicated by the fact that 34/53 required invasive mechanical ventilation (IMV), more than in the lopinavir/r trial. But IMV is not the only indicator for severity. The patients in the lopinavir/r trial were sick, too. Inclusion criteria were CT-confirmed pneumonia and an oxygen saturation of 94 % or less while breathing ambient air.
  5. Age and comorbidities in the lopinavir/r trial looked much the same as in the remdesivir study.
  6. The median duration from onset of disease until treatment initiation was 12 days (IQR 9-15) with remdesivir and 13 days (IQR 11-16) in the lopinavir/r trial. Thus, considering points 4-6, it remains unclear if the remdesivir patients were more severely ill.
  7. If we focus on the 41 remdesivir patients who required at least high-flow oxygen at baseline, clinical improvement at day 14 was seen in only 39%. Is this more than we would have seen with placebo?
  8. This study provided data on ALT, AST, and creatinine. They were “fluctuating” during follow-up. Inexplicably, other easily performed and more important parameters such as lymphocytes, D-dimer and lactate dehydrogenase were unavailable. Several studies have identified these parameters to be highly predictive for clinical progression (Zhou 2020, Ji 2020).
  9. The data quality is poor. For some subjects (31, 32), duration remained (mistakenly?) unclear and subject 43 was discharged while worsening.
  10. The final argument: patients who were too ill for an antiviral agent (the “point of no return”) doesn’t count. Of 17 patients without IMV at baseline, at least 4 patients worsened during remdesivir treatment.

In conclusion, for a number of reasons, this report published in the New England Journal of Medicine is a cautionary tale for “science in a hurry”. Fragmentary data arouse false expectations. This paper is not helpful. Safety of the drug was shown in the Ebola trial. It might have been preferable to postpone the publication – after all, the eagerly awaited results from the randomized clinical trials on remdesivir are just days to a few weeks away.

Will remdesivir be to COVID-19 what AZT was to AIDS? (AZT, approved in 1987, was the first and only antiretroviral drug to treat HIV/AIDS for years. Unfortunately, dosage was complicated and efficacy was minimal.) We hope this is not the case because we badly need efficient drugs in our fight against the SARS-CoV-2 pandemic.

We sincerely hope that our objections to the study by Grein et al. are wrong.


Baden LR, Rubin EJ. Covid-19 – The Search for Effective Therapy. N Engl J Med. 2020 Mar 18. PubMed: Full-text:

Cao B, Wang Y, Wen D, et al. A Trial of Lopinavir-Ritonavir in Adults Hospitalized with Severe Covid-19. N Engl J Med. 2020 Mar 18. PubMed: Full-text:

Grein J, Ohmagari N, Shin D, et al. Compassionate Use of Remdesivir for Patients with Severe Covid-19. N Engl J Med. 2020 Apr 10. PubMed: Full-text:

Ji D, Zhang D, Xu J, et al. Prediction for Progression Risk in Patients with COVID-19 Pneumonia: the CALL Score. Clin Infect Dis. 2020 Apr 9. pii: 5818317. PubMed: Full-text:

Zhou F, Yu T, Du R, et al. Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study. Lancet. 2020 Mar 28;395(10229):1054-1062. PubMed: Full-text: