Contribute Try STAT+ Today

Remdesivir, the only drug cleared to treat Covid-19, sped the recovery time of patients with the disease, but its benefit appeared much more limited in patients who needed mechanical ventilation as part of their treatment, according to eagerly awaited results of a clinical trial. 

Initial results from the study, which led to the drug’s emergency authorization by the Food and Drug Administration, were released late last month. Full data were published late Friday in the New England Journal of Medicine. 

“It’s a very safe and effective drug,” said Eric Topol, founder and director of the Scripps Research Translational Institute. “We now have a definite first efficacious drug for Covid-19, which is a major step forward and will be built upon with other drugs, [and drug] combinations.”


But Topol noted that there was no sign of a benefit in patients who began the study with the most severe baseline status — those who were on non-invasive ventilation, who were intubated on a ventilator, or who were receiving extracorporeal membrane oxygenation, a treatment in which the oxygen is added to the blood outside the body. “We need to get something that works for these patients who have a high mortality rate,” Topol said.

The study of 1,063 patients included 538 who received remdesivir and 521 who were given a placebo. Those who received remdesivir recovered in a median of 11 days, compared to 15 days for those who received placebo. Mortality in the remdesivir group was 7.1%, compared to 11.9% for the placebo group, but this difference was not statistically significant. This is slightly better than previous results.


Remdesivir, developed by Gilead, has been in short supply, so the issue of which patients benefit most could be very important for deciding who receives the drug. The authors of the study argue that the difference in mortality rates resulted from the fact that there were fewer patients whose symptoms were more severe at the beginning of the study, resulting in less statistical power. But the benefit appeared much bigger among those who were less sick.

The severity of patients’ illness was rated on a scale of 1 (not hospitalized) to 8 (dead). The lowest score in the study was a 4, denoting hospitalization, but no need for extra oxygen. The largest group of patients scored a 5, meaning they did need oxygen.

Among patients who scored a 4, there was a 38% benefit in the speed of recovery. Among those who scored 5, there was a 47% benefit. But that benefit fell to 20% among those who scored a 6, meaning they were receiving high-flow ventilation, and a decrease of 0.05% among those who scored a 7, meaning they were intubated or on extracorporeal membrane oxygenation. Until more data are available, doctors and researchers are likely to debate whether to use remdesivir in sicker patients.

There have been concerns that the study’s main goal, originally based on changes in the 8-point score at day 15, was changed before the result was analyzed. Changing goalposts can be a sign that researchers are skewing the results. In the paper, the researchers explain that the concern was that Covid-19 illness was persisting much longer than had been expected. But that original goal, the odds of improvement in the 8-point scale on day 15, was 50% higher in the remdesivir group, a highly statistically significant result.

Some critics saw flaws in the study that could affect how its results are interpreted. The study was stopped after a data safety and monitoring board, a panel of outside experts charged with safeguarding patients in the trial, notified the National Institute of Allergy and Infectious Diseases, which was running the study, that there had been a statistically significant benefit of remdesivir.

The New England Journal paper does not say, as is normally the case, that the study was stopped. As STAT reported two weeks ago, some critics believe that the study should have continued in order to better discern whether remdesivir reduces the death rate.

“It’s clear from the publication that the study was stopped prior to the completion of a significant number of patients,” said Steven Nissen, a cardiologist at the Cleveland Clinic. “By stopping it early the scientific community was deprived of the opportunity to determine whether remdesivir can or cannot reduce mortality.”

According to the paper, as of April 28, 2020, a total of 391 patients in the remdesivir group and 340 in the placebo group had finished 29 days of the study, recovered, or died. But there were still 132 patients in the remdesivir group and 169 in the placebo group who had not recovered or completed their follow-up visits. Nissen’s contention is that if the study had been allowed to finish, it would be more clear whether or not the drug improves patients’ odds of survival.

Topol said that he interprets the survival benefit as “quite likely” and said he is “not bothered” by the fact that the mortality result is not significant. He said he awaits results from future trials.

Peter Bach, a pulmonologist who is the director of the Center for Health Policy and Outcomes at Memorial Sloan Kettering Cancer Center, worried that it appears that slightly more placebo subjects started out in the more severe groups. This, he said, could mean that the overall effect is “potentially exaggerated.”

“I am not concluding from this that the treatment is not reducing time to recovery; it makes me less convinced that there is a mortality benefit when more placebo than intervention ended up in the highest risk group,” Bach said. He pointed to another study of remdesivir, conducted in Wuhan, China, in which the time to benefit was not significant and there was no mortality benefit.

Bach was critical of another study run by Gilead that lacked a placebo group, instead comparing two different durations of remdesivir treatment. He called the decision “frustrating.” Gilead has said that the decision was driven by limitations on its ability to produce vials of placebo that would have resembled the medicine.

In a statement Friday, the company said results from a study that compares remdesivir to the standard of care in patients with moderate illness will be available by the end of the month.

  • Here is the original publication
    My personal summary:
    Significant reduction in the duration of the illness across all patients (Remdesivir 11 days versus placebo 15 days)!
    Overall, a high proportion of critically ill patients in the entire patient group (197 patients with non-invasive ventilation or high oxygen demand / 272 patients with invasive ventilation or ECMO). So 469 patients out of a total of 1036 patients. See Table1 below.
    Death rate remdesivir group 7.1% versus placebo group 11.9% not significant. But i think good under the above conditions. Other studies have shown that over 50% of ventilated patients die.
    Best effectiveness in the group of moderately ill patients (group C with oxygen, but without ventilation, look to Figure 2.)
    My conclusion:
    So let us give Remdesivir earlier before the Patient needs ventilation and the intensive care unit. Seen in this way, it is a drug for a lot more patients. Maybe Remdesivir will be a big seller for Gilead.

  • Topol said that he interprets the survival benefit as “quite likely” and said he is “not bothered” by the fact that the mortality result is not significant.

    Good to know he is director of the Scripps Research Translational Institute.
    Another one reminding me that to make a career in science you must not reason and act like a scientist. Or, at least, like a good one.
    And, even if an educated guess should trump statistics, the favorable outcome would still be very modest at best.

    • What? There was a 40% reduction in mortality. Not a big enough sample to be “statistically significant,” but it’s not a tiny sample, and it’s the best data we have. With more data, we may find that the benefit is much less or much more, but As Topol says, it’s more likely that it will turn out more or less the same. If in two “not statistically signficant” trials of this caliber, one medicine showed a 40% increase in deaths and one a 40% decrease, would you really say, “It’s the same result (no significant difference with placebo) for both, so just give me either one?”

  • This is great news for the global population. My question is: Is these medicines are available in world wide countries? If Not, how developing countries can get it.

  • All of these studies in which treatment is not initiated until patients are hospitalized mean very little.

    I’m sure that mean treatment is initiated 10-15 days into the infection.

    An antiviral is not likely to demonstrate a significant benefit over placebo unless given early in the infection to prevent viral replication and therefore prevent progression of the disease to acute phases.

    All these patients being treated 10-15 days into their infection already have significant lesions and complications, and the factors leading to a full recovery are more dependent upon one’s inherent physiological resilience. Late in the infection the body already has significant antibodies that are also limiting the progress of the disease.

    If Remdesivir was going to be effective, it would be best used by it’s administration immediately upon learning of infection for patients who are in high risk categories to see if it helps prevent them from progressing to a serious state of disease, long before the adaptive immune response has begun.

    These studies aren’t testing what needs to be tested, which is does the early administration inhibit viral replication prior to the body’s adaptive immune response is significantly mobilized.

    We don’t give Tamiflu to flu patients in the hospital, we give it to patients to avoid the hospital.

    • I couldn’t agree more. An oral formulation of remdesivir is needed (if possible) so that patients can be treated at home early in the course of the infection.

    • Seems this anti-viral is best administered early on in the disease process prior to hospitalization. The issue is that this particular agent is only available by intravenous administration, severely limiting its availability outside a hospital setting 😢

  • wow big ups for this strong study otherwise this pandemic is becoming harder

  • A 5% rate of people recovering a few days earlier is not worth the 23% rate of permanent and fatal side effects caused by this drug. To simplify this I would rather take a chance of being sick extra 4 days then have a one-in-four chance of permanent disability or death. The Cure is a lot more deadly than covid-19

    • This is flat wrong. The study showed that the remdesivir patients suffered fewer serious side effects than the patients on the placebo.

  • Also, have Topol and Nissen ever agreed on anything? It’s fun – and informative – to watch their debates, however. And that goes back to Vioxx and before…

  • “Changing Goalposts”? Really? Are we children?? Rick, have someone edit this stuff. Anyone reading STAT knows what an endpoint is, and MOST of us already know that revised endpoints means a failed study.

Comments are closed.