Whipsawed by Remdesivir

In the last few weeks we have been given bizarrely conflicting information about the use of remdesivir to treat COVID-19 infection. This antiviral drug appeared to have the most promise to treat COVID-19, as it demonstrated a well-defined mechanism of action specific for the virus. It also had shown some efficacy in other, similar viral diseases in humans as well as against COVID-19 in tissue culture.

So far, the reports on the clinical use of this drug have been sketchy. Initially, physicians taking care of patients optimistically reported improvements in patients that were given remdesivir. There was then a case series in the New England Journal of Medicine, reported in the news with much fanfare, that showed potential benefit in a few treated patients. I wrote about this in an earlier blog piece that also discussed the drug’s action.

Neither of these studies were either blinded or controlled, meaning there was way to tell if the drug was really making a difference. What was needed to show true value for remdesivir was to compare treated patients with patients that had identical care without the drug. Without this and without the doctors evaluating the patients not knowing who was receiving the drug (blinding) it was impossible to determine that the drug actually benefited the patients.

Yesterday, the World Health Organization mistakenly put an abstract (one paragraph summary) from a paper from China that described a controlled, blinded study that indicated remdesivir had no effect on COVID infection. This was cleverly found my friends at STAT NEWS and reported to the world!

As originally designed, the China study hoped to enroll 453 patients. Once enrolled, the patients were randomized in a 2:1 ratio to either active drug to placebo.  The group was then treated with daily infusions of either remdesivir or a placebo for 10 days.

The primary goal of this study was to show that the patients on the drug had less or improving symptoms within 28 days of starting treatment. The patients were evaluated with a six-point scoring system ranging from hospital discharge (a score of 1) to death (a score of 6). In order to say a patient responded to remdesivir, he/she must had improved by at least two points. Patients could remain hospitalized for the entire the 28-day period of the clinical trial but still improve enough— no longer needing intubation or supplemental oxygen, for example — to count as a responder.

According to the abstract on the WHO website (no one saw the full paper and the WHO took down the abstract when they realized it was openly available), 158 patients received remdesivir and 79 patients were in the control arm. The study was stopped prematurely because of poor enrollment due to decreasing numbers of COVID infections in China. There was no statistically significant improvment in any clinical score that was associated with remdesivir. Unfortunately, since only the abstract was available no firm conclusion on the study can be made other than the fact that it failed to show a difference from the placebo.

Taken together, this group of studies really does not give any insight into whether or not remdesivir will be useful in treating COVID-19. The uncontrolled reports are meaningless, as you cannot tell the effect of a drug without a control group. Blinding is also essential, as physicians are a hopeful group and certainly want to believe that a drug will provide value. That is why clinicians are usually blinded to which patient is receiving drug vs. placebo in a clinical trial.

In some ways the study in China is disturbing, as it is a controlled, blinded clinical trial that would seem to have more weight. It was stopped early because of lack of enrollment and that may be part of the reason it failed to show significance. However, the number of patients is large enough that if there was a significant clinical effect it should have been observed.

There are caveats on this study though. The biggest concern is that the antiviral drug could be started as late as 12 days after patients became symptomatic. This might be a reason that the drug did not work since most antiviral drugs (including those effective for treating flu) need to be started within the first few days of infection to be effective.

There are larger, well controlled blinded clinical trials that are now being run by NIH to assess the clinical benefit of remdesivir. Several of these trials give the drug early in the course of the disease. Any true assessment of the value of this drug in treating COVID-19 awaits the result of those trials.

Sadly, there is no way to speed up the drug evaluation process. Only by independently and rigorously assessing a drug in actual human disease can one decide whether a drug is valuable. Given the emotions associated with the current pandemic, it is easy to understand why people jump at any possible benefit from a treatment. But in any case, whether its hydroxychloroquine or something better founded in science like remdesivir, we require rigorous controlled clinical trials to evaluate every drug. If nothing else, this saga shows that the FDA must keep its bar high in evaluating drugs.