Skip to Main Content

A cheap, readily available steroid drug reduced deaths by a third in patients hospitalized with Covid-19 in a large study, the first time a therapy has been shown to possibly improve the odds of survival with the condition in the sickest patients.

Full data from the study have not been published or subjected to scientific scrutiny. But outside experts on Tuesday immediately embraced the top-line results. The drug, dexamethasone, is widely available and is used to treat conditions including rheumatoid arthritis, asthma, and some cancers.

In a statement, Patrick Vallance, the U.K. government’s chief scientific adviser, called the result “tremendous news” and “a ground-breaking development in our fight against the disease.” Scott Gottlieb, a former commissioner of the U.S. Food and Drug Administration, called it “a very positive finding” in an interview on CNBC. “I think it needs to be validated, but it certainly suggests that this could be beneficial in this setting.”


Atul Gawande, the surgeon, writer and public health researcher, urged caution, tweeting, “after all the retractions and walk backs, it is unacceptable to tout study results by press release without releasing the paper.”

The study randomly assigned 2,104 patients to receive six milligrams of dexamethasone once a day, by mouth or intravenous injection. These were compared to 4,321 patients assigned to receive usual care alone. 


In patients who needed to be on a ventilator, dexamethasone reduced the death rate by 35%, meaning that doctors would prevent one death by treating eight ventilated patients. In those who needed oxygen but were not ventilated, the death rate was reduced 20%, meaning doctors would need to treat 25 patients to save one life. Both results were statistically significant.

There was no benefit in patients who didn’t require any oxygen. The researchers running the study, called RECOVERY, decided to stop enrolling patients on dexamethasone on June 8 because they believed they had enough data to get a clear result.

“Dexamethasone is the first drug to be shown to improve survival in COVID-19,” Peter Horby, one of the lead investigators of the study and a professor in the Nuffield Department of Medicine at the University of Oxford, said in a statement. He added that the drug should now become the standard treatment for patients with Covid-19 who need oxygen. “Dexamethasone is inexpensive, on the shelf, and can be used immediately to save lives worldwide.’”

A different arm of the same study showed on June 5 that hydroxychloroquine, widely touted as a potential Covid treatment, had no benefit in hospitalized patients. Yesterday, based in part on those results, the Food and Drug Administration revoked an Emergency Use Authorization for using hydroxychloroquine in those patients.

From the start of the pandemic in March, researchers have focused on two different stages of Covid-19, which will likely require very different interventions. Some drugs are designed to directly combat the novel coronavirus, SARS-CoV-2, that causes the disease. The first medicine shown to have a benefit, remdesivir from the biotech firm Gilead Sciences, falls into this category, even though, because it must be given intravenously, it has been tested in hospitalized patients. Remdesivir shortens the course of infection, but has not been shown to save lives.

After patients have become profoundly sick, the problem starts to become not only the virus but their own immune system, which attacks the lungs, a condition called acute respiratory distress syndrome, or ARDS. For these patients, doctors have believed, they would need to dampen patients’ immune response even as they fought the virus.

Initially, excitement in this area fell on new and expensive drugs, such as Actemra, a rheumatoid arthritis drug from Roche that is used to treat a similar condition caused by some cancer immunotherapies. But a study in patients who needed oxygen showed no benefit from a similar drug, although another arm in sicker patients is continuing. The National Institutes of Health is conducting a study of an Eli Lilly pill targeting rheumatoid arthritis, an extension of the study that showed remdesivir has a benefit.

Dexamethasone, which reached the market 59 years ago, seemed an unlikely candidate to help these patients; it was seen as too crude a way of tamping down the immune system. In guidelines for physicians treating the disease, the NIH doesn’t even mention the therapy.

Studies that are testing other medicines may now need to incorporate the use of the drug, which could complicate analyzing the results. A spokesperson for Regeneron, which is testing Covid-19 drugs focused on both attacking the virus and dampening the immune system, said the company’s studies are written so that when a new medicine becomes the standard of care, it becomes available to patients in the trial.

Some studies have shown a benefit for using dexamethasone in acute respiratory distress syndrome not related to Covid-19, although the benefit was smaller than in RECOVERY. 

The result, should it hold up to further scrutiny, shows the benefit of the strategy of Horby and Martin Landray, the Oxford researchers who designed the study, leveraging the U.K. health system to start a study of multiple inexpensive potential Covid-19 therapies — including hydroxychloroquine, dexamethasone, and also some older HIV medicines. Several months into the Covid-19 pandemic, two of the most important results come from this single study.

Neither of those results, however, have been scrutinized or published.


  • I gave you 2 examples of a reduction of 50%. One had a NNT of 100, the other 4. It’s not the 35% that matters, it is the two different numbers (NOT included in the summary) that constitute that 35% reduction. As stated earlier, whoever wrote that summary did not include (or the authors did not provide to them) the two numbers, the second of which was 35% less than the first, that would give you the NNT of 8. Need the entire article for that.

    Stated differently, the 35% is the relative risk reduction (rrr) and we need the absolute risk reduction (arr) to calculate the NNT. I presume the authors did the math correctly. Again, they don’t provide what we need for the arr, we just have the rrr. The formula for NNT is 1/arr.

  • The fda is trying to undermine the trump adminstration. Plain and simple

  • The number needed to treat (NNT) is a way of normalizing data to see if it’s “worth the difference” In order to calculate it, you need to know what % of people died without the drug and what % of people died when given it. The formula for NNT to save a life is 1/(difference between mortality in decimal form). They don’t give you enough info in the blurb to see what numbers they had to calculate it. I suppose you could figure it out with some math. I can tell you that 8 is a low NNT, especially for a cheap, readily available med.

    For example: If I had a wonder drug that cut death in half for a disease, but only 2% of people died from it, then my new death rate is 1%. Headline: wonder drug cuts death in half! But doing the math, 1/(0.02-0.01) = 100. So I have to give 100 people the new drug to save 1 life.

    If I had a wonder drug that took a disease with a 50% mortality down to 25%, it could have the same headline. But for this, the NNT would be much better: 1/(0.5-0.25) = 4! Would only need to treat 4 people to save a life. So a NNT of 8 is pretty darn good.

  • It’s nice to know that dexamethasone “reached the market” 59 years ago, because the Highly biased author of the story wants readers to believe that drugs just fell out of the sky back then.

    It was originally synthesized and developed by Merck…

  • Since strong antibody response was associated with disease severity.
    dexamethasone may be working to moderate this immune response.
    Viral Kinetics and Antibody Responses in Patients with COVID-19

    “we observed three types of antibody responses in COVID-19 patients, strong, weak and non-response. Second, we found that the earlier response, higher antibody titer and higher proportion of strong responders for IgM and IgG were significantly associated with disease severity. Third, the weak responders for IgG antibodies had a significantly higher viral clearance rate than that of strong responders. These data indicate strong antibody response is associated with disease severity, and weak antibody response is associated with viral clearance, which resembles SARS and MERS.”

  • Results that demonstrate a 35% reduction in the risk of death does not mean you have to treat 8 patients to have 1 death. Bizarre to present it that way 100 patients at risk of dying 35 patients survived ie reduced the death rate by 35%

    • The investigators report 41% mortality in the standard of care arm and 35% less mortality in the dexamethasone arm. So if 100 patients are treated about 27 die rather than 41, saving 14 lives per 100 treated, which is about 1 life saved for every 8 treated.

  • Where was this published?
    Has it been peer reviewed or even accepted for publication?
    Without peer review and analysis by statistical and medical experts this is nothing ….Yet.

    • Apparently, the odds are pretty good that in the largest group, those that didn’t require oxygen or ventilation, the “wonder drug” results in more deaths. (1.22 point estimate).

  • I was puzzled by the math between saving one life after treating 8 ventilated patients and a 35% reduction in the death rate. Does the math mean that usual number of deaths for 8 ventilated patients is 3 deaths, but with the Dexamethasone treatment only 2 patients die (a 1/3 reduction in deaths)?

  • I’m not sure why this is a breakthrough, I was under the impression that steroids were a standard treatment in multi-system failure.

Comments are closed.