Skip to Main Content

Infusing hospitalized Covid-19 patients with blood plasma from people who recovered from the disease appeared to show a benefit in a nationwide study, but the study’s lack of a placebo group left several experts struggling to interpret the data.

The study, which enrolled more than 35,000 patients, found that quickly administering so-called convalescent plasma had a marked effect on mortality for patients with severe cases of Covid-19. Those who received transfusions within three days of diagnosis had a seven-day death rate of 8.7%, while patients who got plasma after four or more days had a mortality rate of 11.9%. The difference met the standard for statistical significance.


But without a placebo group for comparison, it’s unclear just how valuable the treatment might be. The study, run by the Mayo Clinic and sponsored by the National Institutes of Health, was meant to broaden access to convalescent plasma. It was part of what is known as an “expanded access program,” not designed to definitively test how well the treatment works but to get it to patients while collecting data.

In a statement, the Mayo Clinic said that the Food and Drug Administration has advised it that the expanded access program will continue “while planning is under way to transition smoothly to Emergency Use Authorization of convalescent plasma.”

Peter Bach, director of Memorial Sloan Kettering’s Center for Health Policy and Outcomes, said there’s no way to be sure about the ultimate benefit of convalescent plasma, but the study “checks a few boxes.” For one, it supports the overall theory that convalescent plasma might help some patients, he said. He noted that researchers also found that the quality of the plasma infusion had an effect on patient outcomes, as those who got infusions that were particularly rich with antibodies fared better overall. 


“If we had just done the randomized controlled trials, we would know the answers we are still guessing at,” Bach said, noting the authors of the study were “appropriately cautious.”

The results of the study were published Thursday on a preprint server, meaning they have not yet been peer-reviewed. 

The study enrolled a high proportion of critically ill patients, with about 52% in intensive care units and 28% requiring mechanical help to breathe. About 60% of the study participants were male. Roughly half the patients were white, while 38% were Hispanic, 19% were Black, and 4% were Asian. The study classified more than a quarter of patients as “other or unknown.”

The concept behind convalescent plasma, which dates back more than a century, is that blood from patients who have recovered from an infection will be rich with antibodies against it, making their plasma an effective treatment. A host of pharmaceutical companies are developing lab-grown Covid-19 antibodies that would work much the same way, and the many vaccines in development are meant to spur the production of similar antibodies that would prevent infection in the first place.

As doctors around the world scramble to find effective therapies for Covid-19, studies like this one present enticing data but do little to address the clinical guesswork physicians face when treating the disease.

“It raises the question of what strength of evidence is necessary to treat during a pandemic,” said Harlan Krumholz, director of the Center for Outcomes Research and Evaluation at Yale New Haven Hospital. “The problem is we have yet to resolve what is sufficient evidence to change the treatment paradigm.”

Steven Nissen, a noted clinical trialist at the Cleveland Clinic, agreed: “The lack of high-quality trials in making clinical decisions about how to treat patients with coronavirus infection is a national embarrassment. Here we have another non-randomized study, NIH-funded, and uninterpretable.”

Randomized, placebo-control studies of convalescent plasma are ongoing.

  • With enough information on the patients’ age, medical history, severity of infection, treatment given in addition to convalescent plasma etc. in the database, it would seem to be fairly simple to find good matches in the millions of patients who have been treated for coronavirus (but without plasma) for a great majority of the 35,000 patients. Even just 15,000 would be enough to equal the numbers used for Moderna’s vaccine trials.

    How well does Moderna match people in the vaccinated group and the control group, anyway? Instead of pushing for trials that meet some unspecified standard, this would be a good benchmark for the degree of matching to aim for in a retrospective study. The advantage of such a study over one that is run by rigid adherence to the so-called “scientific method” is that no new people are put to any risk.

    • Lest the word “benchmark” be misunderstood, I should mention that I do not want to set the bar that low unless Moderna did an exemplary job of matching the members of the control group with those given the vaccine. I would like for the kind of study I have in mind to set the standard of excellence for such matchings.

      It may well be that the people in charge of the Mayo Clinic experiment have just such a study in mind. This is suggested by the following excerpt from the article that reported on the experiment:

      It was the intent a priori to create a control comparator group to determine
      potential efficacy using patients hospitalized with COVID-19 infections
      during the same time period. This decision was made after collaboration with
      the US FDA. … A future publication will discuss potential efficacy.

  • We don’t need clinical study this plasma treatment works. It has healed thousands we need a major effort of recruiting donors of plasma with Corona antibodies. Our TV news is not mentioning this treatment nor are they asking for donors.

    • There is a billboard in my subdivision which is letting people recently recovered, or convalescing, from Covid-19 asking them to consider donating plasma to a center whose phone number, etc. are given. Apparently it will take hundreds of thousands of such small ads to reach the millions whose plasma can make the difference between massive lung damage and minor effects — or even between life and death.

      The major outlets are publicizing the efforts to produce a vaccine because there are billions to be made there, and no such money from cures.

      Also, there is a concerted effort, including a recent article in the New York Times, to promote the idea that we will be totally in the dark about the efficacy of convalescent plasma until “randomized control trials” are done. But these trials are not designed to yield what criminal courts call “proof beyond a reasonable doubt.”

      The best we can expect is that “The hypothesis that this treatment improves patient outcome is supported.” The worst [extremely unlikely, by all available evidence] is “The hypothesis that this treatment improves patient outcome is falsified.” But even this unlikely possibility would be very easily found by the matching of patients in all relevant respects without putting anyone at a new risk. That is not true of randomized control trials.

  • We are so hungry for a treatment that we are almost ready to accept results of any type of study. We should stick to randomized, placebo-control studies and avoid losing time, patients and money in uninterpretable studies.

    • Au contraire, I believe the real waste of time would be to drag our feet on the contribution and use of convalescent plasma on the grounds of a naive faith in the degree of certainty we can expect from randomized, placebo-control studies.

      The so-called “scientific method” isn’t designed to achieve near-certainty on the level of “beyond a reasonable doubt” that is required in criminal law. It is designed with a hope to arrive at one of two conclusions: “The hypothesis was supported” and “The hypothesis was falsified.” But even “falsified” carries with it a degree of uncertainty, especially if the experiment was not optimally designed.

      Note the use of the word “support” in the following excerpt from the article that reported on the Mayo Clinic experiment:

      Conclusion: These updated data provide robust evidence that transfusion
      of convalescent plasma is safe in hospitalized patients with COVID-19,
      and support the notion that earlier administration of plasma within the
      clinical course of COVID-19 is more likely to reduce mortality.

  • There are three types of people involved with COVID decisions: scientists, experts, and leaders.

    1) The scientist relies just on data from lengthy studies where they can read the study results for their decisions. Concerning COVID antibodies, they have not run enough proper studies to make conclusions about antibodies. Scientists are very slow to provide useful answers. Right now, they are hesitant about giving helpful information on plasma treatments and about antibody immunity. A good question for a scientist, how long does it take to complete a study, 2022?

    2) The expert can make recommendations based on the current data and historical information. For plasma treatment, the expert may recommend plasma treatments based on existing data that there are promising results with little adverse effects and historically plasma treatments have been effective for 100 years

    3) The leader will make decisions based on the pros and cons. A leader will evaluate if it is better to take unknown risks to save lives now or let people die while waiting for more conclusive data.

    We need more leaders.

  • Oh for Pete’s sake just give it to everybody already! There has been almost no evidence of harm caused by convalescent plasma infusions amongst multiple observational studies and this study provides strong evidence that the treatment works given the dose dependent mortality differences and the early treatment mortality differences. This is on top of countless observational studies for COVID-19, SARS and MERS that convalescent plasma appears to reduce mortality.

    You can always run a retrospective matched cohort study of patients after a period of time of universal treatment of hospitalised patients with convalescent plasma to determine whether there has been a reduction in the case fatality rate amongst hospitalised patients.

    What is a national embarrassment is not the lack of a completed RCT on this treatment but the complete inability of the US (and other western countries) medical profession to get its head around the need to potentially make treatment decisions without the benefit of a RCT. This is bizarre given they seemed happy to throw hundreds upon hundreds of thousands onto ventilators prematurely with absolutely ZERO evidence of efficacy or even safety in COVID-19 induced ARDS.

    Similarly the US medical profession continues to overwhelmingly use face masks for non-invasive ventilation of COVID-19 induced ARDS despite a damning study on the inferiority of face mask based non-invasive ventilation vs. helmet based non-invasive ventilation! How many COVID-19 patients have needed to be intubated and died due to the unwillingness of medical practitioners to change treatment protocols around NIV I wonder???,occurs%20when%20fluid%20builds%20up%20in%20your%20lungs.

    • Yes, 35,000 patients is more than enough to determine that blood plasma works. On top of that it has historically worked on past infections for over 100 years now. It is sad when Big Pharma doesn’t stay out of their influential way of allowing others with treatments that they do not produce. Enough is enough.

  • May be questions but it is a start. Had to determine FIRST that providing the plasma didn’t cause more harm or side effects. And it didn’t cause harm or side effects.
    Husbands oldest daughter is one of the authors of this paper so we’ve been keeping up with the progress.
    Frankly the more ways/tools we have to treat the severe infections the better!

    • SUBJECT: I have a suggestion to do an indirect challenge trial, by extracting convalescent plasma from vaccine recipients in Phase III
      TO: [email protected]; [email protected]; [email protected]; [email protected]; [email protected], [email protected], [email protected];
      Hello !
      First off all I don’t have a medical pedigree, but I do have a strong tendency towards divergent thinking.
      Maybe it’s been tried already, but it seems logical that the plasma collected from vaccine recipients should be the same as convalescent plasma which has proven to be helpful to COVID-19 patients.
      To be 100% safe there are two possible testing approaches:
      1. Inject the plasma from vaccine participants who never had COVID-19 into animals whose immune response is closest to our own, and then expose the animals to COVID-19.
      2. Inject the plasma from vaccine participants who never had COVID-19 into people who are at the earliest stage of COVID-19. The fail safe is if it doesn’t work give them actual convalescent plasma.
      What’s it good for ?
      1. If there is a shortage of convalescent plasma available and the vaccine recipient’s plasma proves a viable substitute you’ve just increased the supply available.
      2. If the vaccine is acting the same as convalescent plasma, helping animals or patients with COVID-19, that is animals used for challenge versus humans who already have early stage COVID-19 then it can provide fairly convincing proof that the vaccine works.
      Of course the scientists who specialize in this will know better than me whether the idea, which seems sound will actually work, but you never know until you try.
      I’m hearing that the vaccine will be approved by mid-2021 to late-2021. It’s all over the map, but this may provide the confidence to move the approval to an earlier date if it works as I propose.
      Best Regards,
      Dan Hearn

Comments are closed.