In 1901, Italian doctor Francesco Cenci discovered that blood infusions from recovered measles patients could keep other people from getting the illness. A few years later, he discovered that the same treatment helped a critically ill child to recover faster. Similar cases were reported with diptheria, and the treatment was then used in the 1918-19 flu pandemic. Scientists would eventually discover that plasma, the yellowish, watery part of the blood in which red blood cells are suspended, contains antibodies to fight off infection.
On Sunday afternoon, hailing so-called convalescent plasma as a “breakthrough,” President Trump announced an emergency use authorization, or EUA, of the technology for Covid-19. The EUA, which allows a treatment to be used despite not going through the typical tests, will “dramatically expand access to this treatment,” Trump said, calling it a “truly historic announcement.”
Setting aside the fact that it’s awfully hard to call a technology that predates penicillin a “breakthrough,” the announcement raised a number of questions. How well does convalescent plasma work against Covid-19? What’s the appropriate dose, how soon does it need to be given in the course of illness, and which types of convalescent plasma work best? No one really knows, and the prospects for sorting it out now seem dimmer than ever. Although everyone wants to prevent more coronavirus deaths, experts worry that the announcement of plasma’s expanded use will upend important research on plasma and antibodies—knowledge about immunology that might have led to a cure or vaccine. They also fear the announcement will set a precedent, particularly as the November election looms, for bringing treatments to market before they’ve been fully tested.
A recent study of 35,000 patients conducted by the Mayo Clinic and sponsored by the National Institutes of Health, or NIH, found that giving patients convalescent plasma early in their illness seemed to reduce their chances of dying. Some patients were given plasma within three days of diagnosis, while others were given the plasma after four days. Those who received it early had an 8.7 percent mortality rate in the next week, while those who received plasma later had an 11.9 percent mortality rate. But the preprint study, which has not yet been published in a journal or peer-reviewed by other experts, didn’t have a control group of patients who received a placebo—making it difficult to judge definitively whether it was the convalescent plasma or some other factor that reduced mortality. Just because plasma transfusions seem to work against measles or the flu doesn’t mean they work well against Covid-19.
When the Mayo Clinic announced two weeks ago that the U.S. Food and Drug Administration, or FDA, would issue an emergency use authorization, the NIH, which sponsored the Mayo study, intervened, saying more rigorous studies need to be done—such as a randomized controlled trial, or RCT, the gold standard in research. The Trump administration pushed ahead anyway.
The president’s announcement does not actually change much about how convalescent plasma is used, experts told me. It was already approved for compassionate use in the United States under an expanded access program, which allows patients access to potentially life-saving treatments before they have been approved by the FDA. The EUA may reduce some of the paperwork for getting the plasma. And Trump’s press briefing may increase the use of the unproven treatment since now, with the public announcement, more people know they can simply ask their doctors for the treatment. That creates a problem.
“He touted it from the podium in the White House as being some kind of miracle cure,” John P. Moore, a professor of microbiology and immunology at Weill Cornell Medical College, told me. “There’s just no evidence that that’s true.” The announcement actually makes it less likely scientists will find that evidence: “The only way you can prove that it works is a randomized controlled trial. But they were already proving to be difficult to complete, and they’re now going to be really difficult to complete because patients won’t volunteer for them.”
Patients in a randomized controlled trial have a fifty-fifty chance of receiving a placebo. “If you’re a patient, why would you volunteer for a clinical trial and risk a 50 percent chance of getting a placebo when basically the physician can give you the same active component under an EUA?” Moore asked. “This announcement has, if not killed, then certainly seriously compromised the ability to prove whether or not convalescent plasma does actually work.”
According to the research that has been conducted so far, Moore said, convalescent plasma seems to have an effect only under certain circumstances: if the plasma has plenty of neutralizing antibodies and if the patient receives an infusion early in the illness. “If you use weak plasmas or if you delay it too long, there’s very little chance of seeing anything beneficial.”
Much of the plasma now being used outside of clinical trials is not well monitored for how many antibodies, and what type, it contains. “Unless you’ve actually gone and measured what’s in it, you don’t know what’s in it, and then you’re kind of keeping your fingers crossed,” Moore said. “Hard data would have been really useful, but now it will become harder and harder to get hard data.”
When it comes to antibodies, Meghan May, a professor of microbiology and infectious disease at the University of New England, told me, “they’re not all created equal.” Some antibodies may be more effective than others at responding to the virus, but scientists don’t know which ones are best yet. “We don’t understand immunity and protection against this virus because we haven’t had enough literal time in history to study it and study recovered patients.” Researchers are able to measure the level of antibodies in a recovered patient’s plasma, she said. “But we don’t know how effective or what proportion of those antibodies are going to be able to neutralize the virus.” That’s what studies were trying to show. “Now that this authorization exists, it becomes impossible to properly do the study.”
In addition to the knowledge gaps, the resource itself is limited. While some recovered patients with high levels of neutralizing antibodies donated earlier in the pandemic, the antibodies eventually wane, making the pool of donors smaller. “We don’t have enough lots of plasma for all the people who would need it,” May said (lots refer to units of plasma). That brings up thorny ethical questions. “How do you decide who gets it? Is it based on payment; is it based on geography; what is it based on?” It’s possible that the emergency use authorization could introduce further inequality to who is able to afford the treatment, she said.
Convalescent plasma doesn’t seem to make Covid-19 worse in those who receive the treatment, as Ebola plasma may do among some patients. And it doesn’t seem to have negative side effects like hydroxychloroquine—the previous medication announced by the FDA and touted by Trump for emergency use in the pandemic—may have. However, if people think convalescent plasma is a tried-and-true cure for Covid-19, they might take fewer precautions to avoid the virus—which could result in more deaths than if the treatment weren’t used at all.
Beyond questions of how well plasma works, May is concerned this week’s announcement may pave the way for using an EUA for future treatments—for instance, a vaccine that has not completed clinical trials yet. “I’m really, really, really worried about the precedent that this set,” she said. “Is this a normalizing of the act of bypassing the FDA? I hate thinking like that, but it worries me.”
It seems to worry most Americans, as well, who say they want treatments to be tested for safety and efficacy first, even if it means waiting. One poll in July found that more than 60 percent of Americans prefer to wait for a fully tested vaccine. If the process is rushed, some people might decide not to get it. “A vaccine that’s not widely used is a vaccine that’s useless,” Moore said. And if a rushed vaccine ends up having side effects, he said, “that’s going to have a huge negative impact on the public in the longer term,” possibly eroding trust in all vaccines.
Joshua Sharfstein, vice dean for public health practice and community engagement at the Johns Hopkins Bloomberg School of Public Health, believes that convalescent plasma has the potential to help Covid-19 patients and emergency use authorizations can be effective at bringing the treatment to people who otherwise wouldn’t have access to it. But he’s sympathetic to worries about the precedent the EUA may set. “I would have wished there’d been a little bit more discussion of that because that’s obviously a concern that people have at NIH,” he told me. More than 50,000 patients have already received the plasma under compassionate use, he added. “By the time you have 50,000 patients getting a treatment, we’d hoped we’d have better clinical trial data.”
At the White House briefing, officials inaccurately portrayed the limited data that is available. “If you are one of those thirty-five out of one hundred people who these data suggest survive as a result of it, this is pretty significant,” said Stephen Hahn, FDA commissioner, discussing the plasma infusions. But that 35 percent, it turns out, referred to the Mayo Clinic study where patients who received the plasma earlier in their illness had a 35 percent better chance of recovering than patients who received the plasma later. That does not mean convalescent plasma reduces mortality 35 percent of the time, as Hahn implied. (In a series of tweets last night, the commissioner walked back the claim.)
“People really need to trust the FDA, and trust comes from independence and commitment to science,” Sharfstein said. “Announcing things with the White House with misleading statistics is neither independent nor scientific.”
Risking public trust in medicine is dangerous on its own. But if the announcement affects research in the way other experts fear or if it clears the way to bypass the regulatory agency entirely, the consequences could be huge.