Science’s COVID-19 reporting is supported by the Pulitzer Center and the Heising-Simons Foundation.
Now that regulators around the world have begun to issue emergency use authorizations (EUAs) for COVID-19 vaccines—the United States authorized a candidate vaccine from the biotech Moderna on Friday—a theoretical debate that has simmered for months has become a pressing reality: Should ongoing vaccine efficacy studies inform their tens of thousands of volunteers whether they were injected with a placebo or the vaccine, and also offer an already authorized vaccine to those who got the placebo?
Vaccinemakers must now quickly decide how to handle this issue, called unblinding. And if they do choose to unblind, they will also need to get regulatory approval. Adding to the pressure: The choice arrives as many trial participants in the United States who are now eligible for an authorized COVID-19 vaccine are dropping out of studies in order to make sure they get immunized.
At the heart of the dilemma is a balancing act. On the one hand, unblinding a vaccine efficacy trial compromises the ability to gather robust scientific data on important issues, such as how long a vaccine protects a person against COVID-19. On the other, withholding a working vaccine from a trial participant who could get it elsewhere is ethically dicey.
Last week, those issues came to the fore at a meeting of a U.S. Food and Drug Administration (FDA) vaccine advisory committee, which was considering Moderna’s EUA request. Lindsey Baden, a principal investigator on Moderna’s vaccine efficacy trial, and Steven Goodman, an epidemiologist from Stanford University, presented two different schemes for handling the delicate matter. Both schemes differ from a third plan put forth by Pfizer and its partner BioNTech, which are already distributing a COVID-19 vaccine authorized for use in the United States, the United Kingdom, and the European Union.
Pfizer, which is still running trials of its vaccine, has already asked FDA for permission to unblind its study, and give volunteers who received a placebo the option of receiving its two-dose vaccine. But the offer would come with a catch. Given the limited supply of vaccine, national or local authorities have said they will first provide it to groups most at risk of becoming seriously ill or of transmitting the virus, such as front-line health care workers or the people living in nursing homes. The Pfizer plan would follow that plan by first unblinding only volunteers in one of the top-priority groups specified by authorities in their location and vaccinating those in the placebo group. (AstraZeneca and the University of Oxford, which jointly developed a COVID-19 vaccine candidate that has confusing efficacy results and does not yet have an EUA, reportedly want to offer a similar unblinding scheme.)
A different approach is advocated by Baden, an infectious disease specialist at Brigham and Women’s Hospital. He prefers an “open label” strategy, in which the Moderna trial is unblinded and the company offers everyone in the placebo group the vaccine regardless of eligibility in their location.
Goodman, meanwhile, favors a “blinded crossover” option. In that approach, all trial volunteers would receive an additional pair of shots, with placebo recipients getting the vaccine and the vaccinated getting placebos. But the volunteers wouldn’t be told which arm they were in for some period of time.
At the FDA meeting, Baden acknowledged that Goodman’s plan has some advantages. A blinded crossover potentially minimizes bias and allows better data collection that could better clarify how long a vaccine protects people. But Baden also thinks the blinded crossover was impractical for Moderna, given where its study now stands.
Goodman contends that both open-label and blind crossover are good choices, but much could be gained by trying to make a blinded crossover work. “The ability to do placebo controlled [COVID-19 vaccine] trials is running out,” Goodman says. “So let’s try to squeeze out of these few as much as we possibly can.”
Peter Marks, who heads FDA’s vaccine division and was the initial architect of the U.S. government’s crash COVID-19 vaccine project he dubbed Operation Warp Speed, says he and his team recognize the merits of the blinded crossover, but also acknowledge the practicality issues. “This would have been an absolutely valid design from the start,” Marks says. “Missing that may have been a casualty of working at warp speed.” His team will decide on Pfizer’s open-label request over the next few days.
ScienceInsider recently spoke with Goodman, Baden, and others to make sense of the ethical challenges and scientific trade-offs of these options. The discussions center on the situation in the United States, although similar debates will play out most everywhere as COVID-19 vaccines are given EUAs and distributed.
What is the proposal from Pfizer and BioNTech?
The proposal from Pfizer and BioNTech is a narrower version of the open-label approach. It is conditional on a person in the study being in a priority group and eligible for a vaccine; health care workers in the United States who joined their trial could be unblinded, for example, and offered the vaccine if in the placebo arm. “We believe we have an ethical duty to create a pathway within the study for interested, eligible participants in the placebo group to have access to the investigational vaccine,” a Pfizer spokesperson wrote to ScienceInsider. People who live in or work at long-term elderly care facilities in the United States have also been prioritized and would therefore now be eligible for unblinding. This weekend, an immunization advisory panel to the U.S. Centers for Disease Control and Prevention recommended the next groups to be vaccinated in the country should be front-line essential workers, like teachers and bus drivers, and people 75 and older. That means many more trial participants will soon be considering their options.
In what Pfizer calls the “vaccine transition option,” it hopes participants who request their status, learn they got a placebo, and then get a COVID-19 vaccine from the trial will continue to report any possible COVID-19 illnesses and provide blood samples and nasal swabs. “Participants should not have to choose between remaining in the study or receiving the investigational vaccine if it is available to them in their community,” its spokesperson says.
But Pfizer stresses that learning one’s vaccination status is optional. The company is encouraging participants to stay in the original study and remain blinded, even if they are vaccine eligible, for 6 months after they receive their second dose of placebo or vaccine. At that point, everyone would learn their status. Vaccine supply presumably will no longer be an issue by then, and all placebo recipients would be offered the vaccine.
What are the research advantages of this proposal, compared with the full open-label scheme?
Because the narrower approach would unblind and vaccinate only those eligible, the number of people in the placebo group would dwindle more slowly, allowing for longer comparisons of the two arms.
Are there ethical reasons to link unblinding and eligibility?
Yes. If trial participants who received the placebo were offered the vaccine and were not yet in an eligible group—many are young, healthy adults who are at low risk—they could be seen as “jumping the queue.” Goodman warned at the FDA meeting that this could create something of a public relations nightmare. “This will get out in the community [and look] very different from the pictures we saw after the authorization last week where it was health care workers and the others.” This, in turn, could erode trust in the allocation system, he said.
How would a blinded crossover approach differ and what are its advantages?
A blinded crossover would give trial volunteers in the placebo group the vaccine, but participants would not know their status until enough time had passed so that everyone in the trial had a few weeks to develop full immunity. This crossover might only occur with eligible participants, but it could also be done with everyone at the same time.
In the Pfizer-BioNTech and Moderna efficacy trials, all participants have received either two placebo or two vaccine shots. A blinded crossover would ask everyone to agree to receive two more shots, but researchers would not reveal whether the participants were initially in the placebo or the vaccinated arm. If they already received the vaccine, the two new shots—spaced 3 weeks apart for Pfizer and BioNtech, 4 weeks for Moderna—will contain placebo doses of saline. Those in the placebo group will get two doses of the actual vaccine.
In the first period of the trial, participants did not know their status, so behavior should have been the same in both placebo and vaccine recipients. Retaining the blind after a crossover avoids the “behavioral disinhibition” problem of people who know their vaccine status—they may drop their guards by, say, not wearing masks as regularly—that would diminish the ability to conduct meaningful comparisons of the two groups.
This strategy could help answer the critical question of how long vaccine-induced immunity lasts. Basically, researchers can compare the number of people who get sick in each group starting the day the second period begins. It takes 2 weeks after the booster for immunity to peak, so adding in the time between shots, there are five to six interesting weeks when the two groups clearly have different immune status. Following the two groups for several more months should reveal when immunity begins to wane in the vacinees from the first period.
Does Baden oppose blinded crossover trials?
Hardly. He co-authored a paper, which came out as a preprint a few days before the Moderna EUA meeting, advocating the idea for COVID-19 vaccine efficacy trials. But he thinks a blinded crossover requires some time to set up and cannot be launched on very short notice. Baden notes the switch for the Moderna trial would need approvals from FDA, institutional review boards that oversee studies, and 99 clinical trial sites. Researchers would have to explain the change to 30,000 or more participants and receive their consent. And hundreds of health care workers—who make up 25% of the participants in the Moderna study—have already dropped out of the trial because they have become eligible for the vaccine through other means.
Launching a blinded crossover study also presents a huge communication challenge to people inside and outside the study. “The more we are changing our approach, the more complex the messaging,” Baden says. “I want to minimize and not add to this confusion as the community trust and volunteer trust is paramount.”
Should Moderna tie the open-label design to eligibility?
No, says Baden, who at the FDA meeting added a surprising detail that alters the ethical calculations: Moderna has “research-labeled vaccine” that was set aside only for the trial and is about to expire. So he advocates rolling out vaccine to placebo recipients as soon as it is logistically possible as part of the ongoing research study.
How did the FDA advisory committee react to the two ideas?
They seemed to favor the open label, which is the traditional scheme used in vaccine efficacy trials, but FDA asked the panel not to vote on the question for Moderna’s trial. Goodman and Baden did not debate each other, and Goodman thinks this led to much confusion among advisory panel members about which scheme would most likely encourage participants to stay in the study. Goodman worries there is a mistaken notion that the blinded crossover withholds vaccine from the placebo group. “It was as though each member of the committee had a different idea of what was being proposed,” he says.
What do the companies think of the blinded crossover idea?
At the 10 December advisory committee hearing for the EUA of the Pfizer-BioNTech vaccine, William Gruber, who heads clinical vaccine development at Pfizer, said the scheme was overly cumbersome, noting there are 44,000 people in the efficacy trial who already made two visits each to trial sites for their pair of shots. “Obviously the considerations that were discussed by Dr. Goodman have been front and center for us in thinking about how we have looked forward,” Gruber said. “The logistics are not trivial because we’ve got to bring in a lot of people to be vaccinated and if we now add that to 88,000 visits, that’s a challenge.”
What will now happen to the many other COVID-19 vaccines still in efficacy trials or nearing them?
The Moderna trial debate, Goodman says, “will look very simple compared to the dilemmas we’re going face in the next few weeks and months with the other vaccines.” That, he says, bolsters the argument for Moderna using the blinded crossover. “I tried to make a very, very, very strong case that what they do with this trial has implications for the future, because if they allow Moderna to unblind and blow up all aspects of the placebo control—even though they will still be able to extract some information—it gives companies in the future permission to do the same thing,” Goodman says.