Covid vaccination in the US has been framed as a binary: People either seek out the inoculation, or they distrust the formula—or the politics that produced it—and reject the shot. People who accept the vaccine get to return to normal life. For the people who don’t, “Your health is in your hands,” as Rochelle Walensky, director of the Centers for Disease Control and Prevention, tweeted in May.

But that binary never accounted for the many people who wanted the shot but couldn’t obtain it: because vaccine campaigns didn’t come to their neighborhoods, or lack of sick leave wouldn't let them risk side effects, or they didn't meet eligibility criteria in the early days when supply was limited. Now it’s becoming clear that this binary also excludes another enormous group: People who received the shot but were not protected by it, because their immune systems didn’t manufacture an adequate defense.

Millions of Americans are immunosuppressed or immune-compromised. That is, they take drugs to make sure that a transplanted organ is not rejected or to tamp down the overactive immunity that produces rheumatoid arthritis and lupus; or, alternatively, they have illnesses that undermine their ability to defend against pathogens. A handful of research papers published over the past few months all find the same result: When these patients receive Covid vaccines, their bodies don’t create as many defensive antibodies as those of healthy people. Some have contracted the disease despite being fully vaccinated—meaning that, to protect themselves, they must continue to behave as though their vaccinations never occurred.

As a result, some are seeking extra vaccinations, arranging for third doses that they hope will act like booster shots. A study published Monday in the Annals of Internal Medicine by a team at Johns Hopkins University School of Medicine documents the experience of 30 people living with organ transplants who sought out a third shot in hopes of boosting their immune responses. After their second shots, none of the 30 had high antibody levels; in fact, only six showed any antibody response at all. After the third shot, 14 out of 30 saw some improvement, and 12 of 30 had antibody levels that the researchers considered protective.

This is an important finding—even though it was made in a small group of self-selected volunteers, something that’s typically thought of as a weak study design—because it might point the way to letting still-vulnerable people rejoin a post-Covid society. It might also help explain some of the rare and not-well-explained “breakthrough” infections that occur in a small fraction of fully vaccinated people. (There was a high-profile cluster of breakthrough infections last month among people who work for the New York Yankees baseball team. No underlying conditions have been linked to them.)

“What we have found is hope,” says Dorry L. Segev, the study’s senior author and a Hopkins professor of surgery. “A lot of transplant patients were very excited about the vaccine, and then they got their antibodies checked and they were very suboptimal, and that was frustrating and demoralizing. We're showing something can be done about that.”

More in a minute on the practicalities of third shots. To understand why the possibility is so alluring—and sticky, in a regulatory sense—it’s worth looking first at what researchers have sussed out about how chronic immune conditions complicate protections against Covid.

Segev’s team was among the first to tackle this issue. Last December, one day after the Food and Drug Administration issued emergency use authorizations for the Pfizer and Moderna vaccines, the Hopkins group put out a call on social media. They asked immune-suppressed people to enroll in a trial that would keep track of their health and document their responses to the vaccines. They signed up almost 1,000 people in the first week, Segev says, and have almost 4,000 in a registry now.

In March, the group published the first data from those sign-ups: Among 436 transplant recipients, only 17 percent showed detectable antibodies after one dose of the Pfizer or Moderna vaccines. That was discouragingly low. By contrast, every participant in the mRNA vaccines’ clinical trials—100 percent—demonstrated antibody production by the 15th or 21st day after their first dose of the vaccine, according to data from the manufacturers and academic researchers.

A follow-up Hopkins study evaluated antibody responses after two doses in 658 transplant recipients from the social media volunteers, 396 of them participants from the first group. It yielded both good and bad news. Good: 259 members of the study group, 39 percent, had no antibodies after their first dose but gained them after their second. Bad: 46 percent of the participants had no antibody responses at all.

The Hopkins group wasn’t the only academic team thinking about this problem. Soon after the vaccines became available, physicians at Columbia University Medical Center in New York who were taking care of kidney transplant patients became worried. Several of their patients qualified for the earliest doses, yet after completing their vaccine courses and passing their “immune by” dates, several of them still came down with Covid. Worried, the nephrology department staff began asking any kidney transplant recipients who received the vaccines to have their antibody levels tested two weeks after their second shot. By the end of March, 28 patients went through that testing. Only 25 percent of them showed any antibody response.

This was disappointing, but, immunologically, it makes sense. After all, the purpose of vaccination is to induce the body to develop a protective response—expressed in antibodies, along with other immune system cells—that defends it against an infection. But the drug regimens that transplant recipients take, to keep their bodies from rejecting donated organs, suppress antibody production. That goal works against what the vaccine is intended to do.

“Patients’ immunosuppression regimens are designed to blunt their immune response to a new pathogen, because typically the biggest pathogen they're being exposed to, so to speak, is the new kidney,” says Syed Ali Husain, an assistant professor of medicine at Columbia and first author of an account of the findings, published in April. “Their immunosuppression regimen is designed to prevent antibody production.”

It’s important to say here that all these findings are mapping new territory. In order to get the cleanest, fastest authorization, people with immune deficits were excluded from Covid vaccine clinical trials. There was a broad assumption among physicians that those patients would have less response to the vaccines—but vaccine manufacturers didn’t attempt to gather that data. The run of studies that began this spring represents the first attempts to define vaccine effectiveness, or continued vulnerability, for this patient pool.

Organ transplant patients aren’t the only people bedeviled by low antibody counts after Covid vaccination. For instance, an April analysis submitted as a preprint (so not yet peer-reviewed or published in a journal) of how people with blood and bone marrow cancers responded to two doses of Covid vaccines showed that 46 percent of them made no antibodies. Some of those patients had their immune systems suppressed by drug regimens, and in others their cancers were to blame for their immune status. Patients with one particular blood cancer, known as B-cell chronic lymphocytic leukemia, were even less likely to make an immune response: Only 23 percent had detectable antibodies, even though most of them were not taking immune-suppressing drugs at the time.

These low percentages feel unnerving, but they may not tell the whole story. Ghady Haidar, the senior author on that analysis and a transplant infectious diseases physician at University of Pittsburgh Medical Center, says a frustration in translating risk to patients is that we still don’t have good metrics—”correlates of immunity” is the technical term—for what constitutes protection against Covid. That is, researchers know what protective antibody counts look like in healthy patients; but how far below that range an immune-compromised person can drop and still be protected isn’t yet known.

Plus, antibodies are not the only defense the body deploys to create immunity: We also make T cells, memory B cells, and others. The vaccine clinical trials didn’t attempt to measure the cell counts required to create an effective defense against the virus. They reported only clinical endpoints, such as whether someone became seriously ill or died from the disease. So focusing on antibodies alone may miss important parts of the immune response.

“I try not to use words like ‘you didn’t respond’ to the vaccine when someone isn’t making antibodies,” says Haidar, who is principal investigator on a larger study that is recruiting people with a range of immune deficits, including HIV, in order to study Covid vaccine response. “I worry that it might drive vaccine hesitancy if the messaging is that the vaccine isn't working for you. I think we have to be a little more nuanced to account for the complexities that other elements of the immune system could have been revved up by the vaccines.”

Even in the few studies done so far, it’s clear that immune response to the vaccines varies, depending on the age of the patient, the type of immune deficit they are experiencing, the type of transplant they received, the specific drugs they take, the length of time since transplant or last dose, and a host of other factors. The likelihood of abundant antibody production appears higher, for instance, in patients who take immune-suppressing medications to treat chronic inflammatory diseases than it does in transplant and cancer patients. Studies done by Segev and team show better rates of antibody production in those patients after one and two doses. But a separate preprint, done by Washington University School of Medicine in St. Louis and UC San Francisco, shows a wide range of responses depending on which drug regimen a patient is taking.

That may provide a clue for managing patients’ vulnerability, so that they can get closer to the kind of immune protection that otherwise healthy people receive from Covid vaccines. “One thing we are telling patients who are in suppression, who haven't gotten vaccinated yet, is to consider holding their medicine,” says Alfred H. J. Kim, that study’s senior author and an assistant professor of rheumatology and immunology at Washington University. “Obviously, if you hold medicines, you risk flares. And if you're going to flare, this could make your vaccine side effects worse, or it could make the vaccine itself less effective. It’s a really tricky situation.”

And, legally, doctors currently can’t advise patients to seek extra doses of the Covid vaccine. The FDA has authorized only one or two doses for all the vaccines it has let enter the US market. For Segev’s team’s study, the doctors didn’t prescribe third doses—the patients found third doses on their own, in ways the study did not specify. The Hopkins team tracked the results.

Still, there’s some evidence in medical literature to support the utility of additional doses. For instance, the French government has recommended a third dose for anyone who is immunocompromised. And in the US, it’s been understood for years that a second dose of seasonal flu vaccine and larger doses of hepatitis B vaccine are required to create immunity in them.

But it will be necessary to gather more data to be sure. The Hopkins team is contemplating a larger trial in which immunosuppressed patients seeking third doses would be enrolled and tracked in a formal way. And despite the allure of higher protection, they’re not urging immune-damaged patients to start freelancing their own third shots. “There are risks to taking third doses,” Segev says. “There is a risk the third dose will activate your immune system and cause either an overt rejection or some sort of subclinical thing, where you start to develop a little more antibodies against your transplanted organ. It's important that people who do go out and get third doses are either part of research protocols or are doing this in collaboration with their doctors who have evaluated the risks and benefits.”

If trials like this can yield data—another one, recently announced, is being conducted by the National Institutes of Health—they could do more than let the immunocompromised get back to everyday life. They could also illuminate aspects of the immune system and its interaction with vaccines that are still not very well understood. And that will be of benefit not just during this pandemic but for whatever we need to protect ourselves against next.