When the immune system goes awry, it’s bad news. A wonky immune system might mean that you’re more likely to catch colds and flus, or be infected by other pathogens—and less likely to shake them off. It might mean that your body fails to detect and destroy growing tumors. It might even mean that the body turns against itself, leading to chronic autoimmune conditions like arthritis or Crohn’s disease. The fallout of immune system dysfunction on the human body is widespread and unpredictable—which is why it was so concerning in 2020 when evidence began to amass that COVID-19 seemed to be disrupting human immunology. So much so, in fact, that John Wherry, director of the Penn Medicine Immune Health Institute, summed it up this way to Kaiser Health News: “COVID is deranging the immune system.”

Most of the early immunological evidence—the evidence that Wherry was referring to—came from patients who died or suffered severe COVID. Now, three years of infections and immunizations later, severe COVID is getting mercifully less common; a brush with the virus may well feel unremarkable. And a new idea about how COVID can affect immunity has emerged: that even mild infections routinely cause consequential damage to our bodies’ defenses. This quiet degradation was memorably termed “immunity theft” by one evolutionary biologist speculating on why this fall’s respiratory virus season seemed more severe than usual.

There are plenty of reasons to not want to get COVID over and over, but the prospect of an increasingly damaged immune system is a particularly compelling one. Throughout the pandemic, scientific evidence has mounted that mild COVID infections may be doing something to our immune systems—prompting many on social media to hyperbolically describe COVID as “airborne AIDS.” But the lessons that scientists are drawing from their research are nuanced—and the larger picture says more about the sturdiness of our collective immunity than anything else.

There are a few ways scientists can probe COVID’s impact on immune systems. One is to investigate how well the immune system can rally against a second go-round with SARS-CoV-2. At the start of the pandemic, Shane Crotty, of the La Jolla Institute of Immunology, published some of the first papers looking at the immune response to COVID. “There was a lot of concern about how strange it might look,” he said. But really, “it looks as we would largely expect for a respiratory viral infection.” Antibodies recognize and subdue the virus, while immune memory cells linger about, ready to gear up for the next infection. A similar response is seen after vaccination. Because of this robust immune response, SARS-CoV-2 infections are now, on average, shorter and milder. So far this year, COVID hospitalizations have not surged, despite high rates of infection. Some of this attenuation may be due to a meeker (arguably) omicron variant, but it’s more likely because, with respect to fending off COVID at least, our immune systems are working just as they’re supposed to.

But there’s another way to think about COVID’s immunological impact. What if SARS-CoV-2 infection fortifies our immune systems in very specific ways such that we can stave off severe COVID, but precipitates subtler, long-term immunological changes that leave us more vulnerable to other infections or even chronic disease? The data here is murkier.

Scientists know that during severe cases of COVID, things go immunologically haywire. A study from the pandemic’s early days, in January of 2020, profiled 41 hospitalized COVID patients in China and found that 63 percent of them had low numbers of lymphocytes, a critical type of disease-fighting white blood cell. A postmortem study found that patients who had died of COVID lacked germinal centers, which teach immune cells to mount a long-lasting response to infection. A few studies—which looked at hospitalized patients, and cells in petri dishes—have claimed that SARS-CoV-2 can directly infect immune cells, and others have found that the virus can stir up “autoantibodies,” or immunological turncoats that blitz the patient’s own proteins and cells. Wherry is a co-author on one such study; this is the kind of “deranging” he was talking about. Immune system derangement appears to be what can make some severe cases of COVID so horrible.

Since then, many other studies have unearthed immunological oddities with worrisome names like “T cell exhaustion” and “dendritic cell deficiencies”; sometimes the oddities are seen in patients with just mild COVID infections. These studies can fuel scary-sounding headlines (“Is COVID prematurely aging our immune systems?”). Their top-line results are often circulated as validation of widespread and ongoing immune dysregulation. But that’s not really true.

To understand why these studies aren’t validation requires going a bit in the methodological weeds. In most of them, researchers take a group of patients who suffered from COVID—often a mix of mild, moderate, and severe cases—and find a group of “matched controls,” or uninfected people with similar health and demographic profiles. Then, they train the power of modern biotechnology on the immune system, scanning hundreds or thousands of cells, genes, and molecules for anything that looks different in COVID patients vs. their never-infected counterparts. The data sets are huge, and sophisticated machine-learning algorithms are often deployed to sift signal from noise.

These are impressive techniques, but it’s important to keep in mind the purpose of these studies, and their limitations. Most are conducted in small groups of patients who may or may not be representative of the population as a whole. And from the small group, researchers collect lots and lots of data. This style of research is exploratory, designed to pick out avenues for future and more robust study. Some of the leads may pan out, but many—if not most—findings will just end up being natural variation between people’s bodies, popping up by chance in those who had COVID.

Few, if any, studies have information on the state of patients’ own immune systems before they were infected with COVID, making an apples-to-apples comparison of what COVID does to a particular person’s immune system impossible. Robust longitudinal data starting prior to the pandemic would show “whether we’ve seen large-scale changes in immune fitness,” Wherry told me—and we just don’t have it. In its absence, “the evidence of a long-term impact on the immune system in fully recovered COVID patients, whether mild or severe, is really pretty thin.”

“There are some diseases where there is a clear immune signature that would make me worried,” Crotty, the immunologist at the La Jolla Institute, said. “For example, if you catch measles, you end up more susceptible to other infections for several months. But I haven’t seen anything in our data or other studies that makes me worried about long-term impacts on immunity to other infections.”

Even if the laboratory studies aren’t conclusive, there is some real-world evidence that mild COVID infections can throw the immune system out of whack. The latest is a large observational study posted on a preprint server just last week by German researchers. At first glance, the findings are scary—they found a whopping 43 percent increase in the onset of autoimmune disease in COVID patients compared with noninfected controls. But it’s important to put that number in context. First, the infections were in 2020, before vaccines. Second, that 43 percent is relative risk. In absolute terms, the study found that 1.1 percent of people developed autoimmune disease after catching COVID; 0.8 percent of controls developed autoimmune disease during the same period. That’s a 0.3 percent difference. The study size was huge, so that small difference could very well be real and a cause for concern on a population level. (The study has not yet been peer-reviewed.) But it’s still a fairly rare outcome.

Less rare is long COVID, and it’s possible immune dysfunction could be linked to this worrisome condition. Scientists who have looked for obvious signs of how immunological dysfunction produces long COVID symptoms haven’t found it. “If there’s chronic immune activation that’s damaging tissue to produce the symptoms, then we should be able to detect that tissue damage. And that was what we couldn’t do,” said Michael Sneller, an infectious-disease specialist at the National Institute of Allergy and Infectious Diseases who is running one of the more comprehensive longitudinal studies on long COVID patients.

But long COVID is far from solved, so scientists are still probing for clues about how and why it comes about—this is where some of that exploratory research based on small sample sizes and broad swaths of data may prove valuable.

In the late 2000s, researchers in Vietnam found that infection with the typhoid bacterium left an imprint on the immune system for at least a year. When COVID hit, David Lynn, a co-author on that paper and a professor at the South Australian Health and Medical Research Institute and Flinders University, put together a grant application to determine whether SARS-CoV-2 did the same. Lynn and his team followed 69 patients in Australia with COVID infections ranging from mild to very severe. In contrast to standard clinical tests, Lynn used newer, sophisticated molecular techniques that can pick up much subtler signals. With a more fine-toothed comb, Lynn was able to find that almost every patient showed signs of a perturbed immune system at the molecular level, compared with healthy controls—but in most patients, these perturbations faded away after a few months. In a third of them, the immune system remained in a state of dysregulation. Many of those patients were later identified as having long COVID. “That was quite remarkable to us,” Lynn told me. Other studies back up Lynn’s findings, including one study posted on a preprint server by renowned Yale immunologist Akiko Iwasaki, whose team was able to predict which patients had developed long COVID by profiling immunological data at the molecular level alone.

One possible explanation for this small-scale perturbation is that long COVID’s symptoms could be the result of a hidden confrontation between the immune system and active virus ensconced deep in the body’s tissues—like combat among soldiers in remote outposts who don’t yet know the war is over. Other explanations include autoantibodies, reactivation of other latent viruses, or difficult-to-detect tissue damage from microclots. But these hypotheses remain speculation. And these newer molecular techniques haven’t been as widely used to investigate the immunology of other pathogens, meaning it’s unclear whether SARS-CoV-2 is even that exceptional. “I think all of the studies, including ours, are just giving hints to what’s going on,” Lynn said. “I don’t think we have a smoking gun at the moment.”

Importantly, since none of these studies have pre-pandemic snapshots of participant immune systems, it can’t be ruled out that patients may be more susceptible to developing long COVID if they already had a less-than-healthy immune system, muddling cause and consequence. Chansavath Phetsouphanh, an immunologist at the University of New South Wales’ Kirby Institute, was lead author on an Australian study that found hints of immunological dysfunction at least eight months after infection, particularly in patients with long COVID. “We know there are preexisting medical conditions that can make individuals more susceptible to developing long COVID,” Phetsouphanh said. This is a clue that “there may already be dysregulation of the immune system prior to getting infected.” As scientists work to untangle cause and effect, their research highlights the importance of health care and public health measures for people with compromised immune systems, whatever the cause.

At this point, the best we can say is that sensitive scientific tests can detect differences in the immune systems of people infected with SARS-CoV-2 compared with the ever-dwindling cohort of those who have never been infected at all. Whether these differences—which vary from study to study—add up to a molecular “immune signature” of COVID remains to be seen. Any clues may be particularly valuable in untangling why a subset of patients have symptoms that last a few months or longer. But to conclude that these detectable differences translate to real-world consequences is overreach. Further, to claim that COVID is destroying all of our immune systems, or is inflicting the direct and intense immune damage of an HIV infection, is absurd.

Wherry, the Penn Medicine immunologist, told me this—I think it’s worth laying out the quote in full:

We’re not seeing evidence that this one virus has changed our immune system’s ability to keep us healthy on a large scale. There are 7 billion people on the planet who are doing fairly well. And we’re not seeing opportunistic infections, we’re not seeing huge increases in cancers that need immune surveillance, we’re just not seeing the kinds of things that we saw in other settings where the immune system was compromised, or dysregulated, because of an infectious disease, or because of toxins, or because of radiation, that we’ve seen in a variety of human events over the past 100 years.

To claim that COVID could potentially discombobulate immune systems isn’t to peddle pseudoscience. People respond to infections very differently, and it’s worth continuing research into what COVID might be doing to some people’s ability to fight future infections, or how it might trigger chronic autoimmune disease. Even if long-lasting immune dysfunction develops in only a very small percentage of people, that still means a lot of individual suffering and a major toll on families and communities. In addition, noted Lynn, from a public health perspective, “that is a large additional burden on health-care systems around the world.” Further, the intense focus on the post-viral immunological effects of SARS-CoV-2 infection may shed light on what can happen after other infections—the possibility of long-lasting immunological changes in response to pathogens may be more common than previously thought. But researchers emphasize that even if COVID routinely tinkers with the immune system, our body’s defenses are stubbornly resilient. “Think of the immune system like a Boeing aircraft,” said Lynn. “For it to crash, you need multiple things to go wrong. Just one, or even a few things, is unlikely to be sufficient to bring the immune system down.”