Risk Factors for COVID Are Only Barely Understood
On the surface, the September 24 announcement from the head of the CDC outlining who, exactly, would be eligible for COVID-19 booster shots seemed like a clarifying moment. But even as the agency’s leader, Rochelle Walensky, declared the need to make “concrete recommendations that optimize health,” the new guidance was hard to parse. It said, for instance, that people as young as age 18 who received the Pfizer vaccine may get a third shot as long as they have any of a list of “certain medical conditions” that might put them at “high risk” or “highest risk” for developing complications from severe COVID.
What are those high-risk conditions? The CDC has compiled a vague and partial list, presented in alphabetical order, that includes cancer, diabetes, liver disease, and smoking among 29 named conditions in all, divided into 17 categories. One of those categories—“immunocompromised state”—is itself a Russian doll of health disorders such as rheumatoid arthritis, lupus, and HIV; transient conditions such as pregnancy; and various treatment-induced vulnerabilities. Even the group most at risk from COVID—elderly people—can be said to be in an “immunocompromised state.” Researchers have termed the gradual weakening of bodily defenses as a person ages “immunosenescence,” and there’s evidence that an older immune system may also get stuck in an inflammatory state, a problem called “inflammaging.” So when we say that someone 85 years or older has 570 times the chance of dying from COVID as a young adult, we’re really using age as a stand-in for some invisible, underlying immune state.
Just as immune function—and its associated protection against severe COVID—tends to wane across our life span, it also varies along a continuum from person to person as a product of genetics. One evidently healthy 30-year-old, for example, could be more predisposed to getting very sick with COVID than another, even if they had the same set of medical conditions as listed on the CDC website. Scientists have been working out the details of these individual differences in immune function, but their findings haven’t yet been brought to bear on the pandemic in any widespread way, let alone considered guidance for the use of booster shots. Instead, as we struggle to set up rival groups of high- versus low-risk people, or immunocompromised versus immunocompetent, we tend to ignore all the gradations of vulnerability that might lie in between. Those who don’t clearly fit into the CDC’s official categories are left to guess at their personal levels of risk, counting their COVID antibodies “like calories” or grabbing booster shots of their own accord.
In the future, we may have more precise ways of gauging our individual vulnerability to COVID, or indeed to all infections. To some extent, we already do.
When I asked Harry Malech, the chief of the Genetic Immunotherapy Section at the National Institute of Allergy and Infectious Diseases, about natural differences in people’s ability to fight off disease, he recalled a case he took on in 2008. A young man with a strange constellation of symptoms had been referred to him for help with a diagnosis. The patient was in his late teens and had endured recurrent infections throughout his life. The roof of his mouth had become soft and was beginning to melt away, which was distressing. The patient’s sister had died in early childhood, so doctors wondered if a genetic factor was to blame.
At that point, scientists had decades of experience uncovering specific mutations that made certain patients unusually vulnerable to infections. The first inherited immune-system deficiency was described in 1950, when the Swedish pediatrician and army doctor Rolf Kostmann published a report of congenital neutropenia, in which babies are born with a shortage of white blood cells known as neutrophils. In the 1970s, a boy named David Vetter, who lived his life separated from the world by a plastic barrier, opened the public’s eyes to other inborn conditions that disable the immune system.
Yet almost 25 years after Vetter’s death, at age 12, clinicians like Malech still didn’t have a cheap, fast way to sequence and analyze the DNA of patients like the ailing teenage boy with mouth ulcers. Instead, they’d sequence tiny chunks of a patient’s genome after developing a hunch of where to look. By luck, one of Malech’s colleagues came across what was then a brand-new report in The New England Journal of Medicine mentioning a rare immune disorder in a few young girls who also had ulcers. “She said, ‘Maybe he’s got this. Let’s sequence this gene.’ And by golly, he had [the] deficiency,” Malech recalled. The teen received a bone-marrow transplant to reset his immune system, and he got better. That’s how things used to go, Malech told me: It would take “a sage diagnostician, coupled with a bit of serendipity and a whole lot of immune tests, to get at the heart of things.”
Now everything is different. “In the last seven to eight years, the entire field has changed,” Malech explained. “The ability to do rapid, cost-effective, high-throughput sequencing of people’s genomes has turned the whole process on its head.” You can easily obtain a patient’s “exome,” which tells you the code for all the proteins in their body. From there, faster and more powerful computers facilitate careful searches through those genetic sequences for mutations of interest.
Thanks to all these changes, scientists are finding DNA mutations with subtler—but still important—effects on the immune system. Today, more than 400 different chronic immunodeficiencies caused by genetic variations have been identified, according to the Immune Deficiency Foundation. Very few of these would land a person in a plastic bubble from infancy, but many could make someone more prone to repeat visits to the doctor’s office for infections as an adult—and perhaps more vulnerable to COVID.
Jean-Laurent Casanova, of Rockefeller University in New York City, co-leads a consortium called the COVID Human Genetic Effort. A year ago, he and his collaborators made headlines with a study in the journal Science that described how mutations affecting certain genes were more common in a group of more than 650 individuals with life-threatening COVID pneumonia than in their control counterparts, who were infected but asymptomatic. Specifically, the scientists found mutations that could disrupt immunity controlled by molecules known as type I interferons. Other scientists have not yet been able to replicate all of the same findings, but they haven’t discounted the importance of interferon in protecting people from severe disease.
More recently, Casanova and his teammates tried to get a sense of the scope of immune deficiencies in relatively young people who get very ill with COVID. In a paper published in August, they offered evidence suggesting that about 1 percent of men under 60 years old who developed life-threatening COVID have a mutation on the X chromosome that affects a receptor known as TLR7, which sits on the surface of immune cells and carries signals about microbe invaders. The scientists found this mutation in 16 of the more than 1,200 people with unexplained critical COVID in the study, but it was totally absent in more than 300 people who either had mild illness or were asymptomatic. The connection between TLR7 and severe COVID has been found by other groups as well. “Given that there have been many studies demonstrating the impact of TLR7 variants on COVID-19 severity, we believe this is likely a true signal,” says Tomoko Nakanishi, a respirologist at McGill University, in Montreal, who was in one such group.
An international collaboration has also uncovered genetic variants associated with severe illness from the coronavirus. Some are thought to diminish levels of an enzyme called oligoadenylate synthase, which normally helps chew up viruses. A paper published in just the past few days joined others in finding that a variant affecting one form of that enzyme is also associated with worse COVID outcomes; its authors note that this variant is common in all people, although less so among those with African ancestry. Yet another common variant—this one found in as many as 15 percent of individuals of European descent—could increase the risk of severe COVID by 70 percent, and by 170 percent in people less than 60 years old, according to a paper this month from Nakanishi and colleagues.
Not every immune-system glitch is necessarily predetermined at birth. Another study from a large, international group—this one including Casanova—skipped over the genome and looked for subtle immune disorders that people might acquire over time. In particular, the group found signs of antibodies that had gone rogue and were attacking the patients’ own immune molecules in about one-fifth of those from its sample who had died from COVID. It also looked at the immune systems of healthy individuals and found that older people were far more likely to have these same autoantibodies—a fact that could help explain why increasing age has such a strong association with COVID disease risk.
Casanova told me that the progress made in understanding immune-system vulnerabilities to COVID already outpaces that which he’s seen for other illnesses. For two decades, he tried to uncover underlying predispositions for tuberculosis, and while he did succeed in finding a seemingly relevant mutation in an immune-system enzyme called TYK2, it could account for only about 1 percent of cases among Europeans. Now scientists around the globe are making much faster headway with COVID, Casanova said. “I still can’t believe it.”
Given the lure of personalized genetic testing, some companies have started offering 23andMe-style diagnostics with a pandemic spin. Earlier this year, an Australian firm called Genetic Technologies collaborated with distributors to release an individualized, $175 assessment in the U.S. called the “COVID-19 Risk Test,” based on age; body mass index; preexisting conditions, such as diabetes; and seven DNA markers. Another testing company, Nutrigenomix, which has focused on nutritional genetic testing in the past, began offering an add-on feature to its 70-gene test this summer: For an additional $79, it tells you the status of your TAS2R38 gene, which codes for a receptor involved in bitter taste that has also been linked to poor COVID outcomes.
Some researchers see promise in personalized testing, reported Jocelyn Kaiser in Science this past June, but many believe that the underlying science remains too murky to sustain widespread genetic testing for COVID risk. Indeed, despite the run of notable findings in the research literature, there hasn’t been much push for understanding how to test and segment the U.S. population according to people’s genetic susceptibility to COVID.
Some wariness may be leftover from a confusing episode at the beginning of the pandemic. In early June 2020, a large team of scientists in Europe put forth preliminary data suggesting that, according to a genome-wide analysis of nearly 2,000 COVID patients in Italy and Spain, people with blood type A had a 50 percent increased risk of experiencing respiratory failure, while those with blood type O were somewhat protected. “A genetic test and a person’s blood type might provide useful tools for identifying those who may be at greater risk of serious illness,” wrote NIH Director Francis Collins at the time. 23andMe quickly followed with its own preliminary data suggesting that people with blood type O were more impervious to COVID. (A full study of the massive dataset from 23andMe data, published in April, confirmed this protective effect.)
But just a month after the initial raft of news about COVID blood types, reports emerged that two similar studies had failed to find a strong connection. People with blood type A were not more prone to falling severely ill from COVID, and any protective effect from blood type O was so small that scientists said it was basically useless. An author of one of these follow-up studies told The New York Times that the case for identifying genetic susceptibilities via simple blood-typing was closed. “I wouldn’t even bring it up,” she said. A year later, is it any wonder that COVID risk tests aren’t getting that much press?
According to Julien Textoris, a vice president of global medical affairs at the diagnostics company BioMérieux, we have a long way to go before people can make a doctor appointment to know if their immune system is up to snuff. Beyond the exceptional cases, such as inborn immune-system mutations or immunosuppression following an organ transplant, “there is no operational definition” of what it means to be immunosuppressed, he told me. For his part, Casanova believes that anyone who gets severely ill or dies from COVID is “immunodeficient” by definition—even if there is no current explanation for why they fared so poorly. Researchers say that continued advances in genetic sequencing will help unravel some of that mystery. “As to who counts as immunocompromised? I think it will be easier in the future to answer this question than it is today,” Wayne Koff, president and CEO of the Human Vaccines Project, told me.
Even if the science were all worked out, it would be an “expensive proposition” to do comprehensive sequencing of people’s coding DNA at a population level, Jeffrey Townsend, a biostatistician at the Yale School of Public Health who has studied COVID, told me. At the moment, we’re in a chicken-and-egg situation in that more genetic sequencing is needed for scientists to sort out which bits of DNA have the greatest influence on our immune function, which would in turn justify the cost of sequencing more people’s DNA. Complicating matters is the fact that our immune system derives from a vast network of genes. The genetic subtleties of COVID risk are intriguing, Townsend said, but we can already try to gauge people’s vulnerability to disease by means of simple SARS-CoV-2 antibody tests. “The evidence so far seems to indicate that antibody level is a major predictor of your level of defense against COVID-19 infection,” he said.
In the meantime, the further development and deployment of genetic methods for determining each person’s individual COVID risk may bring along some dangers of its own. In an essay published in June 2020, the sociologist Richard Milne warned that although such tools “may have potential value,” they could also lead to discrimination. People deemed to be particularly susceptible to COVID because of their DNA “may be advised to continue shielding or self-isolation measures long after the rest of the population,” he wrote, leading to significant psychological and financial hardships.
Even well-established markers for risk could be misleading when taken out of context. Researchers who were looking at the 23andMe data found that having blood type O offered slight protection against developing COVID—but the company’s previous research suggested that the same blood type was also a possible risk factor for seasonal flu. Or consider the discovery in the mid-1990s that a mutation to a certain immune-cell receptor could protect people from getting infected with HIV. A decade later, scientists learned that the same genetic quirk also puts people at higher risk of falling ill from West Nile virus. “One of the things that people don’t appreciate is that the human immune system is a compromise,” Malech said. “If you do better at A, you may be less good at B. There’s no free lunch here.”