A Negative COVID Test Has Never Been So Meaningless

A string of negatives can still presage a clear-as-day positive.

a series of blue coronavirus splotches, the last of which is pink
Getty; The Atlantic

In early May, 27-year-old Hayley Furmaniuk felt tired and a bit congested, but after rapid-testing negative for the coronavirus two days in a row, she dined indoors with friends. The next morning, her symptoms worsened. Knowing her parents were driving in for Mother’s Day, she tested again—and saw a very bright positive. Which meant three not-so-great things: She needed to cancel with her parents; she had likely exposed her friends; a test had apparently taken three days to register what her vaccinated body had already figured out.

Tests are not and never have been perfect, but since around the rise of Omicron, the problem of delayed positivity has gained some prominence. In recent months, many people have logged strings of negatives—three, four, even five or more days in a row—early in their COVID-symptom course. “I think it’s become more common,” says Amesh Adalja, an infectious-disease physician at the Johns Hopkins Center for Health Security.

No one can yet say how common these early negatives are, or who’s most at risk. But if SARS-CoV-2 is rewriting the early-infection playbook, “that makes it really scary,” says Susan Butler-Wu, a clinical microbiologist at USC’s Keck School of Medicine. “You can’t test and get a negative and actually know you’re negative.” Misleading negatives could hasten the spread of the virus; they could delay treatments premised on a positive test result. They also buck the current COVID dogma: Test as soon as you feel sick. The few days around the start of symptoms are supposed to be when the virus inside you is most detectable and transmissible; we built an entire edifice of testing and isolation on that foundation.

Experts aren’t sure why delayed positives are happening; it’s likely that population immunity, viral mutations, and human behavior all have some role. Regardless, the virus is “acting differently from a symptom perspective for sure,” says Emily Martin, an infectious-disease epidemiologist at the University of Michigan. That’s worth paying attention to. The start of symptoms has always been a bit of a two-step: Is it COVID, or not? If SARS-CoV-2 is re-choreographing its moves, we must too—or risk losing our footing.

Right now, experts are operating in a vacuum of evidence: “I don’t even know of any data that systematically evaluates this,” says Yonatan Grad, who’s studying the viral dynamics of SARS-CoV-2 at Harvard’s School of Public Health. But several phenomena could plausibly be muddying the testing timeline.

First, the immunity hypothesis, the most popular idea floated by the experts I spoke with. Perhaps symptoms are preceding test positivity, less because the virus is peaking late, and more because illness is arriving early, thanks to the lightning-fast reflexes of people’s primed immune systems. Sometimes, sickness is direct damage from a virus. But a runny nose, muscle and joint aches, chills, fevers, fatigue—which are common across many respiratory infections—can also be “signs that the immune system is being activated,” says Aubree Gordon, an infectious-disease epidemiologist at the University of Michigan. When the pandemic began, infections happened exclusively in people who’d never encountered the coronavirus before; illness took several days to manifest, as the virus churned itself into a frenzy and the immune system struggled to catch up. “Once people are vaccinated, though, their immune systems kick in right away,” says Emily Landon, an infectious-disease physician at the University of Chicago. (Prior infection, too, could have an impact.) If the body makes fast work of the invader, some people may never end up testing positive, especially on antigen tests. (PCRs are generally more sensitive.) Others may see positives a few days after symptoms start, as the virus briefly gains a foothold.

But some of the experts I spoke with were a little hesitant to give the immune system all the credit. Some unimmunized people have experienced early negativity, too, and many people who have gotten their shots still test positive before falling ill.

SARS-CoV-2 traits, too, could be flipping the sickness script, which brings us to the virus hypothesis. Any member of the Omicron cohort is “just a different beast,” says Ryan McNamara, a virologist at Massachusetts General Hospital. It struggles to penetrate deep into the lower airway, and may not accumulate to the densities that Delta did in the nose, which could make false negatives more likely. A couple of studies have also found that Omicron may, in some people, be detected in the mouth or throat before the nostrils.

In practice, “it’s really hard to separate if all of this is a property of the virus, or a property of the immune system, or both,” says Roby Bhattacharyya, an infectious-disease physician at Massachusetts General Hospital. Take Omicron’s symptom profile, for instance. This variant seems to more often prompt sneezier, head-cold-esque symptoms than those that came before it, and less often causes loss of taste and smell. And, on average, people infected in recent surges have been showing symptoms three days after exposure, far faster than the incubation period of five or six days that was the norm in the pandemic’s early days. But those patterns could be attributable to either the peculiarities of the Omicron clan, or how much more immune the average Omicron host is.

And testing and sickness severity involve “just so many variables,” says Ali Ellebedy, an immunologist at Washington University in St. Louis, similar to the diversity in reactions to vaccines—some people feel side effects, others don’t—or virus exposures. Some people never get infected, even after spending days with infectious people, while others seem ultra-susceptible. People’s vaccination status, age, genetics, even the dose of virus, can affect if, when, or how they feel ill, and whether their infection registers on a test.

Symptoms and test accuracy are also both subject to human bias. People can’t always remember when they started feeling sick. And user error can muddle diagnostics. “Are people really doing 15 seconds in each nostril, and really scrubbing each time?” Landon said. Even super-sensitive tests will miss the virus some of the time. A good portion of the specimens taken from sick people in medical settings “come back negative for everything,” Martin told me.

Coinfections are also possible: People who feel sick and test positive “late” may have actually caught something else first, only to develop COVID later on. “There are easily five to eight other viruses circulating right now,” says Melissa Miller, a clinical microbiologist at the University of North Carolina at Chapel Hill. It’s also allergy season in much of the U.S. And coronavirus tests can spit out false positives—though they’re unlikely to do so several days in a row.

Without more data, it’s hard to know how to best address early negatives. For more clarity, “you’d probably need a human challenge experiment,” in which vaccinated and unvaccinated volunteers are deliberately infected with SARS-CoV-2, then repeatedly tested and monitored for symptoms over time, Gordon said. But based on the stories emerging, the issue doesn’t really seem rare. “I think I’ve had maybe 20 friends in the last five weeks” catch the virus, Jesse Chen, a vaccinated 27-year-old in New York, told me; most of them experienced delayed positivity, including Chen herself.

If this is happening more, then “you cannot trust a negative rapid test at the beginning of illness,” Landon told me. And while the CDC and test makers have long said that negative results can’t rule out a SARS-CoV-2 infection, it’s not clear how these early-illness testing issues fit into diagnostic guidance. Kimberly Modory, a spokesperson for Abbott, which makes the widely used BinaxNOW SARS-CoV-2 test, wrote in an email that “people should continue to follow our test instructions, which is to test twice over 3 days, at least 24 hours (and no more than 48 hours) apart.” Another popular brand, iHealth, offers similar instructions. (CDC and iHealth did not respond to a request for comment.)

Many of these protocols, though, were developed when far fewer people had been vaccinated or infected, and Omicron and its offshoots weren’t yet dominant. And they have missed some recent infections. Furmaniuk, for instance, green-lit herself for dinner with two iHealth tests. And Ellen Krakow, a triply-vaccinated 58-year-old from Long Island, didn’t get a positive antigen result until she took her third BinaxNOW, on her fourth day of symptoms. (A PCR test, collected on her third day, did turn positive first.)

Until experts know more, several researchers recommended that people test with caution. Positives are still reliable, Landon told me. But people who are symptomatic and recently exposed might have good reason to be skeptical of negatives. “If you’re turning symptomatic, assume you’re infectious,” Grad told me—with something, even if it turns out not to be SARS-CoV-2. “People forget that, baked into all the recommendations, is that as soon as you turn symptomatic, you’re supposed to be behaving differently,” Martin told me. “A negative test shouldn’t be a pass to go out.”

Landon also raised concerns about the implications of early negatives for the duration of isolation. In January, the CDC cut recommended COVID isolation time in half, saying that people could stop sequestering themselves after just five days, counting out from when their symptoms started, as long as they masked for the five days following. The agency justified its decision by noting that a majority of people were no longer infectious by that point—but used data that almost entirely predated the Omicron surge.

Based on the evidence that’s emerged since, “five days is ridiculous optimism,” said Landon, who recently ran a study showing that a large fraction of people continue to test positive after their fifth isolation day, raising the possibility that they’re still shedding the virus in gobs. Rebecca Ennen, a vaccinated 39-year-old in D.C., didn’t even get her first positive result until Day Six of her illness, as her symptoms were on their way out. “It was just bizarre,” she told me. “I was on the mend.” So Ennen continued cloistering for another five days, until she finally tested negative again. Others, including Gordon and Furmaniuk, have also waited to test out of isolation; it’s what Bhattacharyya “would do too, if it were me.”

But such a mindset hardly seems sustainable, especially for people who are frequently exposed to respiratory microbes, including parents of very young kids, or who have bad allergies, or who don’t have rapid tests to spare. Isolation still takes people away from work (and income), school, and their families. It’s also emotionally harrowing. Krakow, of Long Island, didn’t test negative again until 14 days into her illness. “I was isolating for beyond two weeks,” she told me.

All of this means that our guidelines and perceptions of the virus may soon need to adjust—likely not for the last time. Butler-Wu, the USC clinical microbiologist, recently advised a friend who had received more than half a dozen negative test results—antigen and PCR—that her respiratory illness probably wasn’t COVID. The friend ended up visiting Butler-Wu, only to test positive shortly thereafter. “That really shook me,” Butler-Wu told me. “It flew in the face of everything I knew from before.” It was a reminder, she said, that the pandemic is still serving up plot twists. “I do this professionally,” she said, “and I still made the wrong call.”