Skip to Main Content

Marianna Parker, a Boston-area pediatrician, wanted to make sure she, her husband, and her toddler didn’t have Covid-19 when they developed a nasty cold around Dec. 18. Over the course of a week she used five Covid rapid tests and took three PCR tests, testing negative. Her husband tested negative on a rapid test on his second day of infection but positive on his sixth, a result that was confirmed via PCR. “If he hadn’t decided to do that random test on day 6 we would never have known,” she said.

Nevertheless, she’s convinced all three of them had Covid, and that the tests simply produced the wrong results. In a Facebook post, she advised friends to remember that if they are vaccinated, their tests might only be positive for a short period of time.


Parker is not alone in puzzling over how to interpret the antigen tests. Social media platforms are rife with anecdotes from people who contracted Covid but who report that rapid tests came back negative. Meantime, a small, 30-patient study this week cast doubts about rapid Covid tests’ reliability early on after infection, showing that the tests took several days after infection before detecting the virus.

Now scientists are trying to determine why the tests are missing cases, many of them of the Omicron variant, and what could be done about it.

No one is arguing that antigen tests should not be used, nor that it is reasonable to expect they will detect every case. They are essential at a moment when getting a PCR often involves waiting an hour for a test and then several days for results. They will be important for identifying patients who should receive new drugs, such as Pfizer’s Paxlovid and Merck’s molnupiravir, that can keep patients out of the hospital if they are given soon after infection. But experts also worry that if they prove unreliable it could affect their usefulness in some situations, such as the common practice of using them before gatherings as an extra barrier against superspreader events or as protection in workplaces.


Growing evidence indicates Omicron collects in the throat, not in the nose, which could be why the nasal swabs used in rapid testing kits often aren’t detecting the variant.

“One of the things that we’re worried about is with all of this rapid antigen testing with nasal swabs, are we missing a lot of early infection?” said Donald Milton, a co-author of a December study showing that saliva tests may work better. “We know that people are transmitting during that period — minus three to plus two days of symptom onset. It’s when transmission happens.”

Abbott Laboratories, maker of the BinaxNOW test, said in a statement that it is well-known that PCR tests are more sensitive than antigen tests. Quidel, maker of the QuickVue test, said in a statement that performance against the new variant appears “similar” as with previous variants, and that the company will remain “vigilant” in evaluating its assays using genetic sequencing and real-world studies of virus samples.

Abbott said that in its continuing surveillance efforts, it has completed extensive testing on Omicron on a total of 74 clinical specimens. “In all cases, our studies confirm that BinaxNOW continues to detect the Omicron variant at comparable viral load levels as all other variants and the original SARS-CoV-2 strain,” the company said.

John Moore, a professor of microbiology and immunology at Weill Cornell Medical College, said there’s still reason for concern.

“Whether there is too little virus on nasal swabs or reduced assay sensitivity, the [rapid antigen tests] are often failing to detect Omicron during early symptomatic infection,” Moore said.

It’s unlikely, many experts say, that the false negatives are the result of mutations in the proteins of the virus, because that would likely affect different rapid antigen tests differently. More likely, any differences are the result of changes in how the virus behaves in the body.

The Food and Drug Administration has authorized antigen tests so long as they are 80% sensitive — meaning they return a false negative 20% or less of the time — and they are at least 98% specific, meaning that false positives are rare. That means that it is expected that false negatives will happen, and the test’s packaging recommends using two tests 24 to 36 hours apart, and getting a PCR test if there is any doubt.

In an interview, a top FDA official emphasized that there is some data indicating the sensitivity of the tests might decrease further against Omicron compared to previous variants, but more data are needed.

“If you are positive on these tests, then there’s a very high likelihood you really do have Covid,” said Jeffrey Shuren, the director of the Center for Devices and Radiological Health at the FDA. “If you’re negative, we’ve always said it’s presumptive.” That means a person who has been exposed or who has symptoms might want to get a PCR when an antigen test turns up negative.

Another uncomfortable fact, according to Shuren: It’s possible to transmit the virus to someone else when there is a negative antigen test. “Rapid antigen tests are not tests of infectiousness,” said Shuren.

All of that was true even before the emergence of Omicron. But Omicron could also change the functionality of the tests in various ways.

Researchers with the National Institutes of Health’s RADx program were able to obtain samples of the Omicron variant that were inactivated by heat. On those samples, the antigen tests performed well. Failure to detect the heat-inactivated virus would have been a red flag, Shuren said, but success does not guarantee the tests function as well in the real world as they have with other variants.

But the FDA made an announcement on Dec. 28 that the tests do detect the Omicron variant but “may have reduced sensitivity” based on “preliminary” data using live virus patient samples.

Other very preliminary data from NIH has shown that, in a small number of samples from Omicron patients, the amount of protein (what’s detected by the rapid antigen tests) relative to genetic material (what’s detected by PCR) may be lower compared to previous strains. But there are limitations to the technique used to determine this, including that protein may be present but bound up in ways that make it more difficult to detect.

That has left even physicians wondering exactly how to use the tests. “I’m afraid I can’t offer much more than a frustrated shoulder shrug here and assurances that I share your questions,” said Craig Spencer, an assistant professor of emergency medicine at the Columbia University Medical Center. “I still do rely on them before small social gatherings, but recognize they might not be the fail-safe we once thought they were.”

“Previously rapid antigen tests have been, OK, they’re less sensitive, but indeed, they catch individuals when they’re infectious, and that’s what we really want to know,” said Anne Wyllie, a Yale microbiologist who is one of the 30-person study’s authors. But, she said, it is “alarming” how much virus was seen in the study in people who tested negative using the antigen tests. “They are such high viral loads in saliva, you cannot tell me that that is not infectious.”

Using saliva samples or throat swabs might be one way around the problem. A December study from researchers in South Africa said that saliva swabs should be the “preferred sample” for PCR tests against the Omicron variant. That’s led to hopes that using throat swabs instead of nasal ones could lead to earlier detection with the rapid tests.

Sanjat Kanjilal, the associate medical director of clinical microbiology at the Brigham & Women’s Hospital, said his interpretation of the study, which has not yet been peer-reviewed, is that the throat approach might be better for Omicron. He said that there is a need for manufacturers and labs to start validating throat swabs.

But not everyone is so sure that will be a complete solution.

“I think it’s just something different about Omicron, and I don’t know how to predict it,” said Blythe Adamson, an epidemiologist and the lead author of the 30-person paper. “I don’t think it’s just time dependency. I don’t think it’s just viral load. And I don’t even think it’s just site of collection, saliva or nasal.”

Adamson is also an employee at Flatiron Health, an affiliate of Roche. Another Roche unit makes PCR tests. One of the co-authors of the paper is the CEO of a company developing saliva-based Covid tests.

Kanjilal said that he is no longer comfortable using the rapid tests as a strategy to ensure gatherings are safe, after his own experience at a family gathering where someone tested positive. Kanjilal himself did not test positive after repeated tests.

“I mean, look, Omicron is out of control,” Kanjilal said. “We’re using rapid tests a lot. It’s still out of control, right?”

The tests will help pick up people who need antiviral treatment. They might help decide when people are no longer contagious and can return to work. But will they catch people who will become superspreaders or make gatherings safe? “For me, I don’t think it’s enough. I wouldn’t feel comfortable with that strategy anymore.”

Megan Molteni contributed reporting.

Create a display name to comment

This name will appear with your comment

There was an error saving your display name. Please check and try again.