A version of this article first appeared in Katelyn Jetelina’s newsletter, Your Local Epidemiologist.
After two long years, antigen tests are finally available to Americans free of charge. On Tuesday, the system secretly went live. More than 900,000 got word and flooded the site in a few hours. Each household can order up to four tests on covidtests.gov. Or you can go on the U.S. Postal Service website and fill out a surprisingly easy government form (oxymoron, I know). You will not be charged for shipping, and they don’t even ask for a credit card number. The tests should show up in your mailbox seven to 12 days after you order. (There are a number of health equity issues with all of this, but that’s for another post.)
And we really need to better leverage antigen tests in our pandemic toolkit. They are one of the most underutilized tools and can help us “learn to live” with SARS-CoV-2 by breaking transmission chains across the nation.
In my last antigen test post, I presented evidence that antigen tests could physically detect omicron (thanks to the N-protein) and that lab studies could detect omicron at the same viral levels as delta. But lab studies done in highly controlled environments are certainly not reflective of the “real world.” We desperately needed evidence in the community. Now we have five studies and each has a unique contribution to the puzzle we’re trying to put together: How well do antigen tests work?
Here is what they say.
Be Careful With Negatives in the Beginning of Infection
The first study followed 30 people in high-risk jobs from Dec. 1 to Dec. 31, 2021, during omicron outbreaks at five workplaces in New York, Los Angeles, and San Francisco. Everyone was fully vaccinated (only some were boosted) and was being tested daily. The scientists looked at how well nasal antigen tests worked compared with saliva PCR tests. They found a few interesting patterns:
• The average time from first positive PCR to first antigen positive was three days.
• Peak viral load in saliva was one to two days earlier than in the nose.
• All individuals developed symptoms within two days of their first positive saliva PCR test.
• It’s possible to be contagious yet have a negative rapid test. Four of the 30 people in this study spread COVID-19 between negative rapid tests.
It’s important to note that this study primarily compared nasal antigen tests with saliva PCR tests. So are these findings because antigen tests aren’t as great in the beginning of infection or because of the location that was swabbed? It seems like the location of the swab is a big part of the answer. Five people in this study also had a nasal PCR. The saliva PCRs consistently had stronger signals (lower Ct values) than the nasal PCRs in the first few days of infection.
This study tells us that we need to be super careful when using rapid tests in the first few days of exposure or infection. To get the most from your rapid test, wait at least 48 hours after symptoms and five days after exposure before taking an antigen test. If you’re negative, test again 24 hours later. You can certainly test sooner, but any negative results will be unreliable. A positive antigen test result, on the other hand, is very reliable right now, especially after exposure or with symptoms.
(For the record, the rapid testing website FAQ section advises testing five days after close contact or as soon as you begin feeling symptoms. We think it’s better to wait a bit longer after symptoms.)
Antigen Tests Work Really Well Thereafter
The second “real-world” study compared nasal antigen tests with nasal PCR tests among 731 people in San Francisco from Jan. 3 to Jan. 4. The scientists evaluated how well the BinaxNOW rapid antigen tests performed compared with PCR at a community-based testing site. What did they find?
• Overall, 296 of 731 (40.5 percent) tested positive on the PCR.
• Among people with a positive PCR with no symptoms (59 people), antigen tests detected 43 positives. This equated to a sensitivity of 89.8 percent.
• Among people with a positive PCR who were symptomatic, the sensitivity of antigen tests was 97.6 percent.
• As expected, the higher the viral load, the more accurate the antigen test.
• The tests were just as accurate among those under 13 years of age compared with those older than 13.
• The tests performed similarly regardless of vaccination status.
• The likelihood of a false positive was very small (432 of 435 PCR negatives were correctly identified as negative by rapid tests).
This preprint study was updated a few days ago. The researchers added data from a pilot study of cheek swabs among 75 people, both by PCR and by rapid antigen test. This is what they found:
• Cheek swabs performed far worse than nasal swabs.
• Among 46 PCR positives, 22 were positive by antigen in the nose and only two were positive from the cheek. These results are disappointing, yet important to know.
In all, this study reinforces the notion that rapid tests are failing in the first few days not because of technical issues, but because omicron is changing disease biology. It seems there is now a lag between symptoms, contagiousness, and viral explosion in the nose. When the virus does take hold in the nose, rapid tests can usually find it, even without symptoms.
You’re Infectious for Longer Than Five Days
The third study evaluated how omicron infection was (or was not) different from delta infection among NBA players—a highly vaccinated population that is tested daily. Specifically, scientists evaluated 97 tests confirmed from the omicron variant and 107 from the delta variant and compared how the tests varied on myriad factors (viral load, length of infection, etc.). Samples were all collected using dual swabs (nose and throat) and evaluated with PCR. What did they find?
• Omicron infection lasted, on average, 10 days. This is comparable to delta infections, which lasted, on average, 11 days.
• Many people cleared the virus faster and others clear the virus much slower.
• Roughly 50 percent of people still had high viral levels at day five (meaning they were likely infectious).
Together the results suggest a broad range of an infectious period, which is all the more reason to use a rapid antigen test-to-exit strategy in the United States.
The U.K. Agrees
A separate modeling study in the U.K. estimated the impact of leveraging antigen tests to end isolation (something the CDC is not recommending). Specifically, the scientists were interested in how many people would be infectious given different policy recommendations. What did they find?
• After a five-day isolation period (and not using an antigen test), there is a 1 in 3 (31 percent) chance you’re still infectious.
• After a seven-day isolation period (and not using an antigen test), there is a 1 in 6 chance you’re still infectious.
• If you use an antigen test on day seven of isolation and it’s negative, there is less than a 1 in 10 chance you’re still infectious. These are the same odds as if you isolated for 10 days without testing.
In England, the isolation policy was updated again this week: “It is now possible to end self-isolation after 5 full days if you have 2 negative LFD [antigen] tests taken on consecutive days. The first LFD test should not be taken before the fifth day after your symptoms started (or the day your test was taken if you did not have symptoms).”
Viral Load Isn’t the Same Thing as Infectious Load, Though
All the previous studies (except the U.K. report) analyzed “viral load,” or the number of virus particles, because it’s the easiest to measure for a quick turnaround study. However, and importantly, the number of viral particles does not equal the number of infectious particles. And the latter is what we are truly interested to answer: Are we infectious?
The final “real-world” study was conducted in Switzerland among 384 symptomatic individuals at a community testing center, who were tested during the first five days of symptoms. The goal was to examine the relationship between viral load (measured by PCR) and contagiousness (measured by lab experiments). What did they find?
• The precise viral level by PCR was not a great predictor of infectiousness: It was a modest 31 percent correlation. This is OK but certainly not fantastic.
• The number of infectious particles was lower among vaccinated people compared with unvaccinated.
• Omicron was not substantially different from delta, either in terms of viral load or contagiousness. In other words, we need to look elsewhere to understand omicron’s mysteriously high contagiousness.
This study did not include rapid antigen tests, but given that we are seeing considerable infectiousness at viral levels lower than what rapid tests pick up, it’s best to assume that a positive rapid test means you are contagious, no matter what testing day you’re on.
Bottom line: Use antigen tests. Use antigen tests. Use antigen tests. Do so wisely.
• Be aware of false negatives in the early stages of infection, and know that it can take several days after symptoms for the virus to take hold in your nose. Once you reach the tipping point, rapid tests are a reliable way to detect and monitor your infection.
• Trust your positive test during the omicron wave.
• If at all possible, do not leave isolation without testing. (I don’t care what the CDC says.) If you can’t access tests, assume you are contagious for 10 days, and act accordingly.