At first blush, you might reasonably dismiss the rapid, saliva-based antigen assay for coronavirus. What possible purpose can a test serve in a pandemic if it misses, say, 30 of every 100 infections it encounters, leaving those cases undetected, untreated, and uncontained? What good is a test that creates that much false reassurance and causes exactly the wrong people to relax their adherence to social distancing? Who would even think of using such a test in a hospital setting, where a false negative result might imperil the safety of health care workers?
You wouldn’t be alone. The failure of the antigen-based assay to match the diagnostic accuracy of the gold-standard test you are probably familiar with—the laboratory based, reverse-transcription polymerase chain reaction (PCR) nasal (or nasopharyngeal) swab—is why the Food and Drug Administration (FDA) has been reluctant to approve it for sale in the U.S.
But what if we re-framed the narrative? What if we instead asked what possible prevention purpose a PCR test can serve in a pandemic if it returns positive results in people who have already cleared the virus and pose no risk of further transmission? What good is a PCR test that identifies non-infectious individuals as candidates for isolation and quarantine? Who would even think of using PCR for outbreak containment if it sends up so many false alarms, leading contact tracers on so many wild goose chases, and undermining public confidence in the role of testing to keep us safe? The rapid, saliva-based antigen test has gotten a bad rap. It is the better, preferred tool for outbreak control. Here’s why:
Infected ≠ Infectious
When you go to the grocery store, eat in a restaurant or play pick-up basketball, the person you are trying to avoid is the person who might transmit the virus to you during that activity. While all spreaders are infected with the SARS-CoV-2 virus, not all persons infected with SARS-CoV-2 are spreaders. Not everyone who has detectable levels of SARS-CoV-2 virus in their airways is shedding virus in sufficient quantity to pose a transmission risk to others. When it comes to preventing the spread of the virus and containing outbreaks, it’s the infectious people you want to identify and isolate.
Testing To Diagnose ≠ Testing To Contain An Outbreak
Suppose you go to the hospital because you are feeling sick, with headache or fever or difficulty breathing. Your caregiver is going to want to investigate the cause of your symptoms. A test for the presence of SARS-CoV-2 infection will help them choose the necessary level of personal protective equipment and cohorting to keep other health care workers and your fellow patients safe. This information will also help them to understand your prognosis, what treatments you should receive, and whether they should continue to pursue an alternative diagnosis. They will conduct a test to determine if you are infected. That’s called “diagnostic testing” and it serves an indispensable clinical function in individual patient care.
But diagnostic testing is not the only testing we do in public health and medicine. Suppose, instead, that you are a student at a residential college. Or perhaps you work as a custodian in a nursing home. Maybe you reside in a high-rise apartment building. Or perhaps you are just wondering if it is safe to play soccer or visit an elderly relative. It doesn’t really matter that you feel no symptoms of illness, because 40 percent of people who are infected are asymptomatic. You and the people living, loving, and working in close proximity to you might reasonably wonder what transmission threat you pose, even in your asymptomatic state. Perhaps you are recently exposed and even though you haven’t yet incubated the infection to the point of symptoms, you actually are infectious. Or maybe you are a “silent superspreader”: infected and infectious but asymptomatic.
Testing to confirm that you pose no risk of inadvertent transmission is called “surveillance testing.” It serves a critical function in containing outbreaks, especially since the other tools at our disposal for this purpose (e.g., masks, hand-washing, social distancing, and quarantine) are all limited in number, adherence, and effectiveness.
When evaluating whether a test for SARS-CoV-2 is up to the task, the first things we need to do is to specify carefully what exactly that task is: Because it turns out that the PCR-based nasal swab your caregiver uses in the hospital does a great job determining if you are infected but it does a rotten job of zooming in on whether you are infectious. By contrast, the rapid saliva-based antigen test with the 30 percent false negative rate does a poor job of diagnosing infection, but it is likely the better tool for judging infectiousness.
Antigen Testing Is Maximally Sensitive When People Are Maximally Infectious
It’s a gross (but useful) simplification: infectiousness depends on “viral burden.” The more virus you have in your airways, we believe the more likely you are to transmit to others. Viral burden rises steeply in the days immediately following exposure. The typical infected person begins to shed large, transmissible quantities of virus starting (on average) around Day 3 after exposure. Symptoms, if they emerge, will follow around Day 5. By Day 8, whether or not symptoms have emerged, viral burden and infectiousness are both in steady decline. All these numbers will vary from person to person, of course, but the general lesson is this: there is a roughly 5-day window of maximum transmissibility, stretching from Day 3 to Day 8 following exposure.
Another useful simplification: viral burden is also associated with the likelihood that an antigen-based test will pick up the presence of infection. The more virus in the airway, the greater the likelihood of a positive antigen result. What that means is that— conveniently—there is a high degree of overlap between the period of maximal infectiousness and the period during which the antigen-based test is most likely to yield positive results. And, just as importantly, antigen-based testing is highly specific: unlike its PCR counterpart, the antigen test isn’t fooled by faint signals of lingering viral matter outside the period of infectiousness.
For purposes of surveillance screening, those antigen-based negatives worrying the FDA aren’t false negatives at all; those are true negatives for disease transmission. Far from being problematic, in the context of outbreak containment, the antigen test’s limited window of sensitivity is a major asset. The antigen test is ideally suited to yield positive results precisely when the infected individual is maximally infectious.
Contrast that with PCR, whose performance barely depends at all on viral load. In principle, PCR can detect just a few copies of viral RNA. PCR starts to pick up the scent on Day 2 and continues to return positive results for as many as 6-12 weeks after exposure. That’s great if you’re interested in knowing if someone was infected in the past several weeks to months; but it’s a problem if you are interested in knowing if someone is infectious right now. The PCR continues to produce positive results long after a person has ceased to pose any real risk of transmission to others, some data say up to 12 weeks longer!
As a test of infectiousness, the PCR test is far too prone to false positives. These false positives clog up the testing and contact tracing infrastructure and needlessly ground a lot of people who pose no transmission risk to others. PCR is the wrong tool for the surveillance job.
There are a thousand little details that our clinical colleagues would add to this narrative. But the main message is this: For purposes of containing outbreaks, the problem is decidedly not false negatives associated with antigen testing; rather, the problem is false positives associated with PCR testing. If your goal is to suss out infectiousness and prevent outbreaks, the antigen test is the tool you want to reach for first.
There is an important downside. Because the period of infectiousness only lasts about 5 days, antigen tests need to be conducted again and again. A negative antigen result received 10 days ago tells you virtually nothing about your infectiousness today. That’s why we and others have been recommending screening every 2-3 days on college campuses. Fortunately, other attributes of saliva-based antigen testing make this feasible: unlike PCR, it doesn’t require trained personnel, specific chemical supplies, or centralized, expensive laboratory instrument; it is less vulnerable to the supply chain bottlenecks that plague PCR-based labs; it returns results in a matter of hours; it costs as little as $1 per test; and it could be as convenient as sucking on a piece of
Rapid, saliva-based antigen testing is an essential weapon in the fight to resume many of the activities and reopen many of the venues that comprise what we used to call “normal life.” It is practical, convenient, cost-effective, and easily scaled. It is time to stop asking what possible use we might have for a test with a 30 percent false negative rate. Instead, it is time to start asking why it is taking the FDA so long to approve this essential prevention technology.