Insensitive and Inefficient The Illusion of Rapid Antigen Tests

Germany is hoping that rapid antigen tests will help navigate the country out of the third coronavirus wave currently gripping the country. But they aren't up to the task. Other strategies could prove more effective.
A still from a video advertising the efficacy of rapid antigen tests in schools

A still from a video advertising the efficacy of rapid antigen tests in schools

Foto: Augsburger Puppenkiste / youtube

A middle-aged couple is leaning against a high table outside a gardening store in the Charlottenburg neighborhood of Berlin, both of them making strange, throaty noises. "Should we ask someone?" The woman seems annoyed.

She and her husband are there to buy a few plants, but before they can begin examining the greenery on this spring weekend, they must first take a rapid antigen test to see if they are infected with SARS-CoV-2. The test is from a company called Hygisun, available for 5 euros to those who can't otherwise present a certified corona test at the entrance.

The instruction leaflet, plastic funnels and pipettes are piled on the table in front of the couple. On the instructions, printed in tiny font, is a picture of a red, wide-open mouth. Beneath the image, it says: "Cough deeply. Make a noise like 'kuuua.'" And then: "Collect 2 ml of saliva from deep in your throat."

"Kuuua? What's that?" the man asks.

Surrounding the couple are other perplexed, coughing and spitting customers, all gathered at the entrance to the gardening store. The small, plastic test kit is their ticket for entry – their hope for a bit of gardening diversion on another shutdown-weekend in Germany.

People across the country currently find themselves confronted with similar riddles. Schoolchildren are shoving cotton swabs up their noses, daycares are handing out test "lollipops," while factories are asking their workers to take swabs of their mucous membrane before their shift. In professional testing centers, medical workers are performing tests on those who want one.


The article you are reading originally appeared in German in issue 16/2021 (April 17, 2021) of DER SPIEGEL.

SPIEGEL International

There is great hope that the rapid tests might allow the country to return to some semblance of normalcy before vaccinations make significant headway. The hope is that the test can help ward off one of the most dangerous aspects of SARS-CoV-2: that people might unknowingly have contracted the virus and begin spreading it to others. The rapid tests promise that they can identify infections within 15 minutes – thus putting the brakes on the pandemic.

Numerous German state governments are relying on the small, plastic kits. And the federal government's strategy for reopening the country depends heavily on testing, hoping to keep schools open up to an incidence rate of 200. In most states, there is already a de facto testing requirement for schools and many companies are planning on testing more frequently in the future as well.

Yet the antigen tests aren't actually suitable for screening in schools, companies or the population at large. It turns out that they are actually quite bad at fulfilling the hopes that have been pinned on them. On average, they are only able to identify 58 percent of asymptomatic infections, according to a review performed by Cochrane Collaboration, which analyzed studies focusing on the most common rapid antigen tests on the market. In the tests they considered, the samples were taken professionally. And tests on adults were among them.

Among children, the shortcomings are even more pronounced. Around three-quarters of infected schoolchildren aged 14 and younger were missed by the twice-a-week tests, according to an estimate resulting from a third iteration of a monitoring study in Austrian schools, as Michael Wagner, a microbiologist at the University of Vienna and co-head of the study, told DER SPIEGEL two weeks ago. Wagner estimates that "around 40 percent of them are infectious."

Berlin-based virologist Christian Drosten also warned this week of the tests' deficiencies, saying on his Tuesday podcast, aired by the German public broadcaster NDR, that "three of eight infectious days are overlooked."

Thorsten Lehr, of Saarland University, agrees, saying: "Rapid tests will certainly not lead us out of the pandemic." As such, the researcher says, they should not be seen as "the key to more freedom."

Virologist Oliver Keppler, from Ludwig Maximilian University of Munich (LMU), is even more dismissive of the idea of using rapid antigen tests as an effective tool in the fight against the pandemic. "It's lunacy," he says, "and in no way evidence based." The tests, Keppler says, "were propagated with lots of optimistic rhetoric." But there is a deep divide between the hopes that have been invested in them and the reality they deliver.

That can be seen in Austria. On Feb. 8, schools and shops in the country were opened, with a rapid test requirement being introduced in some economic sectors. "Nevertheless, the incidence rate has continuously risen," says Michael Wagner. Oliver Keppler adds: "If you want to be really cheeky, you could even ask if the incidence has risen precisely because of these tests – because everyone feels so safe."

Nevertheless, rapid testing is now being expanded to companies in Germany. Workers must be offered at least one rapid antigen test per week. Numerous models for opening up ("test and shop") also depend heavily on the antigen tests. Many experts, though, believe such models are destined to fail for as long as the case numbers remain as high as they currently are.

The Robert Koch Institute (RKI), Germany's leading institution for disease control, issued a warning early on about the unreliability of the tests, reporting in a January issue of its weekly Epidemiological Bulletin that a rapid test used frequently in Germany only identified 39 percent of asymptomatic infections at the emergency room of a large hospital. Oliver Keppler and his colleagues within the research network B-FAST have composed position papers, outlined concepts and sent emails. "It's not as if we scientists haven't informed policymakers of this problem," Keppler says. "But at some point, apparently nobody wanted to listen to us anymore."

The problems with the rapid tests begin with their approval. Test manufacturers do not have to prove, for instance, that their products are also able to detect variants. And if the tests are only able to identify a single part of the virus, "they would no longer be able to identify an infection" in the case of a mutation, "despite the spread of a new variant," says Wagner. "We would suddenly be blind without knowing it."

Beyond that, the manufacturer's own specifications alone are sufficient for approval. "And they sometimes tell tall tales about the sensitivity of their product," says Keppler, "which they frequently just sell as intermediaries." Claims, for example, that they identify far more than 90 percent of infections. In reality, as shown by the Cochrane study, most tests don't even meet the standard set by the World Health Organization (WHO), which demands that the tests identify at least eight out of 10 infections.

Famous German television personality Günther Jauch experienced last week just how easily rapid antigen tests can miss a SARS-CoV-2 infection, even in people exhibiting symptoms. Jauch was already suffering from a headache and achy joints when his rapid test showed that he was negative. But he felt even worse the next day, and a PCR test finally gave him his diagnosis: COVID-19.

Placing too much trust in the rapid tests could even result in superspreader events. If federal and state governments place their trust in the tests, the population is likely to trust them too. And how are schoolchildren who tested negative at school supposed to know that they could still be carrying the virus?

Not only that, but the rapid tests frequently indicate an infection when there isn't one. If one in every 200 people is infected with SARS-CoV-2, then – according to calculations by the Cochrane Collaboration for a specific rapid test brand – that would mean that almost three-quarters of the positive test results among asymptomatic people were wrong. That value will become even worse when case numbers are lower.

According to a recent report in the Guardian, the British government is thus considering a radical reduction in the huge number of rapid tests that are currently being performed as part of the country's opening.

The problem with the high rate of false positives will become particularly apparent in schools. If 11 million students and teachers are regularly tested, then tens of thousands of girls, boys and teachers who have no infection whatsoever will be sent into quarantine until they are able to produce a negative PCR test two days later.

And what about a child whose morning test at school shows an infection? "Is the child supposed to raise their hand and say, teacher, there's something here?" wonders Wagner, whose study involves PCR tests being performed on students and teachers. "Or is there too much pressure to test negative?"

"The problem with the rapid tests in schools is primarily how they are carried out," says Eva Rehfuess, professor for public health at LMU. Her daughter, the researcher says, has told her something that many other parents have no doubt heard as well: "When children take samples from their noses, many often start sneezing." Rehfuess says she wonders if it is really a good idea for schoolchildren to all test together in the classroom. "From my perspective: No!"

Researchers are also concerned that if the samples aren't treated with the necessary care, then the tests themselves could provide the virus with an excellent opportunity to find a new host – if, for example, the swabs aren't immediately disposed of in the proper way. In such instances, the testing events, which are actually supposed to slow down the pandemic, could turn into superspreader events.

And the quality of the results depends strongly on the quality of the samples taken, says virologist Keppler, something he sees every day in the clinic. "We get completely different results depending on whether nurse Raimund or nurse Hildegard took the samples – and they are both competent professionals." Keppler says he simply doesn't understand how this job can be completely turned over to laypersons. With such an approach, he says, "we should completely forget about required mass testing with rapid antigen tests.

A famous puppet theater in Augsburg even made a video illustrating how to carry out rapid tests

A famous puppet theater in Augsburg even made a video illustrating how to carry out rapid tests

Foto: Augsburger Puppenkiste / youtube

Still, it remains clear that coronavirus testing programs are necessary to slow the resurgence of case numbers following the lockdown. But to ensure that such programs are effective, they have to be designed completely differently than is currently the case in Germany. "With the rapid tests, we are only able to catch the big fish out of the lake that we are particularly afraid of when we go swimming," says Michael Wagner from the University of Vienna. "The small ones stay in the lake and keep growing, so we have to keep catching them. With PCR tests, we catch all of them, and can go swimming with no worries."

That means that the following things are necessary for a good testing program:

  • Highly sensitive testing procedures. PCR tests, for example, are more than 1,000 times more sensitive than antigen tests, which means that infected people can be identified and isolated before they can infect others.

  • Systematic, regularly performed testing. Instead of tests performed by people who don't know what they are doing in front of the gardening store, a system of regular tests performed several times a week in schools and companies should be introduced. That, too, would contribute to identifying infections early on. It would also allow for the discovery of infection clusters in families that might not otherwise test themselves.

  • Easy testing. Tests should be painless, using gargling, for example, to ensure that people are more accepting of the practice.

  • Contact tracing in the case of positive test results. It's the only way to prevent further infections.

A broad consensus has developed in Germany and Austria for how mass testing might look, including cleverly designed PCR testing regimes and innovative new ideas for mass screening. "The rapid tests were good as a kind of transitory technology," says Wagner. Together with colleagues from other universities, he has developed a concept for screenings at schools and presented it to the Austrian government. It calls for gargling PCR tests that can be performed at home. "It involves gargling a salt solution for half a minute before breakfast and bringing the sample to school, where it will be collected. The whole thing is then performed three times a week," he says. A few model schools have been chosen to test the concept and they are to start following the current lockdown.

A large monitoring program has now also started in Vienna after months of preparation. Since the end of March, everyone living or working in the Austrian capital has had the possibility to be tested at no cost with a PCR gargle test. Each week, 1.2 million test kits for the collection of samples are handed out. The gargle samples can be handed in at any REWE supermarket. The results are available within 24 hours.

Model calculations show that if 150,000 Viennese households hand in samples each day – capacity is 200,000 per day – then it should be enough to either prevent a new rise in case numbers or to significantly slow that rise. It is apparently especially effective if all households with children participate.

Johannes Zuber, from the Research Institute of Molecular Pathology at the University of Vienna, and Julius Brennecke, from the Institute of Molecular Biotechnology of the Austrian Academy of Sciences, were both instrumental in designing the program. "We do it alongside our research projects," Zuber says, adding that they are convinced "that with such a monitoring program, it could really be possible to control the pandemic, if people participate. This one, and the next."

At the Vienna Biocenter, they claim, the reliance on a gargle test program among the 1,500 employees has resulted in 93 positive cases, but not a single transmission on campus. But they have much larger aspirations: that of testing all of Austria, all of Germany or even all of Europe twice a week for SARS-CoV-2 using PCR tests.

"We got the idea from the northern Italian village of Vo," says Brennecke. In February 2020, the town, with a population of 3,400 people, was one of the very first in Europe to be sealed off due to the novel coronavirus. But in contrast to the rest of the country, all Vo residents were quickly tested using PCR tests and those who tested positive, even absent symptoms, were isolated. That enabled Vo to get the virus under control.

"It is a potentially excellent technology because it is quick, sensitive and scalable."

Oliver Keppler, virologist from Ludwig Maximilian University of Munich

The problem in Germany doesn't just come from political reliance on the less sensitive antigen tests on the path to opening back up, Zuber believes. Rather, he says, the opportunities presented by large-scale corona monitoring programs are "talked to death," because people only see the potential problems with the programs.

For example, when it comes to test capacity. Many ideas are, in fact, rejected in Germany with the argument that there is insufficient PCR testing capacity. But if several samples are analyzed together, a practice known as "pooling," with individual analyses only being carried out if the pool tests positive, then test capacity immediately becomes several times greater – with no significant sacrifice of sensitivity.

And there is also a third testing method, the LAMP test, which is almost as sensitive as the PCR test and vastly more sensitive than antigen tests. The procedure has the significant advantage that large testing capacities can be developed relatively quickly and affordably.

"It is a potentially excellent technology because it is quick, sensitive and scalable," says Oliver Keppler. The procedure doesn't yet have a CE marking – indicating approval in the EU – and isn't yet commercially available for that reason. But it has proven its efficacy in research projects, such as the Virus Finder Study in Heidelberg. In that analysis, more than 17,000 test kits for gargle samples were distributed to randomly selected households in the region. Participants sent their gargle samples back to the researchers, who then examined them for the presence of the SARS-CoV-2 virus using the LAMP test.

The goal was to arrive at a more precise assessment of the number of symptomatic and asymptomatic infections in order to react more quickly to the possibility of hospitals being overrun. Around 0.4 percent of the samples were positive.

"What we definitely showed, is that such a large-scale testing project is logistically workable," says project leader Andreas Deckert. The LAMP method, he says, proved itself. He believes that if samples are pooled, the technique could surely be used to develop greater testing capacity.

"It is important to test as many people as possible as often as possible," says Deckert's colleague Simon Anders. "And to do that, we have to make testing as simple and as comfortable as possible: gargling instead of swabs, mail-in tests or self-tests instead of having to go to testing centers."

"It sometimes seems to me that we are a bit afraid of creative approaches," Deckert says. To motivate people to get tested, he suggests, all those tested could perhaps be included in a kind of lottery, for example.

Michael Wagner from Vienna estimates that it would take metropolitan regions around three months to develop the infrastructure for pooled PCR tests. His Munich-based colleague Keppler believes that the opportunity to do so could soon present itself. "It could very well be that we will have to go into lockdown soon. Then we would have a few weeks to get PCR pool testing under way."

Note: In an earlier version of this article, there was an infographic on the topic of rapid tests that could lead to misunderstandings due to different data sources. We have therefore removed this graphic.