Antigen tests could change Covid-19 screening culture

Interrupting transmission of the virus is the goal for Becton Dickinson.

Trial Results

Becton Dickinson, armed with trial data suggesting that its Veritor antigen test is better at identifying infectious Covid-19 patients than the gold standard of PCR, has begun clinical trials in the next obvious setting: people who show no signs of the disease, but do in fact carry the coronavirus. 

With more contagious variants of the virus circulating, using cheap and simple antigen tests to screen whole populations could be a crucial way of swiftly identifying and quarantining local outbreaks. This is particularly useful if they can identify individuals at a high risk of passing the virus on.

BD’s recent trial data showed that the Veritor test, which was authorised by the FDA in July, had a higher positive predictive value than Quidel’s Lyra PCR test – 90% versus 74% respectively – when used to test 251 people with symptoms consistent with Covid-19. 

But this raises an obvious question: if PCR testing for viral RNA is the yardstick against which all antigen and antibody test are measured, how can BD’s test have beaten it? In fact, the trial used an even more rigorous control: viral cell culture. This is even more expensive and time-consuming than PCR testing, so much so that it is functionally useless as a diagnostic for Covid-19. But a positive result means a person likely was, at the time the sample was taken, highly contagious. 

Accuracy of assays compared with viral cell culture
Performance values BD's Veritor antigen test Quidel's Lyra PCR test
Positive percentage agreement (%)  96.4 100 
Negative percentage agreement (%)  98.7 95.5
Positive predictive value (%) 90.0 73.7 
Negative predictive value (%)  99.5


Overall percentage agreement (%)  98.4 96.0 
Culture positive, test positive 27  28 
Culture negative, test positive 10 
Culture positive, test negative
Culture negative, test negative 220  213 
Source: Clinical Infectious Diseases.

In the trial, 28 samples were positive by cell culture, and Veritor detected 27 of them. The Lyra PCR test, however, reported 38 positives, implying that it had identified 10 non-infectious individuals. The greater overlap with Veritor suggests that antigen tests could be better than molecular assays at pinpointing those who are likely to be shedding active virus, with none of the drawbacks of viral culture. 

Antigen tests might, therefore, have continued utility even as vaccines roll out. Astrazeneca in particular is looking at transmission rates among vaccinated people. 

Chuck Cooper, BD’s vice president of medical affairs for its diagnostics unit, says that the central value of antigen testing is its swiftness. 

“That rapid result allows you to apply infection control measures such as quarantine and contact tracing very quickly, rather than having a delay,” he tells Evaluate Vantage. “Particularly in this pandemic, where you have a virus that is highly infectious, and people are infectious even before they develop symptoms, having that time-to-result window is really important for public health measures to try and interrupt transmission in the community.”

Next step

But antigen tests are only going to be useful as a way of screening entire communities if they are also accurate in patients with no symptoms of the disease – some of whom will in fact have the virus, and will be passing it on to others. 

Last year evidence emerged that Abbott’s PanBio antigen test could work in these patients. Subsequently, a CDC trial indicated that BinaxNow, Abbott’s other antigen test, missed two thirds of asymptomatic cases. 

BD, meanwhile, is going ahead with Veritor.

“It’s quite possible that there are different performance levels for different assays in asymptomatic populations. For that reason, we’re conducting some very rigorous clinical studies to understand how our assay performs in that patient population,” Mr Cooper says.

It will take some time for these data to emerge. The challenge with asymptomatic clinical trials, Mr Cooper explains, is that the prevalence of infection is so low – much lower than the 5-20% positivity rate seen in symptomatic trials. That said, Mr Cooper says BD has managed to enrol “a significant number of people who are positive for Covid-19 despite being asymptomatic”.

Until these trials report, the utility of Veritor as a screen remains unproven. But BD deserves credit for seeking to find out. 

Share This Article