Asymptomatic COVID-19 testing reaches inflection point
U.S. policymakers and regulators usher in an era of more widespread and decentralized COVID-19 testing
As U.S. policymakers and regulators usher in an era of more widespread and decentralized COVID-19 testing, questions arise on how to gauge test performance and control.
With the pandemic raging, regulators are breaking down the distinction between symptomatic and asymptomatic COVID-19 testing, and opening the door to individual self-screening.
Communications from FDA and HHS, and the authorization of two home antigen tests this week, solidify the U.S.’s transition from a framework focused on clinical diagnosis of patients with symptoms to one that encourages asymptomatic screening, both on- and off-label.
The most obvious change came on Tuesday, when FDA issued its first emergency use authorization (EUA) for an over-the-counter, at-home COVID-19 antigen test.
Developed by Australian biotech Ellume Ltd. with support from NIH, the test will be sold without a prescription and can be used in any individual over the age of two, regardless of whether there is reason to suspect they might be infected.
A second home antigen test authorization followed Wednesday, this time for Abbott Laboratories (NYSE:ABT), whose manufacturing capacity for its BinaxNOW test is an order of magnitude larger than Ellume’s.
Like the first FDA-authorized at-home COVID-19 test from Lucira Health Inc., home use of BinaxNOW requires a prescription, and is only indicated for diagnostic testing of “individuals suspected of having COVID-19 by their healthcare provider.”
The more subtle shift is in the way officials are defining the phrase, which is used in the majority of EUAs for COVID-19 tests.
“Individuals suspected of having COVID-19 by their healthcare provider” is now explicitly stated to include asymptomatic individuals with recent exposure to a person who tested positive, healthcare workers, residents in congregate facilities and — at a provider’s discretion — those in areas of high community spread, according to Nov. 16 updates to FDA’s FAQ page for COVID-19 testing.
Officials are also endorsing off-label use of diagnostic tests for asymptomatic screening, citing data that suggests SARS-CoV-2 loads are similar in symptomatic and asymptomatic individuals. Only a small fraction of tests with EUA are authorized for screening, which requires conducting clinical studies in asymptomatic populations.
“The entire framework of asymptomatic versus symptomatic authorization is wrong. It does not obey the biology,” said Assistant Secretary for Health Brett Giroir in a press briefing Thursday.
Comments this week from Giroir and FDA’s Timothy Stenzel expressed support for broad use of diagnostic tests for asymptomatic screening, and cited CMS’s support as well.
The harmonization across government agencies is a stark contrast to the mixed messages sent in the summer, and a necessary step for widespread COVID-19 screening to go mainstream. While off-label use of tests was always permitted at a doctor’s discretion, the main change is in the government’s messaging about who should be tested.
This week also saw two public calls for more investment in asymptomatic testing: a Dec. 16 report from the Rockefeller Foundation and a Dec. 15 letter to Congress from 50 infectious disease physicians, epidemiologists, scientists, and community leaders that both called for government-supported mass asymptomatic screening campaigns, the latter with a focus on rapid, point-of-care antigen tests.
CMS, FDA and HHS have each voiced support for asymptomatic antigen testing, with CMS saying it would temporarily exercise enforcement discretion for the tests in congregate settings during the pandemic.
And at a virtual town hall meeting Wednesday, Stenzel said FDA believes it’s acceptable to use any authorized antigen test off-label for asymptomatic screening, based on data it has seen suggesting the tests perform adequately in people without symptoms. Stenzel is director of the Office of In Vitro Diagnostics and Radiological Health in FDA’s Center for Devices and Radiological Health (CDRH).
Antigen tests, which detect viral proteins, cost less than PCR tests for viral RNA, and are more readily deployable as rapid tests at the point-of-care or at-home.
But questions remain about whether these immunoassays, which are generally less sensitive than PCR tests, will continue to show sufficient accuracy when rolled out to the larger public, and whether the uncertainty around their results will be properly understood by users, particularly in at-home settings.
Another emerging issue for antigen tests is the distinction between their performance in individuals with low versus high viral loads. While the tests generally underperform in the former, proponents say what really matters is their ability to catch the latter, who are thought to be at highest risk of spreading infection.
A cautionary tale from the U.K. illustrates the risks of conducting mass asymptomatic screening programs with insufficiently validated antigen tests, including blowback about the discrepancy between a test's stated intended use for symptomatic individuals and its use in practice.
A rapid, point-of-care antigen test from Innova Medical Group Inc. deployed in the U.K. government's Operation Moonshot mass testing scheme was reported to have 96% sensitivity in its instructions for use document, but later shown to have an overall sensitivity 76.8%, and 57.5% when used by self-trained members of the public.
Though a Liverpool Mass Asymptomatic Serial Testing (MAST) trial using Innova's test did catch hundreds of infections that would have otherwise gone undetected, the lower than expected sensitivity raised alarm and undermined confidence in the testing program.
Beckman Coulter Inc. thinks its high-throughput COVID-19 antigen test, launched Friday at $4 per sample, threads the needle on accessibility and accuracy by providing a lab-based test that is cheaper and speedier than PCR, but more sensitive and scalable than point-of-care antigen tests. “I look at the role of this for serial mass testing programs,” said CMO Shamiram Feinglass.
The company’s first EUA application is for use in symptomatic patients. Feinglass said FDA’s announcement doesn’t change the company’s plans to wait until its asymptomatic trials are done to go after that market. “The way our clinical trials are being run, this is the order that were doing it.”
Degrees of DIY
Ellume and Abbott’s EUAs this week highlight the trade-offs between scalability and control for home tests, and ups the ante on communicating uncertainty around results.
BinaxNOW, which has the largest production capacity of any point-of-care COVID-19 antigen test, has been heavily deployed by the Trump Administration in government-supported testing programs. The company plans to deliver and administer 30 million at-home tests in 1Q21 and another 90 million in 2Q21.
In opting for a prescription path, facilitated by digital health partner eMed, Abbott faces less uncertainty about how individuals will interpret and act on test results than it would in an over-the-counter setting.
eMed determines an individual’s eligibility, facilitates test delivery to their home, guides the self-collection process and interpretation of results through a live video visit with a trained provider, reports results to public health agencies, and creates a digital results “certificate” in Abbott’s NAVICA app, which the company says could be used to facilitate access to in-person venues.
Ellume is targeting delivery of 20 million of its over-the-counter COVID-19 tests to the U.S. market in 1H21. The Australian company, which has its second headquarters in California, received a $30 million grant from NIH’s Rapid Acceleration of Diagnostics (RADx) initiative to scale up manufacturing.
The test comes with a smartphone app that provides user instructions, displays results and facilitates reporting to public health agencies, healthcare professionals, employers and educators. The lateral flow device, which uses optical detection of fluorescent nanoparticles, electronically analyzes its own results and communicates them to the app over Bluetooth.
Taking healthcare providers out of the equation makes it all the more important that manufacturers and regulators clearly communicate the possibility of false positives and negatives, a step FDA took when it announced Ellume’s EUA.
“For patients without symptoms, positive results should be treated as presumptively positive until confirmed by another test as soon as possible. This is especially true if there are fewer infections in a particular community,” the agency said. “Individuals who test negative and experience COVID-like symptoms should follow up with their health care provider as negative results do not preclude an individual from SARS-CoV-2 infection.”
Ellume founder and CEO Sean Parsons told BioCentury that FDA required the company to prove subjects could understand the results they received, and take the correct next step afterwards.
Part of the reason Ellume’s test was more conducive to over-the-counter use than Abbott’s was the fact that it checks itself against internal controls and interprets results for the reader, similarly to a digital pregnancy test, while BinaxNOW consists of a stand-alone lateral flow “card” that requires human interpretation.
The trade-off is between interpretation control and scalability, as Ellume’s test is more expensive to manufacture than Abbott’s, a difference seen in their respective prices of $30 and $5. But for consumers, the price difference is evened out by the prescribing, interpretation and reporting services that accompany Abbott’s test, which bring its total cost to $25.
HHS’s “COVID-19 Diagnostics At-Anywhere Design-a-thon,” a competition that includes apps to communicate COVID-19 results, could provide companies developing at-home tests with off-the-shelf tools. On Wednesday, the agency announced the top 16 entries; Oracle Corp. (NASDAQ:ORCL) placed first.
“These capabilities will be necessary for data capture and reporting when at-home, non-prescription and over-the-counter tests are authorized and begin saturating the market,” an HHS statement said.
Sorting out sensitivity
As part of their EUA announcements, both Abbott and Ellume provided details on the products’ at-home performance in specific patient subsets.
Abbott broke the study subjects down by viral load.
When at-home results from 52 individuals were compared with an FDA-authorized, lab-based PCR test, BinaxNOW had a 91.7% positive percent agreement with an authorized PCR-based test, a measure of sensitivity, and 100% negative percent agreement, or specificity, in people seven days or less post-symptom onset.
In a subset of those individuals whose PCR readouts had cycle threshold (CT) counts of 33 or below, the sensitivity jumped to 100%. CT counts reflect the number of thermal cycles a PCR machine undergoes before the target RNA is amplified enough to be detectable; the lower the CT count, the higher the starting concentration of viral RNA.
A separate study by University of California San Francisco researchers showed BinaxNOW had 93.3% sensitivity and 99.9% specificity in 15 individuals with CT counts of 30 or lower, six of whom were asymptomatic. The test did not detect infections in 11 individuals with CT counts >30, six of whom lacked symptoms.
Abbott cited emerging research showing SARS-CoV-2 can’t be replicated from samples with CT values in the 30s, suggesting individuals with those levels of viral load wouldn’t be infectious, and said frequent antigen testing will be key to catching these individuals when they become capable of infecting others.
Ellume focused on comparisons between symptomatic and asymptomatic subjects.
In a simulated home-use study with 198 participants, the test had an overall sensitivity of 95% and an overall specificity of 97%.
For symptomatic COVID-19 patients, the sensitivity was 96% and specificity was 100%, while for asymptomatic individuals, the sensitivity was 91% and specificity was 96%.
Direct comparisons of COVID-19 antigen test performance have been challenging because companies report clinical sensitivity and specificity in non-identical subject populations, and conduct limit of detection (LoD) studies, a measure of analytical sensitivity, with different protocols. FDA is conducting its own harmonized LoD studies of authorized tests and updating results on its website.
Company-reported data indicate Ellume’s test has a higher LoD, and therefore lower sensitivity, than other antigen tests, including Innova’s.
Clinical performance metrics are also not static. Abbott, which had previously reported a 97.1% sensitivity for point-of-care use of its test, provided updated data in its release Wednesday showing an overall sensitivity of 84.6% in people seven days or less post-symptom onset at all CT counts; sensitivity was 95.6% in people seven days or less post-symptom onset with CT counts of 33 or below.
While there is no consensus on thresholds for acceptable test performance, there is growing agreement that tests with lower sensitivity can still be valuable, particularly when used as part of a serial testing program.