As the FDA continues to grant Emergency Use Authorizations (EUAs) for rtPCR tests to diagnose COVID-19, public health experts, scientists and industry groups are calling on the agency to lower requirements around test sensitivity for rapid antigen tests.
There has been growing interest in the use of rapid antigen tests to screen individuals at nursing homes, employees returning to work and students at schools and universities (for a terrific discussion of the argument for widespread, cheap rapid antigen testing, listen to Harvard University T.H. Chan School of Public Health epidemiologist Dr. Michael Mina on the podcast, This Week in Virology). Governors from seven states banded together to purchase 3 million of the tests earlier this month.
The tests deliver results within 30 minutes and can be, in theory, produced cheaply and en masse. Criticism of the tests revolves around their poorer performance at detecting low levels of viral RNA, but public health experts such as Mina point out that they can be ideal for identifying individuals who are most infectious.
On Aug. 5, the WHO published draft Target Product Profiles (TPP) for COVID-19 diagnostics that detail “desirable and minimally acceptable profiles” for four different types of tests: rapid point-of-care antigen tests; molecular diagnostics such as rtPCR tests for acute or subacute SARS-CoV-2 infection; serology tests for situations with low-volume needs and serology tests with high-volume needs.
According to the WHO’s guidance, COVID-19 antigen tests would need to have a sensitivity of at least 70% to be considered “acceptable” and equate to a false negative rate of no greater than 30%. The specificity would need to be at least 97%, meaning only 3% of the time would an individual receive a “positive” result and actually be negative for SARS-CoV-2. The sensitivity and specificity criteria for an rtPCR test are much more stringent at 95% and 99%, respectively.
As of Aug. 11, the FDA has granted EUAs for 210 tests, including 171 molecular tests, 37 antibody tests and 2 antigen tests.
The sensitivity for the authorized tests range from 84% to 98%.The FDA requires sensitivities of at least 95% for antigen tests in asymptomatic individuals and at least 80% for tests in symptomatic individuals.
As demand mounts for more rapid diagnostics for SARS-CoV-2, it may not be out of character for the agency to lower the bar for rapid antigen tests.
According to the FDA’s FAQs for COVID-19 diagnostics, screening tests that are looking for the occurrence at the individual level, even if there is no reason to suspect infection, such as known exposure, are regulated by the FDA as in vitro diagnostics.
These could be assays to test all employees before returning to work or as part of testing plans for a school to test students and faculty, the agency stated. In general, not all screening tests regulated by the FDA are required to meet clinical sensitivity thresholds and the criteria vary even within classes of devices.
For example, over-the-counter pregnancy tests are regulated as class II devices and while they are required to establish analytic sensitivity related to the limit of detection, they are not required to establish clinical sensitivity like what is required of COVID-19 tests. Conversely, nucleic acid tests for tuberculosis, also class II devices, are expected to meet predefined sensitivity thresholds of at least 72%.
More stakeholders are turning to fast, point-of-care tests for COVID-19. CMS has partnered with a test provider to administer an antigen test at nursing facilities and the UK government said on Aug. 3 that it would start using two rapid tests in nursing homes in September.
A challenge with current rtPCR tests used in the US is the time it takes to process samples and return a positive result. Waits vary widely, from a few days to weeks.
The delays are compounded in some areas due to limited supplies. In a mid-July survey of Americans diagnosed with COVID-19 conducted by Harvard University, Northeastern University and Rutgers University, the mean waiting time nationally was just over four days; the median waiting time was three days. Thirty-seven percent of people received test results within two days but 21% waited more than five.
A study published in June by researchers at Harvard University and the University of Colorado found that frequency of testing was more important than test sensitivity. Tests conducted more frequently had a greater effect on limiting spread, “with only a small margin of improvement provided by using a more sensitive test.”