A patient having blood drawn for a Covid-19 antibody test
With people stuck at home due to Covid-19, one thing that may be key to allowing them to return to offices, eateries and entertainment venues is widespread testing for antibodies against the virus that causes it, to find who has been previously infected and may be immune. But although several serological antibody tests have already reached the market, experts pointed to a number of challenges that persist.
Last month, the Food and Drug Administration issued a policy allowing developers of serological tests to market or use them without agency approval as long as they validate them as accurate and reliable, notify the FDA of that validation and appropriately label them, stating that they are not to be used as a sole basis for diagnosis. In an April 18 statement, FDA Commissioner Stephen Hahn said the agency encourages developers to seek emergency use authorization, while it is also working with the National Institutes of Health to enable the NIH to evaluate the tests.
In other words, the FDA does not require an EUA for serological tests, but will review them if developers submit them. One example of a company that got an EUA is Cellex, a firm based in Research Triangle Park, North Carolina, that received such an authorization earlier this month.
Illustrating the potential usefulness of serological testing, New York Gov. Andrew Cuomo last week referred during a press conference to a study of 3,000 state residents, including 1,300 in New York City, which found that one-fifth of New Yorkers may have had Covid-19. If the study results can be extrapolated to the state as a whole, that would mean as many as 2.7 million of its 19.5 million residents. However, the World Health Organization warned Friday that there is no evidence that having antibodies after recovery from Covid-19 can protect against a second infection.
The most common means of testing for Covid-19 is to use reverse trascription-polymerase chain reaction, or RT-PCR testing, which typically involves using a cotton swab to detect active infection in the nose and throat. Antibody testing using blood samples is not so effective at detecting active infection, but it can detect whether some was infected before, even if they had no symptoms.
But the fact that serological tests don’t necessarily need FDA authorization could make it more challenging to scrutinize them.
“FDA approval is not needed to market serology tests, so a lot of these were developed very quickly, which is understandable,” said Nina Babic, associate professor of pathology and laboratory medicine at the Medical University of South Carolina, in a phone interview. “But what that means is that the available validation data must be reviewed and evaluated with scrutiny to understand what the shortcomings and shortfalls are.”
Covid-19 serology tests, Babic said, should have at least 99% specificity – meaning a false-positive rate of no more than 1%. Higher false-positive rates could be a problem, and anecdotally she has heard of rapid, point-of-care serology tests providing false positives due to cross-reactions with other coronaviruses that cause the common cold rather than the one that causes Covid-19, SARS-CoV-2.
She said she would prefer to have 1,000 people tested to provide enough data, though she would feel comfortable with 500 subjects. “For tests where high clinical specificity is needed, I look for at least 500 people being sampled to feel comfortable with stated test performance data,” she said.
The explosion of tests from many manufacturers has been a challenge, wrote Dr. Michael Borowitz, a professor of pathology at Johns Hopkins University, in an email, as it has led to concerns about variability of results.
“Serology tests are rarely perfect,” Borowitz wrote, pointing to the problem of false positives and false negatives. “While some of the serology tests have been well-tested and are known to be highly, if not perfectly specific, there are others out there that have not undergone the same degree of scrutiny.”
A lack of standardization is another part of the problem, said Dr. Amesh Adalja, a senior scholar at the Johns Hopkins Center for Health Security, in a phone interview. “There are many kits out there, but we don’t know how they compare to each other,” he said.
What’s also unclear is whether a positive result for SARS-CoV-2 antibodies means someone is immune.
“When people are vaccinated against a virus, like hepatitis or polio, they will develop antibodies, and we know that those antibodies confer immunity to the disease,” Borowitz wrote. “But having antibodies doesn’t necessarily mean that a person is immune, and this simply isn’t known yet about the antibodies produced in response to Covid-19, although a lot of studies are underway.”
Borowitz explained that it’s additionally unknown whether people with antibodies can still spread the virus, and in particular whether those with detectable SARS-CoV-2 virus are actually carrying it or simply carrying leftover RNA from dead viruses.
Population size can affect reliability as well.
“There many questions about what those antibodies mean on an individual level, but on a community level they have a role that’s easier to establish,” Adalja said.
On a community level, the percentage of people being tested can affect reliability as well, Babic said, giving the hypothetical example of testing in New York City versus Charleston, South Carolina. In other words, it’s easier to rely on the accuracy of a positive result in a place with a high population prevalence of the disease, such as New York – which had more than 141,000 cases as of Friday – versus one with a low prevalence, like Charleston, which had 412 cases.
“So the more population you test, the better idea you’re going to have about the prevalence of a disease, which, in turn, will affect how reliable that result is,” Babic said.
Photo: Pier Marco Tacca, Getty Images