By The Washington Post · Laurie McGinley · NATIONAL, HEALTH, SCIENCE-ENVIRONMENT
Antibody, or serological, tests are designed to identify people who may have overcome covid-19, including those who had no symptoms, and developed an immune response. They are not designed to detect active infections. Some officials tout the blood tests as a way to reopen the economy by identifying individuals who have developed immunity and can safely return to work. But many scientists, as well as the World Health Organization, say evidence is lacking that even high-quality antibody tests can prove someone has immunity to the novel coronavirus.
Now, the emergence of dozens of tests never reviewed by the FDA - many of which are being aggressively marketed - could confuse doctors, hospitals, employers and consumers clamoring for the products, according to critics who say the agency's oversight has been lax. The questions are taking on special importance as federal and state officials debate various strategies, including using serological testing, to help determine when they can end state and local lockdowns.
"A test is only as good as its results," said Kelly Wroblewski, director of infectious disease programs at the Association of Public Health Laboratories, which has been urging the FDA to take a close look at the unapproved antibody tests. "Having many inaccurate tests is worse than having no tests at all."
Prodded by such concerns, the FDA recently has stepped up warnings and is joining other agencies, including the National Cancer Institute, to try to determine whether the unvetted tests actually work.
On Thursday, during an interview with Washington Post Live, FDA Commissioner Stephen Hahn said "people should be very cautious" about tests that have not undergone the rigorous process of getting emergency use authorization from the agency. So far, only four tests have gotten such authorizations: Cellex, Chembio Diagnostic Systems, Ortho Clinical Diagnostics and Mount Sinai Laboratory.
The new validation of unauthorized tests will be done primarily by NCI, working with the Centers for Disease Control and Prevention, the National Institute of Allergy and Infectious Diseases and academic laboratories. Much of the effort will occur at the cancer institute's human papillomavirus (HPV) serology testing lab at Frederick National Laboratory. The assessments also will be used for some tests seeking FDA emergency use authorization, said Hahn in a statement Saturday evening. The goal is to have "greater confidence in test performance," he said.
Meanwhile, however, the unreviewed tests remain on the market. In mid-March, the FDA, wanting to ensure quick access to serological tests, said manufacturers could sell them after validating results themselves and simply notifying the agency. The labels on the products must state that they aren't FDA-approved or meant to be used as the sole basis for diagnosing active infections. The agency lists on its website the companies and laboratories that have submitted those notifications. In response to concerns that the list might be seen as an endorsement, the agency updated its website to make clear which tests are authorized by the FDA and which aren't, Hahn said.
Testing experts warn that the risks of inaccurate tests are high. A wrong result could, for example, indicate that individuals have immunity against the virus when they don't, potentially resulting in behavior that would endanger themselves and their family. The peril of buying unvetted tests was underscored by the British government's recent $20 million purchase of antibody tests from China that didn't work. Even for high-quality tests, scientists say, they don't yet know the level of antibodies needed to make someone immune from the coronavirus or how long any protection might last.
Many of the unreviewed tests being sold in the United States are made in China or elsewhere in Asia, and are being marketed heavily online. Distributors staunchly defend their products as a way to accelerate testing in a country that has been hobbled by shortages and to provide individuals, employers and governments with critical information.
Jonathan Cohen, president of 20/20 GeneSystems, based in Rockville, Maryland, said the test he is selling to health-care professionals, named "CoronaCheck," has been approved by the Chinese version of the FDA and verified by multiple hospitals in the United States. "These tests are quite good," he said.
Paul Kahlert, owner and president of ARCpoint Labs of Edina, Minnesota, said customers are eager to know whether they were infected and also whether they may be able to donate blood that can be used to manufacture "convalescent plasma," which is being tried as a potential treatment for people with covid-19. "People are so appreciative we are doing this," he said. "They want to help."
But an information sheet about the ARCpoint antibody tests raises concerns, said Scott Becker, chief executive of the Association of Public Health Laboratories. The sheet said the presence of certain antibodies suggests that a person has "functional immunity" and can "discontinue social distancing." Becker said scientists don't know enough to say that about antibody tests for the coronavirus.
Amanda Harley, a spokeswoman for ARCpoint Labs, a franchise company based in Greenville, South Carolina, said the information sheet was created as a guide for franchisees and is no longer used. John Constantine, the company's president and chief executive, said the firm also includes disclaimers on its information, such as that test results are not a guarantee of immunity. "We would never say, 'You are good. Go see grandma without a mask,' " he said.
Antibody tests typically come in either rapid versions that use a small amount of blood from a finger prick and may cost as little as $20, or others that rely on traditional blood draws and cost considerably more. Many of the finger-prick ones are being marketed for use in doctors' offices and pharmacies, even though the FDA requires that all unauthorized tests be run in so-called CLIA laboratories certified to conduct highly complex tests, Becker said.
The agency's willingness to allow tests on the market without review means "you are getting a lot of folks where this isn't their core competence entering the market, and there is an opportunity for flawed results in that environment," said David Morgan, senior vice president for Eurofins U.S. Clinical Diagnostics. A Eurofins subsidiary called Boston Heart Diagnostics is starting to offer an antibody test to hospitals and is seeking FDA authorization, he said.
Some state and local health departments also have warned consumers and doctors to avoid tests that haven't been authorized. In San Diego County, officials shut down a pop-up antibody testing site at a local community college.
Some of the unreviewed tests were developed by high-quality academic medical centers and aren't a source of concern among public health experts. Emory University, for example, developed a test it is planning to use for staff and patients to see how broadly the virus has been transmitted in its community.
Marybeth Sexton, an Emory infectious disease specialist, recently got the test because she wondered if a bout of illness in February was caused by the coronavirus. The test detected antibodies to the virus, but because she is still uncertain whether she's immune, she plans to continue practicing social distancing and wearing a mask. "I won't do anything different," she said.
In New York City, periodontist Scott Froum said he is exploring offering antibody tests to staff and patients when he reopens for routine care, as an extra layer of protection against the virus. But he's wrestling with regulations and choices. "You have to be careful," he said. "Some of the tests don't look kosher."
The accuracy of any of the tests is a big issue. Accuracy is typically determined by two measures, called sensitivity and specificity. The sensitivity of a test is the true positive rate, and a test with high sensitivity will have few false negatives. The specificity test is the true negative rate, and a test with high specificity will have few false positives.
But even for a very accurate test, its predictive value depends on how prevalent the disease it's designed to detect is in the population. Say a test was 95 percent sensitive and 95 percent specific - a fairly accurate test. But if only 1 percent of the population were infected, the test would identify far more false positives than real positives. That's because when there are very few people in the population with the disease, the false positives can outnumber the true positives.
A positive result would have a slim chance of being accurate. But if 5 percent of the population were infected, a positive result would have a 50-50 chance of being a real positive.
The FDA is expected to authorize more tests in coming days and weeks. But to get the volume of serology testing the nation needs, officials say, big manufacturers and laboratories will have to be involved. One big manufacturer, Abbott, recently launched a serology test and plans to produce 20 million tests a month by June. The company said it will apply to FDA for official authorization.