As the world fights to bring the coronavirus pandemic under control, new blood tests that identify people with possible immunity are being touted as a key tool to return life to normal. But almost none of the dozens of tests that hospitals and clinics across the US are now rushing to obtain are being verified for accuracy by regulators.
These tests are not the kind that spot a current infection. Instead, they indicate if a person’s immune system had produced antibodies to fight the virus in the past. Widespread use of such tests could answer important questions about how much of the population has been infected by the virus, shedding more light on how deadly it really is. And some argue the tests could also hold the keys to restarting the economy: In the US and some European countries, the possibility of “immunity certificates” has been gaining traction as a way to help determine who can safely return to work.
But the accuracy of the tests in the US is currently far from guaranteed.
Over the last two months, the FDA has been loosening its regulations in an effort to fast-track available testing. In mid-March, the agency started allowing coronavirus antibody tests, also known as serology tests, to be sold and used without going through any review, as long as the test’s developer could show proof that they had validated their test. More than 95 antibody tests are now available, but only three have received official “emergency” authorization from the FDA, with the latest ones announced Wednesday.
The US is still suffering from a catastrophically short supply of nasal- and throat-swab tests that detect an ongoing infection, causing delays that critics say wasted the first precious weeks when the disease could have been contained. That crisis is now causing skyrocketing demand for antibody tests to identify how many people have recovered.
But experts fear that unanswered questions about these tests could breed the next testing crisis. Scientists don’t know enough about the new virus to determine whether survivors are really immune — or if most of these tests can reliably show whether they are or not.
“A test is only as good as its result, and having many inaccurate tests is worse than having no tests at all,” said Kelly Wroblewski, infectious disease director of the Association of Public Health Laboratories. According to the group, federal health agencies intend to start validating the dozens of serology tests on the market in the next week.
The UK has shown what can go wrong with antibody tests. Prime Minister Boris Johnson called them a “game changer” and the government ordered millions from China, only for them to later be deemed unfit for widespread use.
In the US, doctors and scientists say that without federal guidance, they are on their own when it comes to combing through and researching an ever-growing flood of tests. The result is a testing hodgepodge. In Alabama, California, Georgia, Michigan, Minnesota, Missouri, and Texas, nine hospitals, clinics, labs, and universities are using or planning to use at least 11 different antibody tests — none of them verified by the FDA, a BuzzFeed News review found.
Sold online to medical facilities with names like “CoronaCheck” and “CoronaCide,” manufacturers are marketing their test kits’ ease of use and ability to quickly generate results. One company, BioMedomics, boasts that its test can be “deployed in businesses, schools, airports, seaports and train stations, etc., giving it the potential to become a compelling force in the fight against this global threat.”
In this free-for-all environment, some companies are trying to edge out the competition by falsely touting their tests as “FDA-approved” or “FDA-authorized,” the agency said last week.
One manufacturer, SafeCare Canada, is selling a $450 rapid-detection coronavirus antibody test and falsely claiming that the product is “FDA certified,” BuzzFeed News found. Asked for comment, SafeCare Canada's co-chief executive officer, Darrell Bauschke, said by email: "Not interested in your news format. Or Target audience." The webpage for the test was taken down after BuzzFeed News published this story. An FDA spokesperson declined to comment on the company.
Health care providers are making decisions about which tests to provide with very little information. Last week in Huntsville, Alabama, Synergy Wellness Clinic started offering tests from two companies in China. Michael Chang, a staff doctor, said that he pored over studies from several manufacturers before settling on the ones with the most promising claims about accuracy rates. But he said he did not have time to independently verify if the tests work as advertised, as he ordinarily would.
“You take it at face value,” he said. “We don’t live in normal times. It’s a leap of faith.”
Chang acknowledged that he and others are using tests that have not been thoroughly vetted. But he also pointed out that, given that the virus has been around less than four months, no test right now will be backed by perfect data. And if patients are warned about caveats upfront, antibody testing will likely still be worth doing.
“Maybe we’ll look back and say the tests are not very sensitive. We’ll know that in hindsight,” he said. “The question is, are we going to wait until hindsight to do what we think should have been done months ago? That’s the thing people are currently grappling with.”
Antibody testing is being touted as a key to putting America back in business.
The coronavirus outbreak is nowhere close to ending, in the US or elsewhere. More than a half-million Americans have been confirmed to be infected, with many others missed by the lack of testing for the virus. Hospitals, particularly in hard-hit spots like New York City, New Orleans, and Detroit, are overwhelmed by scores of sick and dying patients. As of Wednesday, more than 26,000 Americans had died from COVID-19.
But there are signs that the pandemic is at least starting to slow in some places. This week, governors of states on the West Coast and in the Northeast said they’ve started discussing when and how they will work together to reopen the economy.
With millions out of work, government officials are hopeful that widespread antibody tests can be part of a strategy to get back in business without worsening the outbreak. In late March, the CDC and local public health agencies began the first of a series of communitywide serological testing surveys, starting in Washington state and New York City before expanding to more hard-hit states and some households, a CDC spokesperson said by email. And the idea of “immunity certificates” that would allow coronavirus survivors to officially reenter society is being discussed at various levels in the US, Germany, Italy, and the UK.
“How do you get people back to work as quickly as possible? It’s gonna come down to testing,” New York Gov. Andrew Cuomo said last week. “You’re going to have to know who had the virus, who resolved the virus, who never had it.”
And Dr. Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, recently told CNN, “If their antibody test was positive, one can formulate strategies about whether or not they would be at risk or vulnerable to being reinfected,” which would be particularly important for health care workers. An immunity passport “might actually have some merit under certain circumstances.”
But there are still big questions about how much immunity people who have recovered from COVID-19 really have.
Antibody tests analyze a person’s blood for antibodies to a particular infection. Generally for infectious diseases, immunoglobulin M (IgM) antibodies, which are the first type of antibody produced by the immune system in response to a pathogen and stay in the bloodstream for weeks, indicate that someone was recently infected. Immunoglobulin G (IgG) antibodies stay for years and are a sign of a past infection. Serology tests generally check for both antibodies in a blood sample.
For other infectious diseases, IgGs can signal that a person is immune for a year, as in colds, to a lifetime, like measles. That pattern has inspired researchers to start collecting COVID-19 survivors’ donated blood plasma, on the hunch that their antibodies could help heal sick patients.
But SARS-CoV-2 is so new that researchers don’t know with certainty yet if that principle holds true here. They also don’t know how long immunity might last, or what level of antibodies might be needed for immunity.
One hopeful clue about how the virus might behave is that in survivors of the coronavirus behind the 2003 SARS outbreak, immunity seemed to last up to three years.
Other reports are less hopeful. South Korea reported Monday that at least 116 patients who had been cleared of COVID-19 tested positive again. And a preliminary study of 175 recovered patients in China found that nearly a third had antibody levels so low that they might not offer protection.
There are also big questions about the accuracy of the antibody tests.
Experts say that widespread testing will be important to understand how much of the population has gotten infected with the virus, giving us a clearer picture of how fatal it can be. On an individual level, however, interpreting a positive or negative test result is more complicated.
You might get a false negative because your body has not yet produced antibodies in the first few days of the infection, or even at all. Alternatively, you may get a false positive because the test picks up antibodies from a different infection, in which case you might mistakenly think you’re immune and it’s safe to go outside.
Complicating matters further, each test has its own sensitivity and specificity rates. Sensitivity is the percentage of times the test correctly identifies people who have the condition being tested for, and lets you know the rate of false negatives you can expect. And specificity is the rate the test correctly identifies people who don’t have that condition, making clear how often it generates false positives.
For example, Cellex — the first company to receive FDA emergency authorization for its SARS-CoV-2 antibody test — says its test is 93.8% sensitive and 96.4% specific. Ortho Clinical Diagnostics says its test, authorized Wednesday, is 100% sensitive and 83.3% specific. A test from Chembio Diagnostics, also authorized Wednesday, says it is 93.5% sensitive and 90.2% specific.
While these varying accuracy levels may be acceptable for figuring out how much of the population has been infected, since scientists can correct for the errors, they present more problems for patients trying to interpret their individual results. When a smaller percentage of the population is infected, for example, the chance of an individual getting a false positive is much higher than the test accuracy number alone would suggest.
And accuracy numbers for the flood of unauthorized tests are even less clear, since they haven’t been independently validated. For example, a test from BioMedomics, which doesn’t have FDA authorization, says it is 88.66% sensitive and 90.63% specific.
Scott Becker, executive director of the Association of Public Health Laboratories, said that although having a variety of test options is better from a supply chain perspective, having an unknown number of unreliable tests among the 95-plus non-FDA-reviewed antibody tests means that “we will be chasing down a bunch of rabbit holes.” The organization would prefer that all test makers seek emergency use authorization, as Cellex, Ortho Clinical, and Chembio did.
Labs and health care providers are having to validate the tests themselves — or pick at random.
But for some small labs, clinics, and hospitals looking to quickly ramp up antibody testing, the unauthorized tests have additional appeal. Though Cellex’s test has been authorized since April 1, many institutions aren’t using it, in part because its use is restricted to labs certified to carry out moderately and high-complex tests. Other sites are using different tests because they are worried about supply chain issues. Still others said the unauthorized tests were more appealing because they’re cheaper.
Tiffanny Blythe, who started offering antibody testing this week at Blue Lotus Family Medicine, her primary care clinic in Kansas City, Missouri, said she researched 10 tests, looking for ones that had high detection rates and, more importantly, explained how those rates had been derived.
“Not all the companies offer where they get sensitivity and specificity rates from — they just kind of state it as fact,” Blythe said. She ended up ordering tests from Confirm BioSciences, which are now sold out, so she bought a new test called CoronaCheck.
On its second day of testing, her clinic got 400 calls from interested people.
Because of the accuracy issues, Blythe said she still sends patient samples to local public health laboratories for further confirmation and encourages everyone to continue to socially distance, wear masks, and wash their hands properly.
A representative for the Beaumont Health Research Institute, in Michigan, said it was using tests by the manufacturer PerkinElmer because they can process many more samples at once and provide information about the levels of antibodies, not just whether they are present. For the same reasons, Kara Lynch, codirector of the clinical chemistry laboratory at Zuckerberg San Francisco General Hospital, said she plans to start testing patients and staff with a test made by the company ET Healthcare.
And Metroplex Medical Centres in Dallas is using a test by Hangzhou Clongene Biotech Company because it’s cheaper for patients than the Cellex test, said physician assistant Sneha Patel.
Some institutions may also be wary about depending on a single supplier, given how equipment shortages have hindered diagnostic testing in the US so far.
Already, a high-profile attempt by United Biomedical to serologically test every person in San Miguel County, Colorado, has reportedly been delayed by a lack of staff and supplies, according to county officials.
The University of Minnesota made its own antibody test because “the goal was to be less dependent on potentially problematic supply chains,” according to a spokesperson. Stanford University has done the same, with a representative citing a desire to avoid the “multiple supply chain failures” it had initially encountered with its diagnostic tests. Other academic institutions developing their own tests include Emory University.
Sorting through the tests is a challenge, said Elitza Theel, director of the Mayo Clinic’s infectious diseases serology laboratory in Rochester, Minnesota. “We don’t know how well most of those perform,” she said.
The Mayo Clinic considered five antibody tests before settling on two, based on the results of internal validation tests run on patient samples. (It did not disclose which tests it was using, or which ones ultimately failed their validation screen.) That process took three to four weeks before the clinic started to test patients this month. That’s much faster than the 8 to 10 months it typically takes to evaluate and implement a new test, Theel said.
Also in a matter of weeks, the University of Southern California and Los Angeles County teamed up to test the blood of 1,000 randomly selected people for COVID-19 antibodies last weekend. While individuals’ results will be fairly limited — they will simply indicate whether or not they have the antibodies — the researchers say the study will be more useful in providing a snapshot of the virus’s presence in the community.
The study used tests made by Premier Biotech, a Minnesota company, because they were donated and their data seemed reliable, said Neeraj Sood, a health policy and economics researcher at the University of Southern California. At worst, it appears to have a false negative rate of 30%, which his team’s statistical analysis will take into account when assessing the population data.
“The fact is if the FDA hadn’t allowed that, we wouldn’t have had access to these tests, we wouldn’t have been able to do the study,” Sood said. “I would have liked if FDA had done the validation last week. But at least they made the test available so we could start doing our own validation and we could put the study in the field — because we need this information right now.” ●
This story was updated to include a statement by SafeCare Canada's co-CEO and the fact that the webpage for the test was taken down after the publication of this story.