Antibody Tests Aren't Accurate, But You Can Still Buy One

(Bloomberg Opinion) -- Doctors, medical administrators and consumers are falling for misleading marketing of Covid-19 antibody tests. And the FDA has exerted almost no oversight over a hodgepodge of different tests that have come on the market since April, promoted with a blitz of direct emails to clinics and hospitals that tout the tests’ ability to detect past infections of SARS-CoV2.

The FDA made the opposite mistake with the diagnostic tests meant to flag active infections. Those rolled out too slowly, when even imperfect tests would have saved lives at that early stage. There’s no reason to rush out antibody tests since there are no clear lifesaving benefits to detecting past infections — and obvious risk of harm if the tests give misleading results.

Though much remains unknown, the problem is not that the human body fails to mount some protective antibody response after infection. There’s evidence that it does. It’s that the test results are not what the marketing has led people to expect. They don’t measure a clear indication of past infection.

Recent news reports announced that these tests were “wrong half the time,” attributing this to a CDC report. The big problem reported by the CDC is false positives, and the fact that real positive cases are rare enough that the majority of positive tests could be incorrect.

But the issue is more complicated than that. There’s a misunderstanding of what a false or true positive means.

Human blood is full of antibodies left over from different viruses that can easily trigger a test designed to pick up antibodies for SARS-CoV2. The tests work by taking proteins — usually one kind of protein — from the virus and placing it on a surface. Once a blood sample comes in contact with the test strip, any antibodies to the virus should stick to those proteins. 

Duane Wesemann, a Harvard professor who is studying immunity to the coronavirus, says he uses more complex tests in his laboratory that can measure the quantity of antibodies and look for one that binds to different parts of the virus. 

When applied to people who who’ve had a documented infection — in other words, they tested positive while symptomatic — their blood shows lots of antibodies that latch onto several different proteins associated with SARS-CoV2. People who haven’t had a documented infection still sometimes make antibodies that bind to proteins associated with the virus, and are thus capable of triggering a positive response in some antibody tests. Those people might have had a past, under-the-radar infection of SARS-CoV2 … or not

Test results can also be complicated by the fact that there are several classes of antibodies, including the faster-responding IgM antibodies and the slower and more persistent IgG antibodies. IgM antibodies not associated with Covid-19 can nevertheless potentially stick to the viral proteins in a test and register as positive. 

“There are so many tests and so many levels by which things become positive,” Wesemann says. “That’s why this is such a mess.”

All this means many people who never had the virus could show some antibody response to it — coming from some of their more promiscuous antibodies. Will some of those other antibodies turn out to affect the course of disease, perhaps playing a role in explaining its wide variety of symptoms? This remains an open question, although researchers have expressed doubts. 

Despite all the unknowns, there’s a heavy promotional effort behind these subpar tests — even though there’s no special urgency for them, as there was for diagnostic tests that would detect active cases. While the FDA was too slow to greenlight those tests, now, “the pendulum swung in the other direction,” says David Grenache, chief science officer for TriCore Reference Laboratories in New Mexico and president of the American Association of Clinical Chemistry. The FDA opened the floodgates for antibody tests, requiring companies to do some accuracy testing but not reviewing their data.

The demand was driven by the idea promoted in April that antibody tests might allow some lucky people to get “immunity passports” and re-enter normal work and life. Soon after, “my inbox exploded with blanket marketing messages,” Grenache says, coming from testing companies, distributors, or forwarded by doctors and hospitals. “I was getting forwarded emails from hospital administrators saying… ‘Hey I got this email and this looks like something we should be doing here.’” 

“Many of them were companies I’d never heard of, and I’ve been in the laboratory medicine industry for 30-plus years,” he says. “People who didn’t have any expertise or background in laboratory diagnostics were getting pinged with all this marketing information that was deceptive.” Some companies even falsely claimed their tests had FDA approval, he says.

Most of these early tests were known as rapid-response. Doctors were told they could administer the tests right there in the office while a patient waited for results. But such tests were intended to be performed in clinical laboratories capable of doing complex analysis and following government regulations.

Later, he says, reputable companies such as Abbott labs and Roche made more reliable tests which were given “emergency use authorization,” which means somewhat more FDA oversight. These tests aren’t perfect, but their false positive rates are at least more accurately stated. But even so, there’s still no “emergency” need for these tests.

Given what he’s learned, Harvard’s Wesemann remains skeptical of the various antibody surveys of large swaths of the Bay Area, New York, and other regions, the results of which still may depend on the accuracy of the test more than the actual prevalence of past infections. Even now, surprisingly little is known about silent infections — how many have really occurred and what role they play.

These open questions shouldn’t be overinterpreted to mean people who have had the virus and recovered aren’t likely to be protected by their immune systems, or that vaccines are hopeless. And more and better tests are coming, says David Walt, a Harvard professor of pathology.

But for doctors, scientists, and consumers to be able to make any use of them, they have to be accurate. And that means they need to get a very close look from the FDA.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

Faye Flam is a Bloomberg Opinion columnist. She has written for the Economist, the New York Times, the Washington Post, Psychology Today, Science and other publications. She has a degree in geophysics from the California Institute of Technology.

©2020 Bloomberg L.P.