Antibody tests do not pick up people who had mild coronavirus, Oxford study suggests

Jennifer Rigby
·4-min read
Tests
Tests
Coronavirus Article Bar with counter ..
Coronavirus Article Bar with counter ..

Antibody tests may be missing large numbers of people who contracted Covid-19 because they don't work for people who had a mild infection, new research from Oxford University suggests.

A study of more than 9,000 healthcare workers suggested significant numbers of people were getting 'negative' test results, despite probably having had the virus.

The work has major implications for government health policy, and scientists said it might also mean reviewing where the threshold between negative and positive results lies.

Antibody tests are currently used to map the outbreak among a population, but they are also central to potential plans for "immunity passports" - an idea mooted to allow people to go back to normal life. It is not yet clear whether antibodies, the protective proteins produced to help fight off the virus, provide any long lasting immunity to Sars CoV-2.

The Oxford University study compared the results of antibody tests among healthcare workers who had also reported losing their sense of taste or smell, a key coronavirus symptom, as shown below.

Of the 903 people who tested positive for antibodies on one test, 47 per cent reported a loss of their sense of taste or smell. But among those whose test results fell just below the threshold for a positive antibody result, meaning that they would currently be classed as not having had the virus, 30 per cent also reported a loss of sense of taste or smell.

"You can see that below the cut-off, there is a rising proportion of people who report a loss of their sense of smell or taste, and this suggests that the test threshold is missing people with mild disease,"  said Dr Tim Walker, one of the study's authors. "Of course there will be plenty of people, too, who will have had no symptoms whatsoever and will still have antibodies."

The background rate of people who would report these symptoms for another reason - from seasonal colds to other conditions - is around three per cent, researchers added.

A further 387 people also tested just below the threshold in terms of antibodies for a positive test, but did not have symptoms. They may have been asymptomatic coronavirus patients, but the researchers could not say with confidence that they had the virus.

The study used several antibody tests, including the Abbott diagnostic, which is one of the four main commercial tests currently used in the UK. The results suggest these were around 11 per cent less sensitive than is currently believed - around 98 per cent - in part because the tests were developed using samples from symptomatic, often hospitalised, patients.  

The team suggested that samples from mild and asymptomatic patients with confirmed infections should be included in the evaluation process in order to investigate further whether mild illness is associated with a mild immune response - responses that are currently being missed by the antibody tests.

Dr Walker said that antibody tests have been designed to make sure that people who did not have antibodies were not mistakenly told that they did, giving them a false sense of security about the virus. Because of this, some people with antibodies would be missed, he said. 

Professor Will Irving, a virologist at the University of Nottingham who was not involved in the study, said: "It is very difficult, in fact impossible, to know to what extent we are underreporting positives. But we are missing some patients, and what is shown in this paper is that one particular reason for this is that the cut-off is set too high."  

In the future, if scientists prove that having antibodies does provide some kind of immunity, there is a more concerning issue - thousands who do actually have antibodies may be forced back into lockdown, by their negative antibody results, when they do not need to be.

"It depends if you would rather wrongly deny people passports or give too many out to people who are susceptible, and you have to live with the consequences of both of those decisions," said Dr Walker. 

A Department of Health and Social Care spokesperson said: “We do not yet know whether antibodies indicate immunity from reinfection with coronavirus or if they prevent transmission. However, antibody testing is an important part of our testing strategy because it helps us understand how COVID-19 is spreading across the country.”

Abbott said its tests had showed strong performance rates, with 100 per cent sensitivity ratings on 73 samples.

Protect yourself and your family by learning more about Global Health Security