An Oxford University study of more than 9,000 health-care workers suggests antibody tests are not effective for people who had only a mild COVID-19 infection, and that a significant number of people received negative results despite likely having had the virus.
The researchers looked at results of several commercial antibody tests on the market in the U.K. among health-care workers who reported losing sense of taste and smell – a symptom predictive of coronavirus infection.
Results from one test found that 47% of 903 people who tested positive for antibodies reported loss of their sense of taste or smell, but 30% of those whose results fell just below the threshold for a positive result also experienced this symptom, and likely had a mild COVID-19 infection. Researchers noted that reported rate of loss of smell and taste for other conditions, such as seasonal colds, is only around 3%.
“When we adjust for the likely proportion of staff reporting a loss of taste or smell due to reasons other than COVID-19, it suggests that the sensitivity of the antibody tests are 9 to 11% lower than had previously been reported,” explained study coauthor Tim Walker, MRCP. “It means that we need to consider those individuals with the full spectrum of mild to severe infection when we evaluate these tests. We could also consider reporting ‘high negative’ antibody results as ‘equivocal.’ What the test is being used for should of course influence any such decision.”
The researchers suggest that the reduced sensitivity of these antibody tests is related to how the tests are developed: They often use samples from symptomatic, and often hospitalized, patients with COVID-19. They recommend samples from mild and asymptomatic patients with confirmed infections be included in the evaluation process for antibody tests to increase accuracy and investigate whether mild illness is associated with a mild immune response.
“Our dilemma is that some people with a mild form of the disease can fall into an ‘equivocal’ zone that is currently classed as ’negative’ result,” Dr. Walker said. “Lowering the threshold to reclassify some false negatives as ‘positives’ will inevitably mean that some true negatives will then wrongly be categorized as positive.”