Accuracy of COVID-19 antibody tests questioned

A systematic review of COVID-19 antibody tests has raised concerns about the poor quality of results they offer during the early stages of the disease (Cochrane Database of Syst. Rev . 2020, DOI: 10.1002/14651858.CD013652).

The first phase of the ongoing review looked at 54 studies of antibody tests published before the end of April, which were mostly based on test results from hospitalized patients in China. Some 61% of the studies involved laboratory-based methods such as ELISA, with the remainder carried out with handheld test kits that rely on lateral-flow strips.

The review found that these tests offered widely different sensitivities, meaning they varied in their ability to provide a genuine positive result. However, that’s not necessarily due to the quality of the tests, says Jonathan J. Deeks at the University of Birmingham, the review’s lead author: “It is largely driven by when the samples are taken from the patients.”

Tests given during the first week of COVID-19 symptoms only identified an average of about 30% of positive cases. After 2–4 weeks, the same tests caught 90% of cases.

The review team called for larger and more rigorous studies of antibody tests on patients with milder symptoms, and on people whose symptoms appeared over 1 month before testing. More recent antibody studies will be rolled into the review in the coming months.

The review is being coordinated by Cochrane, a UK–based charity renowned for its assessments of healthcare interventions.