University of Birmingham researchers have found that the accuracy of antibody tests for Covid-19 is highly dependent on timing.

By combining data from 54 studies, the researchers found that antibody tests are not accurate enough to confirm whether someone has had the Covid-19 virus if used in the first two weeks after symptoms appear.

The tests are only likely to be useful in detecting historic Covid-19 infection if used at least 14 days after symptom onset, and will still miss around one in ten cases. The researchers also fear accuracy will be lower from samples collected in the community, as the test samples evaluated in the studies were largely collected from hospitalised patients, making it unclear whether the tests can detect the lower antibody levels associated with milder cases of the disease.

There was also a lack of data available for samples taken later than 35 days after symptom onset, meaning the researchers cannot be sure about the accuracy of the tests long-term following infection.

However, the team has cautioned against over-reliance on its research, as the studies they evaluated were small, poorly reported and carried out in select patient groups.