Researchers based at Kings College in London have discovered that commercially available antibody tests for COVID-19 vary widely in accuracy.
The ability of the 10 antibody tests analyzed to detect those who had not been infected varied between 82% and 100% and to detect those who had been infected from 61% to 87%. But the team acknowledged that waiting 20 days or more from symptom onset improved the accuracy of the tests.
Notably, the findings revealed that some fast, lateral flow immunoassay tests (LFIAs), initially assumed to be less accurate, actually produced high quality results.
“We found that some of the quick single-use kits (LFIAs) are as accurate as our sophisticated laboratory technologies. Encouraged by these findings we are piloting LFIAs in the hospital to give doctors a quick reliable answer in a range of clinical settings,” said Jonathan Edgeworth, M.D., Ph.D., a professor at Kings College and senior author on the study, which was published in the journal PLoS Pathogens.
The main tests used to diagnose current infection with the SARS-CoV-2 virus are PCR-based tests that can detect viral RNA in throat and nose swabs. Antibody tests can’t accurately pick up current infections, as it takes days for antibodies to the virus to begin forming in the body of infected individuals. However, they are able to detect if someone has already had the virus. They could also theoretically assess how likely someone is to be immune to future infections, although the degree of protection accorded by antibodies against COVID-19 is currently unclear.
There has been intense activity in the diagnostics sector to produce a variety of COVID-19 tests, and there are a range of antibody tests currently being used around the world. In this study, Edgeworth and colleagues tested the accuracy of 10 serological antibody tests – a chemiluminescence-based platform, two ELISAs and seven colloidal gold LFIAs.
Using highly specific, pre-validated in-house tests they compared the accuracy of the commercial tests using 110 patient samples known to be positive for mild-severe SARS-CoV-2 infection and 50 negative control samples. The samples were taken from 1-30 days after suspected onset of infection.
Overall, the specificity to detect negative controls and sensitivity to detect true cases of infection varied widely between tests. In samples taken at 20 days or more after suspected infection the accuracy was higher for all tests with most achieving a sensitivity of more than 95%.
Three LFIA tests – Accu-Tell, SureScreen and Spring – had the highest sensitivity at earlier testing times, with specificities of 98-100%.
The researchers are now validating the tests that scored most highly on a wider scale at Guy’s and St Thomas’ Hospitals in London.
“Next generation antibody tests may improve on those currently being trialed, but our results demonstrate that LFIAs may have utility in a hospital setting as of now, particularly if deployed where a rapid result could aide a clinical pathway or decision in real time, such as ward location or prioritization of further diagnostics and follow up,” write the authors.
The London researchers are not alone in recognizing issues with antibody test variability. Siemens Healthineers announced earlier this month that it will collaborate with the U.S. Centers for Disease Control and Prevention and the Joint Research Centre of the European Commission to develop a process for standardizing SARS-CoV-2 assays.
