ABSTRACT
Rapid antigen detection (RAD) tests are commonly used for the diagnosis of SARS-CoV-2 infections. However, with the continuous emergence of new variants of concern (VOC), presenting various mutations potentially affecting the nucleocapsid protein, the analytical performances of these assays should be frequently reevaluated. One hundred and twenty samples were selected and tested with both RT-qPCR and six commercial RAD tests that are commonly sold in Belgian pharmacies. Of these, direct whole-genome sequencing identified the strains present in 116 samples, of which 70 were Delta and 46 were Omicron (BA.1 and BA.1.1 sub-lineages, respectively). The sensitivity across a wide range of Ct values (13.5 to 35.7; median = 21.3) ranged from 70.0% to 92.9% for Delta strains and from 69.6% to 78.3% for Omicron strains. When taking swabs with a low viral load (Ct > 25, corresponding to <4.9 log10 copies/mL), only the Roche RAD test showed acceptable performances for the Delta strains (80.0%), while poor performances were observed for the other RAD tests (20.0% to 40.0%). All the tested devices had poor performances for the Omicron samples with a low viral load (0.0% to 23.1%). The poor performances observed with low viral loads, particularly for the Omicron strain, is an important limitation of RAD tests, which is not sufficiently highlighted in the instructions for use of these devices.