Diagnostic Performance of Rapid Antigen Testing for SARS-CoV-2: The COVid-19 AntiGen (COVAG) study

BACKGROUND: Rapid diagnostic testing for SARS-Cov-2 antigens is used to combat the ongoing pandemic. In this study we aimed to compare two RDTs, the SD Biosensor Q SARS-CoV-2 Rapid Antigen Test (Roche) and the Panbio COVID-19 Ag Rapid Test (Abbott), against rRT-PCR. METHODS: We included 2,215 all-comers at a diagnostic center between February 1 and March 31, 2021. rRT-PCR-positive samples were examined for SARS-CoV-2 variants. FINDINGS: Three hundred and thirty eight participants (15%) were rRT-PCR-positive for SARS-CoV-2. The sensitivities of Roche-RDT and Abbott-RDT were 60.4 and 56.8% (P < 0.0001) and specificities 99.7% and 99.8% (P = 0.076). Sensitivity inversely correlated with rRT-PCR-Ct values. The RDTs had higher sensitivities in individuals referred by treating physicians (79.5%, 78.7%) than in those referred by health departments (49.5%, 44.3%) or tested for other reasons (50%, 45.8%), in persons without any comorbidities (74.4%, 71%) compared to those with comorbidities (38.2%, 34.4%), in individuals with COVID-19 symptoms (75.2%, 74.3%) compared to those without (31.9%, 23.3%), and in the absence of SARS-CoV-2 variants (87.7%, 84%) compared to Alpha variant carriers (77.1%, 72.3%). If 10,000 symptomatic individuals are tested of which 500 are truly positive, the RDTs would generate 38 false-positive and 124 false-negative results. If 10,000 asymptomatic individuals are tested, including 50 true positives, 18 false-positives and 34 false-negatives would be generated. INTERPRETATION: The sensitivities of the two RDTs for asymptomatic SARS-CoV-2 carriers are unsatisfactory. Their widespread use may not be effective in the ongoing SARS-CoV-2 pandemic. The virus genotype influences the sensitivity of the two RDTs. RDTs should be evaluated for different SARS-CoV-2 variants.