Comparison of the quantitative DiaSorin Liaison antigen test to RT-PCR for the diagnosis of COVID-19 in symptomatic and asymptomatic outpatients.

BACKGROUND: We evaluated the quantitative DiaSorin Liaison SARS-CoV-2 antigen test in symptomatic and asymptomatic individuals consulting their general practioner (GP) during a period of stable intense virus circulation (213/100,000 habitants per day). METHODS: Left-over RT-PCR positive (n=204) and negative (n=210) nasopharyngeal samples were randomly selected among fresh routine samples collected from patients consulting their GP. Samples were tested on Liaison XL according to the manufacturer's instructions. Equivocal results were considered positive. RESULTS: Overall sensitivity and specificity of the Liaison antigen test compared to RT-PCR were 67.7% [95% confidence interval (CI): 60.9%-73.7%] and 100% [CI: 97.8%-100%]. Sensitivity in samples with a viral load ≥105, ≥104 and ≥103 copies/mL was 100% [CI: 96.3%-100.0%], 96.5% [CI: 91.8%-98.7%] and 87.4% [CI: 81.3%-91.5%], respectively. All samples ≤103 copies/mL were antigen negative. The ratio of antigen concentration to viral load in samples ≥103 copies/mL was comparable in symptomatic and asymptomatic individuals (p=0.58). The proportion of RT-PCR positive participants with a high viral load (≥105 copies/mL) was not significantly higher in symptomatic than in asymptomatic participants (63.9% [CI: 54.9%-72.0%] vs. 51.9% [CI: 41.1%-62.6%], p=0.11), but the proportion of participants with a low viral load (<103 copies/mL) was significantly higher in asymptomatic than in symptomatic RT-PCR positive participants (35.4% [CI: 25.8%-46.4%] vs. 14.3% [CI: 9.0%-21.8%], p<0.01). CONCLUSIONS: Sensitivity and specificity in samples with a viral load ≥104 copies/mL was 96.5% and 100%. The correlation of antigen concentration with viral load was comparable in symptomatic and asymptomatic individuals.