Development and use of a method based on the anti-N reactivity of longitudinal samples to better estimate SARS-CoV-2 seroprevalence in a vaccinated population

BackgroundEmerging evidence suggests that COVID-19 vaccination decreases the sensitivity of anti-nucleocapsid (N) serologies, making them less reliable to assess recently-acquired infections. We therefore developed and tested a new approach based on the ratio of the anti-N absorbance of longitudinal samples to overcome this limitation. MethodsPreviously vaccinated repeat plasma donors provided at least one pre-infection (reference) and one post-infection (test) sample. All samples were tested using an in-house anti-N ELISA. Seropositivity was determined based on the ratio between the anti-N absorbance of the test and reference samples. The ratio approach was tested in a real-world setting during three cross-sectional serosurveys carried out among plasma donors in Quebec, Canada. ResultsUsing a cut-off ratio of 1.5, the approach had a sensitivity of 95.2% among the 248 previously vaccinated and infected donors compared with 63.3% for the conventional approach. When tested in a real-world setting, the ratio-based approach yielded an adjusted seroprevalence of 27.4% (95% confidence interval [CI]=23.8%-30.9%) at the latest time point considered, compared to 15.1% (95% CI=12.2%-18.0%) for the conventional approach. ConclusionsThis article describes a new and highly-sensitive approach that captures a significantly greater proportion of vaccinated individuals with a recent history of SARS-CoV-2 infection.