Fig. 2

Anti-S antibody titres and dynamics in vaccinated and unvaccinated subjects previously infected by SARS-CoV-2. a–d Observed antibody titres in unvaccinated and vaccinated subjects infected by SARS-CoV-2 and tested in May 2020, November 2020 and June 2021 by DiaSorin (vaccinated n=38, P < 0.0001 from November 2020 to June 2021; unvaccinated n=33, P = 0.0063 from November 2020 to June 2021) and micro-neutralisation assays (vaccinated n=38, P < 0.0001 from November 2020 to June 2021; unvaccinated n=32, P = 0.0053 from November 2020 to June 2021). The horizontal line represents the median, the vertical line represents the 95% confidence intervals. e–h Observed individual-level paired antibody titres in subjects infected by SARS-CoV-2 and tested in May 2020, November 2020 and June 2021. In June 2021, 59.4% (19 out of 32 unvaccinated subjects, 95% CI 40.6–76.3%) and 6.3% (2 out of 32 unvaccinated individuals, 95% CI 0.8–20.8%) had antibodies more than 15 months post infection according to DiaSorin and micro-neutralisation, respectively. Subjects with increasing titres are coloured in green, while subjects with a negative result in June 2021 are presented in red. i, j Estimated antibody decay rate distributions calculated among the unvaccinated subjects infected by SARS-CoV-2 in February/March 2020, tested in May 2020, November 2020 and June 2021. Each bar represents the frequency of each slope (in units of days), calculated on the logarithm of individual-level sequential titres. We estimated a median half-life of 214 (95% CI 168–288) days and 174 (95% CI 146–202) days for the antibodies detected by the DiaSorin and micro-neutralisation assays, respectively. To estimate the median half-life, only subjects with no doubling antibodies from May 2020 were considered. Asterisks indicate *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Statistical significance of antibody levels was evaluated by Friedman test followed by Dunn’s multiple comparisons test