Fig. 4

Anti-SARS-CoV-2 immunoglobulin G (IgG) response in Binding Antibody Units (BAU) per millilitre (ml) after first (open circles) and second (filled circles) immunisation with the vector vaccine AZD1222 (green), the messenger ribonucleic acid (mRNA)-based vaccine BNT162b2 (blue) and after a heterologous vaccination scheme, starting with AZD1222, followed by an mRNA-based vaccine boost (BNT162b2 or mRNA-1273; red) with regard to the detection of virus-neutralising antibodies (VNA). The latter were measured in a Vero-cell-based neutralisation test (cVNT) using the SARS-CoV-2 variant of concern B.1.1.7 (alpha). Cut-off values for positivity of the anti-trimeric spike (S) IgG assay (A), anti-S IgG assay (B) and anti-receptor binding domain (RBD) IgG assay (C), respectively, and the cVNT cut-off value for the presence of VNA are indicated by black dashed lines. The Spearman correlation coefficients of log(reciprocal titre) were calculated with 0.86, 0.86 and 0.88, respectively. The probability of detecting VNA at a given BAU/ml in the anti-SARS-CoV-2 IgG assays was calculated by logistic regression (D–F): VNA were present in 95% of samples when IgG concentrations of 886 BAU/ml (anti-trimeric S IgG), 323 BAU/ml (anti-S IgG) and 448 BAU/ml (anti-RBD IgG), respectively, were measured (green dashed lines; 95% confidence intervals (CI) 59.4 to 99.6%). Vertical black dashed lines represent the threshold values set by the manufacturers of the antibody assay; red dashed lines represent the BAU/ml concentrations (anti-trimeric S IgG: 350 BAU/ml; anti-S IgG: 119 BAU/ml; anti-RBD IgG: 174 BAU/ml) with a 50% probability of VNA detection. The distribution of the cVNT titres, the medians, and the 95% CIs between the three plotted thresholds (dashed black, red and green lines in A–F) are shown (G–I)