Distinct receptor binding domain IgG thresholds predict protective host immunity across SARS-CoV-2 variants and time

Abstract

SARS-CoV-2 neutralising antibodies provide protection against COVID-19. Evidence from early vaccine trials suggested binding antibody thresholds could serve as surrogate markers of neutralising capacity, but whether these thresholds predict sufficient neutralising capacity against variants of concern (VOCs), and whether this is impacted by vaccine or infection history remains unclear. Here we analyse individuals recovered from, vaccinated or with hybrid immunity against SARS-CoV-2. An NT50 ≥ 100 IU confers protection in vaccine trials, however, as VOC induce a reduction in NT50, we use NT50 ≥ 1000 IU as a cut off for WT NT50 that would retain neutralisation against VOC. In unvaccinated convalescent participants, a receptor binding domain (RBD) IgG of 456 BAU/mL predicts an NT50 against WT of 1000 IU with an accuracy of 80% (95%CI 73-86%). This threshold maintains accuracy in determining loss of protective immunity against VOC in two vaccinated cohorts. It predicts an NT50 < 100 IU against Beta with an accuracy of 80% (95%CI 67-89%) in 2 vaccine dose recipients. In booster vaccine recipients with a history of COVID-19 (hybrid immunity), accuracy is 87% (95%CI 77-94%) in determining an NT50 of <100 IU against BA.5. This analysis provides a discrete threshold that could be used in future clinical studies.

Fig. 1. Correlation between RBD and NT50.

Two-tailed Spearman correlation between RBD and NT50. a Unvaccinated convalescent participants and WT-B NT50 (exact p-value < 2.2e-16). b Participants who had received only a primary vaccine series and WT-B.1.177.18 NT50 (exact p-value < 2.2e-16). c Participants who had received only a primary vaccine series and Beta NT50 (exact p-value = 1.381e-07). d Participants who had received one booster vaccine dose and had a history of infection and WT-B.1.177.18 NT50 (exact p-value = 3.99e-15). e Participants who had received one booster vaccine dose and had a history of infection and Omicron NT50 (exact p-value = 1.042e-06). Source data are provided as a source data file.

Fig. 2. Receiver operating characteristic curves.

a Receiver operating characteristic (ROC) curve demonstrating the performance of RBD in discriminating a wild type (WT-B) NT50 < 1000 IU in unvaccinated participants. b ROC curve demonstrating the performance of RBD in discriminating a wild type (WT-B.1.177.18) NT50 < 1000 IU in participants with one booster vaccine and a history of infection within the Omicron period. Source data are provided as a Source Data file.

Fig. 3. RBD and NT50 levels between groups.

a Boxplots showing RBD levels in each group of participants (unvaccinated n = 190, 2 vaccine dose n = 55, booster vaccine hybrid n = 70 biologically independent samples). b Boxplots showing NT50 against WT in each group of participants (unvaccinated n = 190, 2 vaccine dose n = 55, booster vaccine hybrid n = 70 biologically independent samples). c Boxplots showing NT50 against WT-B.1.177.18 and Beta in the two-dose vaccine dose group (n = 55 biologically independent samples). d Boxplots showing NT50 against WT-B.1.177.18 and Omicron in the booster vaccine hybrid group (n = 70 biologically independent samples. The centre of the boxplots represents the median, the bounds of the box represent the interquartile range, and the whiskers represent the minimum and maximum values no more than 1.5 times the interquartile range. Where the minima and maxima lie outside the whiskers they are represented as dots. Source data are provided as a Source Data file.