Neutralizing antibody correlate of protection against severe-critical COVID-19 in the ENSEMBLE single-dose Ad26.COV2.S vaccine efficacy trial

Abstract

Assessment of immune correlates of severe COVID-19 has been hampered by the low numbers of severe cases in COVID-19 vaccine efficacy (VE) trials. We assess neutralizing and binding antibody levels at 4 weeks post-Ad26.COV2.S vaccination as correlates of risk and of protection against severe-critical COVID-19 through 220 days post-vaccination in the ENSEMBLE trial (NCT04505722), constituting ~4.5 months longer follow-up than our previous correlates analysis and enabling inclusion of 42 severe-critical vaccine-breakthrough cases. Neutralizing antibody titer is a strong inverse correlate of severe-critical COVID-19, with estimated hazard ratio (HR) per 10-fold increase 0.35 (95% CI: 0.13, 0.90). In a multivariable model, HRs are 0.31 (0.11, 0.89) for neutralizing antibody titer and 1.22 (0.49, 3.02) for anti-Spike binding antibody concentration. VE against severe-critical COVID-19 rises with neutralizing antibody titer: 63.1% (95% CI: 40.0%, 77.3%) at unquantifiable [<4.8975 International Units (IU)50/ml], 85.2% (47.2%, 95.3%) at just-quantifiable (5.2 IU50/ml), and 95.1% (81.1%, 96.9%) at 90^th percentile (30.2 IU50/ml). At the same titers, VE against moderate COVID-19 is 32.5% (11.8%, 48.4%), 33.9% (19.1%, 59.3%), and 60.7% (40.4%, 76.4%). Protection against moderate vs. severe disease may require higher antibody levels, and very low antibody levels and/or other immune responses may associate with protection against severe disease.

Fig. 1. SARS-CoV-2 variants causing the severe-critical COVID-19 endpoints.

Variants are shown by calendar date of severe-critical COVID-19 occurrence and are broken out by geographic region (a, Latin America; b, USA; c, South Africa) and treatment assignment. Endpoint counts do not require having D1 and D29 antibody marker data (see the flowchart provided as Supplementary Fig. 2). As in Sadoff et al., “Reference” refers to the index strain (GenBank accession number: MN908947.3) harboring the D614G point mutation.

Fig. 2. D29 antibody marker level by COVID-19 outcome status (moderate COVID-19 case, severe-critical COVID-19 case, or non-case).

a 50% inhibitory dilution neutralizing antibody (nAb-ID50) titer, (b) anti-Spike IgG concentration, and (c) anti-RBD IgG concentration. Data points are from baseline SARS-CoV-2 seronegative per-protocol vaccine recipients. Violin plots contain interior box plots with upper and lower horizontal edges the 25th and 75th percentiles of antibody level and middle line the 50th percentile, and vertical bars the distance from the 25th (or 75th) percentile of antibody level and the minimum (or maximum) antibody level within the 25th (or 75th) percentile of antibody level minus (or plus) 1.5 times the interquartile range. Each side shows a rotated probability density (estimated by a kernel density estimator with a default Gaussian kernel) of the data. Positive response frequencies (Freq.) computed with inverse probability of sampling weighting. Positive response definitions: Spike IgG, IgG>10.8424 BAU/ml; RBD IgG, IgG>14.0858 BAU/ml. ULoQ: Spike IgG, 238.1165 BAU/ml; RBD IgG, 172.5755 BAU/ml. Positive response for nAb-ID50: D1 nAb-ID50 titer <LLOQ (LLOQ = 4.8975 IU50/ml) with detectable D29 nAb-ID50 ( ≥ LLOQ), or D1 nAb-ID50 > LLOQ with at least a fourfold increase in D29 nAb-ID50. ULoQ: ID50, 844.7208 IU50/ml. Moderate cases are baseline SARS-CoV-2 seronegative per-protocol vaccine recipients with the moderate COVID-19 endpoint (moderate COVID-19 with onset both ≥ 7 days post D29 and ≥28 days post-vaccination) up to 181 days post-D29 but not past data cut (July 9, 2021). Severe-critical cases are baseline SARS-CoV-2 seronegative per-protocol vaccine recipients with the severe-critical COVID-19 endpoint (severe-critical COVID-19 with onset both ≥ 7 days post-D29 and ≥28 days post-vaccination) up to 170 days post-D29 but not past data cut (July 9, 2021). Non-cases are baseline seronegative per-protocol vaccine recipients sampled into the immunogenicity subcohort with no evidence of SARS-CoV-2 infection up to the end of the correlates study period, which is up to 181 days post-D29 but not past data cut (July 9, 2021). BAU binding antibody units, IU international units, LLoQ lower limit of quantitation, Pos.Cut positivity cut-off, ULoQ upper limit of quantitation. Source data are provided as a Source Data file.

Fig. 3. Severe-critical COVID-19 risk and moderate COVID-19 risk by D29 antibody marker tertile.

Plots show covariate-adjusted cumulative incidence of (a, c severe-critical COVID-19 or b, d) moderate COVID-19 by Low, Medium, and High tertiles of D29 (a, b) 50% inhibitory dilution neutralizing antibody titer (nAb-ID50) or (b, d) anti-Spike IgG concentration in baseline SARS-CoV-2–seronegative per-protocol vaccine recipients. e Covariate-adjusted hazard ratios of severe-critical COVID-19 or of moderate COVID-19 across D29 antibody marker tertiles. Endpoint counts for (a–d) calculated by inverse probability of sampling D29 marker weighting. The overall p value is from a two-sided generalized Wald test of whether the hazard rate of the designated COVID-19 endpoint differed across the Low, Medium, and High subgroups. Analyses adjusted for baseline behavioral risk score and geographic region. BAU binding antibody units, CI confidence interval, FDR false discovery rate, FWER family-wise error rate, IU international units, Pt. Est. point estimate. Source data are provided as a Source Data file.

Fig. 4. Vaccine efficacy against severe-critical COVID-19 or against moderate COVID-19 by D29 antibody marker level.

Vaccine efficacy estimates against (a, c) severe-critical COVID-19 and against (b, d) moderate COVID-19 through 170 (severe-critical) or 181 (moderate) days post-D29 were obtained using a nonparametric implementation of the method of Gilbert et al.. Each point on the curve represents the estimated controlled vaccine efficacy at the given D29 antibody marker level: (a, b) 50% inhibitory dilution neutralizing antibody (nAb-ID50) titer and (c, d) anti-Spike IgG binding antibody concentration. Dotted lines indicate bootstrap pointwise 95% CIs. The green histograms are frequency distributions of D29 marker level, with maroon dots representing marker levels of individual cases. Analyses adjusted for baseline behavioral risk score and geographic region. Curves are plotted over the nAb-ID50 titer range from unquantifiable to the 90th percentile (30.2 IU50/ml) and over the Spike IgG concentration range from negative response to the 90th percentile (125 BAU/ml). The horizontal gray line is the overall vaccine efficacy against (a, c) severe-critical COVID-19 or against (b, d) moderate COVID-19 through 170 (severe-critical) or 181 (moderate) days post-D29, with the dotted gray lines indicating the 95% CIs. BAU binding antibody units, CVE controlled vaccine efficacy, IU international units, LLOQ lower limit of quantitation, nAb-ID50 50% inhibitory dilution neutralizing antibody. nAb-ID50 LLOQ = 4.8975 IU50/ml; Spike IgG positivity cutoff = 10.8424 BAU/ml. Source data are provided as a Source Data file.

Fig. 5. Exposure-proximal vaccine efficacy against severe-critical COVID-19 or against moderate COVID-19 by current antibody marker level.

Analyses were performed in baseline SARS-CoV-2 seronegative per-protocol vaccine recipients. Exposure-proximal vaccine efficacy estimates against (a, c) severe-critical COVID-19 and against (b, d) moderate COVID-19 through 170 (severe-critical) or 181 (moderate) days post-D29 by current antibody marker level were obtained using the method of Huang and Follmann, with “current” referring to the true underlying antibody marker level not subject to technical measurement error, in a hypothetical scenario in which the value was available from serum samples collected every day over the follow-up period (see “Methods”). Each point on the curve represents the vaccine efficacy at the given current antibody marker level: (a, b) 50% inhibitory dilution neutralizing antibody (nAb-ID50) titer and (c, d) anti-Spike IgG binding antibody concentration. The dashed lines are bootstrap pointwise 95% CIs. Analyses adjusted for baseline behavioral risk score and geographic region. Curves are plotted over the range from negative binding antibody response (or unquantifiable neutralizing antibody titer) to the 97.5th percentile of each current antibody marker level: Spike IgG, negative response to 352 BAU/ml; RBD IgG, negative response to 486 BAU/ml; nAb-ID50, unquantifiable to 43.4 IU50/ml. Positivity cutoffs: 10.8424 BAU/ml for Spike and 14.0858 BAU/ml for RBD; nAb-ID50 LLOQ = 4.8975 IU50/ml. BAU binding antibody units, CI confidence interval, IU international units, LLOQ lower limit of quantitation, nAb-ID50 50% inhibitory dilution neutralizing antibody. Source data are provided as a Source Data file.