Table 5.
Scorecard for ranking D29 antibody marker performance in each of three categories of immune correlate-quality criteria
| Category 1: correlate of risk (CoR) | Category 2: correlate of protection (CoP): VE modification | Category 3: CoP: VE mediation | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
HR per SD (Cox, quant.) |
HR High vs. Low tertile (Cox) | Cat. 1 Mean rank | Fold-increase in CVE1 (Cox) | Fold-increase in CVE1 (NP) | E-val. Marg. RR Pt. Est. | Cat. 2Mean rank | Proportion mediated | Cat. 3 Mean rank | |||||||||||
| Comp. | D29 Marker | COVID-19 endpoint | Pt. Est. (95% CI) | Rank | Pt. Est. (95% CI) | Rank | Fold-increase | Rank | Fold-increase | Rank | E-val. | Rank | Pt. Est. (95% CI) | Rank | 95% LCL | Rank | |||
| Spike IgG (BAU/ml) | Sev-crit | 0.83 (0.59, 1.17) | 1 | 0.62 (0.27, 1.42) | 1 | 1* | 1.7 | 1 | 1.9 | 2 | 2.6 | 1 | 1.33* | 0.243 (−0.214, 0.700) | 2 | −0.214 | NA | 2 | |
| A | Spike IgG (BAU/ml) | Mod | 0.83 (0.69, 1.01) | 1 | 0.64 (0.40, 1.01) | 2 | 1.5 | 1.7 | 1 | 2.1 | 1 | 2.5 | 2 | 1.33* | 1.03 (−0.022, 2.08) | 1 | −0.022 | NA | 1* |
| RBD IgG (BAU/ml) | Sev-crit | 0.90 (0.63, 1.30) | 2 | 0.67 (0.28, 1.57) | 2 | 2 | 1.3 | 2 | 1.8 | 2 | 2.4 | 2 | 2 | 0.179 (−0.282, 0.639) | 2 | −0.282 | NA | 2 | |
| A | RBD IgG (BAU/ml) | Mod | 0.80 (0.65, 0.97) | 1 | 0.62 (0.39, 0.99) | 1 | 1* | 1.9 | 1 | 1.9 | 1 | 2.6 | 1 | 1* | 0.852 (−0.051, 1.76) | 1 | −0.051 | NA | 1* |
| nAb-ID50 (IU50/ml) | Sev-crit | 0.59 (0.36, 0.95) | 1 | 0.21 (0.07, 0.67) | 1 | 1* | 3.2 | 1 | 7.6 | 1 | 9.0 | 1 | 1* | 0.286 (0.085, 0.487) | 2 | 0.085 | NA | 2 | |
| A | nAb-ID50 (IU50/ml) | Mod | 0.73 (0.58, 0.91) | 2 | 0.43 (0.25, 0.75) | 2 | 2 | 2.0 | 2 | 1.7 | 2 | 4.0 | 2 | 2 | 0.505 (0.008, 1.00) | 1 | 0.008 | NA | 1* |
| Spike IgG (BAU/ml) | Mod to sev-crit | 0.83 (0.69, 1.00) | 3 | 0.63 (0.40, 0.99) | 3 | 3 | 1.7 | 3 | 2.1 | 1 | 2.5 | 3 | 2.33 | 0.723 (0.051, 1.39) | 1 | 0.051 | 2 | 1.5* | |
| B | RBD IgG (BAU/ml) | Mod to sev-crit | 0.80 (0.66, 0.98) | 2 | 0.62 (0.39, 0.98) | 2 | 2 | 1.8 | 2 | 1.9 | 2 | 2.6 | 2 | 2 | 0.603 (0.015, 1.19) | 2 | 0.015 | 3 | 2.5 |
| nAb-ID50 (IU50/ml) | Mod to sev-crit | 0.71 (0.57, 0.88) | 1 | 0.40 (0.23, 0.69) | 1 | 1* | 2.1 | 1 | 1.9 | 2 | 4.3 | 1 | 1.33* | 0.415 (0.088, 0.742) | 3 | 0.088 | 1 | 2 | |
| Spike IgG (BAU/ml) | Sev-crit | 0.83 (0.59, 1.17) | 2 | 0.62 (0.27, 1.42) | 2 | 2 | 1.7 | 2 | 1.9 | 2 | 2.6 | 2 | 2 | 0.243 (−0.214, 0.700) | 2 | −0.214 | 2 | 2 | |
| C | RBD IgG (BAU/ml) | Sev-crit | 0.90 (0.63, 1.30) | 3 | 0.67 (0.28, 1.57) | 3 | 3 | 1.3 | 3 | 1.8 | 3 | 2.4 | 3 | 3 | 0.179 (−0.282, 0.639) | 3 | −0.282 | 3 | 3 |
| nAb ID50 (IU50/ml) | Sev-crit | 0.59 (0.36, 0.95) | 1 | 0.21 (0.07, 0.67) | 1 | 1* | 3.2 | 1 | 7.6 | 1 | 9.0 | 1 | 1* | 0.286 (0.085, 0.487) | 1 | 0.085 | 1 | 1* | |
1Fold-increase calculated on the [1-VE(unquantifiable or negative)]/[1-VE(90th percentile)] scale.
*Antibody marker(s) with the best performance, within each comparison, within each category.
Bold font indicates the three categories of criteria and the mean D29 antibody marker rank within each category.
Baseline covariates adjusted for: Baseline risk score, geographic region (Latin America, South Africa, United States). Maximum failure event time 181 days (moderate COVID-19), 170 days (severe-critical COVID-19), or 181 days (moderate to severe-critical COVID-19) post D29. Cases were counted starting 7 days post D29. All serological assay readouts were expressed in WHO International Standard units (see “Methods”). The Proportion mediated is not a true proportion in that it can take values outside of the interval [0, 1]. Proportion mediated = 1 is equivalent to Non-marker-mediated VE = 0% and Proportion mediated = 0 is equivalent to Non-marker-mediated VE = Overall VE. BAU binding antibody units, Cat. category, CI confidence interval, Comp. comparison, E-val. E-value, HR hazard ratio, IU international units, LCL lower confidence limit, Marg. marginalized, mod. moderate, nAb-ID50 50% inhibitory dilution neutralizing antibody titer, NP nonparametric, Pt. Est. point estimate, RR risk ratio, SD standard deviation, sev-crit severe-critical.