Calibration of Two Validated SARS-CoV-2 Pseudovirus Neutralization Assays for COVID-19 Vaccine Evaluation

Abstract

Vaccine-induced neutralizing antibodies (nAbs) are key biomarkers considered to be associated with vaccine efficacy. In United States Government-sponsored phase 3 efficacy trials of COVID-19 vaccines, nAbs are measured by two different validated pseudovirus-based SARS-CoV-2 neutralization assays, with each trial using one of the two assays. Here we describe and compare the nAb titers obtained in the two assays. We observe that one assay consistently yielded higher nAb titers than the other when both assays were performed on the World Health Organization’s anti-SARS-CoV-2 immunoglobulin International Standard, COVID-19 convalescent sera, and mRNA-1273 vaccinee sera. To overcome the challenge this difference in readout poses in comparing/combining data from the two assays, we evaluate three calibration approaches and show that readouts from the two assays can be calibrated to a common scale. These results may aid decision-making based on data from these assays for the evaluation and licensure of new or adapted COVID-19 vaccines.

Figure 1:. Distributions of ID50 (Panel A) and ID80 (Panel B) titers measured by the Duke and Monogram labs of convalescent patient samples, vaccine recipient samples, and the WHO International Standard sample.

For the convalescent samples, the assay lower limit of detection (LLoD) is 1:20 for Duke and 1:40 for Monogram. For the vaccine samples collected at Day 29 and Day 57, 4 weeks post the first and the second Moderna doses, respectively, and the WHO IS sample, the assay LLoD is 1:10 for Duke and 1:40 for Monogram. Circle-shape points indicate titers > LLoD; triangle-shape points indicate titers ≤ LLoD.

Figure 2:. Scatterplots of calibrated neutralizing antibody readouts demonstrating performance of the three calibration approaches using vaccine recipient samples collected at Day 29 (turquoise) and Day 57 (orange), 4 weeks post the first and the second mRNA-1273 vaccine doses, respectively.

The arithmetic mean-based calibration factor was used in Approach 1. The concordance correlation coefficient (CCC) and 95% confidence intervals indicate the level of agreement between the x- and y-axis values.

1. cID50, cID80: ID50, ID80 titers calibrated to the WHO International Standard, expressed in International Units per ml (IU/ml) (Approach 1).

2. cBIVN-ID50, cBIVN-ID80: ID50, ID80 titers from each lab calibrated to a common scale using an independent pool of clinical samples, assuming a bivariate normal distribution for the readouts from the two labs (Approach 2).

3. cREG-ID50, cREG-ID80: ID50, ID80 titers from the non-reference lab calibrated to the reference lab using an independent pool of clinical samples, based on regressing titers from the reference lab on the non-reference lab (Approach 3).