Introduction

High levels of anti-SARS-CoV-2 neutralizing antibodies (nAb) have been shown to correlate with protection against severe infection with SARS-CoV-2 [1]. Messenger RNA (mRNA) vaccines elicit the production of nAb, but the ability to reliably detect nAb is limited to research laboratory settings. The FDA is in the process of deciding the optimal vaccination schedule for SARS-CoV-2 with a goal of simplifying the vaccination regimen [2]. Neither the CDC nor FDA currently recommend using commercially available antibody testing to assess post-vaccination immunity or to determine the timing of booster doses [3,4]. However, using commercially available assays that reliably predict nAb levels in clinical settings could allow for rapid and customizable decisions for patients regarding the optimal SARS-CoV-2 vaccination schedule and timing of booster doses.

Although no immunoassay has yet been thoroughly validated for post-vaccination serological testing, the Abbott Architect chemiluminescent microparticle immunoassay for SARS-CoV-2 IgG anti-S1/RBD (IgG-II CMIA) was used to study serological response after mRNA vaccination [5]. According to the manufacturer, values of >590.72 BAU/mL (>4160 AU/mL) correspond to a 95% probability (95% confidence interval [CI]: 78%–99%) of neutralization capacity against WA-1 [6]. The correlation of this assay’s results with the neutralizing capability of sera post-vaccination and the impact of patient clinical factors on this correlation have yet to be fully explored. The goal of this study is to measure the correlation between the SARS-CoV-2 IgG anti-S1/RBD and the sera neutralizing antibodies in a vaccinated US adult population, taking into account prior COVID-19 infection and individual clinical factors which may impact immune function.

Methods

Results

We obtained 127 sera samples from 100 unique subjects (Table 1). Log₁₀IgG values demonstrated a strong correlation with log₁₀IC₅₀ values (R² = 0.83, R_adj = 0.70, p < 0.0001) (Fig 1), and significantly predicted log₁₀ IC₅₀ values (beta = 0.56, p = 0) on univariable analysis. We found that a IgG of > 25,000 predicted a robust IC₅₀ of > 400 in all participants. A sensitivity analysis demonstrated an improvement in the adjusted R² value of the correlation to 0.80, suggesting that the linear model in Fig 1 is the most conservative estimate of the true correlation between IgG and nAb IC_50.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Correlation between Log₁₀ IC₅₀ and Log₁₀IgG.

Illustrated above is the correlation between the log₁₀ IC₅₀ values against pseudotyped WA-1 and the log10 of serum IgG anti-S1/RBD value. IgG anti-S1/RBD values beyond the cut-off of 25,000 AU/mL (>log₁₀ = 4) were treated as 25,000 AU/mL. R² = 0.83, R²_adj = 0.70, p-value < 0.001.

https://doi.org/10.1371/journal.pone.0289713.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Cohort demographics and morbidities.

https://doi.org/10.1371/journal.pone.0289713.t001

The following clinical factors had a significant impact on this correlation: COVID-positive status (beta = 0.38, p < 0.01), and GFR 0–30 mL/min/1.73 m² (beta = 0.26, p < 0.01). Age (p = 0.06) and female sex (p = 0.07) trended towards significance on univariable analysis. Malignancy, race, and ethnicity were not significant predictors. Log₁₀IgG (beta = 0.54, p < 0.01), COVID-positive status (beta = 0.38, p < 0.001) and GFR 0–30 mL/min/1.73 m² (beta = 0.19, p < 0.05) remained significant predictors of log₁₀IC₅₀ after adjusting for age and sex on multivariable analysis. Compared to using the log₁₀IgG results alone, including COVID status and GFR with the log₁₀IgG result provided a significantly better prediction model of log₁₀ IC₅₀ (partial F = 11.20, p < 0.0001) (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Comparison of prediction models for Log₁₀IC₅₀ with and without clinical variables.

https://doi.org/10.1371/journal.pone.0289713.t002

Discussion

Our results demonstrate a robust correlation between IgG anti-SARS-CoV-2 S1/RBD and anti-SARS-CoV-2 neutralization antibodies, suggesting that IgG anti-S1/RBD could be used to predict neutralizing antibody immune response. These findings may open avenues for further use of commercially available antibody testing in lower-resourced research settings, where rapid diagnostic testing is generally preferred [10]. Antibody testing will not provide a guarantee against infection, but could help guide clinical decisions, such as the timing and frequency of re-vaccination with a goal of preventing severe infections. Antibody testing is already being used sporadically in primary care settings, although guidelines and evidence for this use are limited [11,12]. Further research into appropriate cut-off values which indicate immune protection, and the impact of host factors will be important to refine and better guide rapid antibody testing in clinical settings.

Immunity is a dynamic process dependent on individual clinical factors and a simplified yearly vaccination schedule may not be appropriate for everyone, potentially resulting in an unacceptably high rate of breakthrough infections over time in vulnerable individuals [13]. Factoring prior COVID-19 infection status and an estimation of renal function based on glomerular filtration rate when using IgG anti-S1/RBD strengthened the prediction of anti-SARS-CoV-2 neutralizing antibodies, while a history of malignancy and age did not significantly impact the correlation between nAb IC₅₀ and IgG anti-S1/RBD. This study highlights the importance of accounting for host clinical factors when analyzing immune response.

The neutralization antibody target in this study was WA-1. Several studies have suggested that post-vaccination neutralizing antibody response may be lower against the newer circulating omicron variants than WA-1 [14,15]. Given the robust correlation between WA-1 neutralizing antibody titers and anti-SARS-CoV-2 IgG, further research into how well the IgG correlates with immune protection against severe disease due to omicron variants is warranted. The clinical interpretation and application of this correlation may also be improved by updating commercial platforms with IgG against newer variants.

This study investigated a limited number of clinical factors in a vaccinated population and will need to be cautiously generalized to the broader population. Although we found that a prior history of malignancy, race, ethnicity, and sex were not a clinically significant factors in this study, we had a limited sample size and ethnic distribution. Larger correlation and modeling studies on populations with additional clinical factors and immunocompromising conditions will need to be conducted to better personalize the IgG anti-S1/RBD results.

Conclusion

There is a robust correlation between sera IgG anti-S1/RBD and neutralizing antibody IC₅₀ titers, which is impacted by certain clinical variables such as prior COVID infection and renal function. Further understanding of the clinical factors that impact this correlation will be important for ongoing clinical or research applications of commercial antibody platforms. Commercially available antibody testing, such as the CMIA assay described in this study, can play an important role in lowering the cost and infrastructure barriers to assessing the levels of clinically active neutralizing antibodies. An understanding of neutralizing antibody titers may enable more nuanced patient conversations about SARS-CoV-2 immunity and frequency of vaccine booster doses, although further research will be needed to guide this use. We should leverage the tools we have available to refine rapid diagnostic antibody testing, lower barriers to SARS-CoV-2 research, and add clinical relevance to the optimal frequency of SARS-CoV-2 re-vaccination, especially in the presence of certain clinical factors.

Supporting information

Acknowledgments

We thank our laboratory staff who worked on our pseudotyping and commercial assay, as well as our subjects for donating their samples.