Longitudinal Qualitative and Quantitative Evaluation of SARS-CoV-2 Antibodies in Immunized Health Care Workers

Health care professionals (n = 72) volunteered serum samples (n = 339) before, during, and after vaccination with either BNT162b2 or mRNA-1273. Administration of the first vaccine dose was designated as day 0 and occurred between December 2020 and March 2021 (Supplemental Figure 1; see supplemental digital content containing 2 figures at https://meridian.allenpress.com/aplm in the February 2024 table of contents). Samples were collected −185 ± 72.7 days (average ± SD; range, −272 to −2; n = 43) before vaccination (T0) to determine baseline antibody reactivity. T1 samples were collected 18.0 ± 2.9 days after the first vaccine dose (range, 6–23; n = 68) and before the second vaccine dose. Additional samples were collected following the second vaccine dose: T2 (day 48.4 ± 7.0; range, 33–72; n = 68), T3 (day 147.4 ± 15.6; range, 90–180; n = 68), and T4 (day 255.0 ± 25.8; range, 194–280; n = 53). T5 samples were collected from participants who received a third booster vaccine dose (day 339.2 ± 27.8; range, 269–385; n = 41). This study was approved by the Colorado Multiple Institutional Review Board (Aurora, Colorado).

T0-T4 samples were tested by the VITROS anti–SARS-CoV-2 IgG qualitative assay (Ortho Clinical Diagnostics), which results in a signal-to-cutoff (S/C) for anti-spike IgG (S/C, reactive if >1.0). T1 samples were also tested by the ARCHITECT Abbott SARS-CoV-2 IgG assay, which qualitatively detects anti-nucleocapsid IgG (S/C, reactive if >1.4). T4 and T5 samples were tested by the VITROS anti–SARS-CoV-2 IgG quantitative assay (Ortho Clinical Diagnostics), which measures IgG against the S1 subunit of the spike protein. Quantitative results are reported in binding antibody units (reportable range, 2.0–200.0 BAU/mL [binding antibody units per milliliter]; reactive if >17.8), using calibration materials traceable to the WHO international standard.²  All tests were performed in a high-complexity Clinical Laboratory Improvement Amendments (CLIA)–certified laboratory according to manufacturer instructions for use and in a manner identical to patient testing. During this study, the clinical laboratory first transitioned from qualitative to quantitative testing and eventually discontinued all anti–SARS-CoV-2 antibody assays.

Data analysis was performed by GraphPad Prism for Mac OS version 9.3.1 and Microsoft Excel for Mac version 16.65. Nonparametric t tests were used to compare differences in IgG S/C values. Nonlinear least squares regression was performed to model partial linearity observed between anti-spike IgG S/C values and anti-S1 IgG BAU/mL values.

Longitudinal serum samples were obtained from study participants before (T0), during (T1), and after (T2–T4) the 2-dose vaccination schedule. Additional samples were collected from individuals who received a third booster vaccine dose (T5) (Supplemental Figure 1). Study participants spanned adulthood (45.2 ± 13.7 years; range, 22–74), were more representative of women than men (69% versus 31%), and were more likely to be vaccinated with BNT162b2 than mRNA-1273 (90.3% versus 9.7%).

Anti-spike IgG testing was performed to evaluate antibody response across the vaccination series (Supplemental Figure 2). Baseline T0 samples demonstrated nonreactivity (S/C = 0.1 ± 0.2) in 95.3% of individuals (n = 41 of 43). Two individuals with a history of COVID-19 had low-level IgG reactivity before vaccination (S/C = 2.3 and 2.7).

Vaccination with the first dose marked day 0. All individuals had seroconversion at T1 (S/C = 12.7 ± 5.1, n = 68). Testing of T1 samples for anti-nucleocapsid IgG (n = 63) demonstrated reactivity in 6 individuals (9.5%), 4 of whom had a known history of COVID-19 infection 20 or more days before T1. Individuals with anti-nucleocapsid IgG reactivity had increased anti-spike IgG reactivity at T1 (S/C = 18.5 ± 4.9), compared to individuals without anti-nucleocapsid IgG reactivity (S/C = 11.7 ± 4.2, P = .02). These data suggest that most of the study population had not yet been infected by SARS-CoV-2 before vaccination and that a history of COVID-19 was associated with the presence of anti-nucleocapsid IgG and an increased anti-spike IgG response following vaccine dose 1.

IgG reactivity further increased following the second vaccine dose at T2 (S/C = 20.1 ± 2.5, n = 68). A progressive decline in IgG reactivity was observed at approximately day 150 (T3, S/C = 15.1 ± 5.0, n = 68) and day 250 (T4, S/C = 9.3 ± 4.2, n = 53). Differences in S/C values were significant between sequential time points (T0–T4, P < .001). These data suggest a dose-dependent increase and time-dependent decline of IgG levels following vaccination.

IgG reactivity was compared in individuals vaccinated with BNT162b2 versus mRNA-1273. Similar IgG reactivity was seen at T1 (S/C = 13.5 ± 6.9 and 12.6 ± 5.1, P = .83) and T2 (S/C = 21.6 ± 2.2 and S/C = 19.9 ± 2.5, P = .17). However, the response to mRNA-1273 was more sustained at T3 (S/C = 20.5 ± 5.2 and 14.6 ± 4.7, P = .04) and T4 (S/C = 13.3 ± 2.2 and 8.9 ± 4.2, P = .01). These limited data may suggest that mRNA-1273 elicits a more durable humoral immune response than BNT162b2.

The first, and currently the only, quantitative SARS-CoV-2 antibody assay to receive emergency use authorization from the FDA uses calibrators traceable to WHO international standard reference materials to produce a value in BAU/mL. Testing of serum samples by both the qualitative anti-spike IgG and quantitative anti-S1 IgG assays (n = 79) revealed partial linearity of the qualitative assay for S/C values below 9 (Figure; R² = 0.94). However, S/C values greater than 9 were not linearly associated with the BAU/mL values and in most cases were beyond the measurable range of the quantitative assay (>200 BAU/mL).

Quantitative testing was also used to assess changes in IgG levels before and after the third booster dose. T4 samples demonstrated an antibody level greater than 200 BAU/mL in 55% of individuals (n = 29 of 53) and 75.6 ± 48.3 BAU/mL in the remaining 45%. T5 samples (n = 41) collected 64.2 ± 32.0 days following a third booster vaccine dose (day 339.2 ± 27.8 after the first dose) demonstrated IgG levels greater than 200 BAU/mL in 100% of individuals. These data may indicate waning IgG reactivity in part of the study cohort that is restored following a third booster vaccine dose.

Determining a correlate of immune protection from COVID-19 has been an elusive target.⁴  Widely available and scalable serologic assays for binding antibodies are the leading class of candidates to fulfill this need. However, such assays are highly variable, and results cannot be compared across assays, thereby limiting our progress and understanding of immune protection provided by antibodies.

To study this question effectively across multiple studies, SARS-CoV-2 antibody tests must be standardized. Currently, this can only be achieved by using the WHO International Standard for anti–SARS-CoV-2 immunoglobulin, which is now in its second iteration to address the impact of SARS-CoV-2 variants.³  Semiquantitative assays generally use non-WHO calibrators to produce numerical results, but even these results are difficult or impossible to compare across studies. Qualitative assays produce an S/C value based on antibody reactivity, which is then compared to an empirically determined threshold. Although S/C ratios may have numerical value, they cannot be reliably compared between different laboratories or assays and are subject to drift over time.

The purpose of this study was to evaluate vaccinated individuals longitudinally by commonly used qualitative and quantitative assays to understand IgG dynamics and to guide interpretation of such assays. Anti-nucleocapsid IgG testing indicated that most of the prospective study cohort was COVID-19–naive at the time of vaccination. Anti-spike IgG reactivity according to S/C values demonstrated 100% seroconversion after the first vaccine dose, peak IgG reactivity after the second vaccine dose, and progressive decline in IgG within 8 months after vaccination. Similar trends have been observed in other studies of both SARS-CoV-2 vaccination and infection.^17,18  Individuals receiving mRNA-1273 mounted a more durable antibody response than those receiving BNT162b2, which has been observed in other studies,¹⁹  but this study was not designed to compare IgG levels in response to vaccination comprehensively.

Comparison of the qualitative and quantitative assays revealed partial linearity at the low end of the qualitative assay. S/C values above 9 were not directly related to the IgG concentration, and in fact many were associated with values above the AMR of the quantitative test (>200 BAU/mL) for most individuals throughout the vaccination and postvaccination period. Therefore, although IgG reactivity trends in qualitative assays are logical and perhaps meaningful, proportional changes in IgG reactivity cannot be evaluated. These findings suggest that qualitative assays should generally be avoided when considering a correlate of protection from COVID-19 in healthy individuals, especially in the setting of vaccination. In select clinical contexts, for example, in those involving immunocompromised patients, changes in low S/C values may be helpful to gauge the humoral immune response.

There are several limitations to this study. First, the SARS-CoV-2 antibody assays studied use different capture antigens, were produced by the same manufacturer, do not necessarily demonstrate antibody neutralization activity, and may not be generally representative of all clinical assays. Second, although the observed trend in IgG reactivity across the vaccination schedule is consistent with that in other studies,¹⁸  this study evaluated a relatively small population of health care workers, favoring women and vaccination with BNT162b2. Third, the AMR of the quantitative assay could have been effectively extended by sample dilution, but this was not performed owing to discontinuation of SARS-CoV-2 antibody testing in our laboratory.

Until there is broader acceptance of WHO standards into manufactured assays and consensus guidance for anti–SARS-CoV-2 antibody levels as a correlate of protection, clinical laboratories and research studies should avoid the use of numerical values from qualitative assays. Expansion of WHO standards to include higher antibody concentrations and expansion of the AMR would increase our understanding of antibody dynamics associated with COVID-19 infection, vaccination, and treatment.