Comparison of Antibody Response Durability of mRNA-1273, BNT162b2, and Ad26.COV2.S SARS-CoV-2 Vaccines in Healthcare Workers

Introduction

To date, more than 820,000 Americans have died of COVID-19. Vaccination with FDA approved/authorized COVID-19 vaccines is imperative to control the pandemic. A number of studies have shown robust similar or higher early anti-spike (anti-S), anti-S-receptor binding domain (RBD), and neutralizing antibody (nAb) responses with mRNA-1273 than BNT162b2 and higher responses with both mRNA vaccines than Ad26.COV2.S.^1-7 The pattern for early vaccine effectiveness (VE) is similar, mRNA-1273 being about 90-95%, BNT162b2 about 90%, and Ad26.COV2.S about 70-85%, with lower and higher percentages being against infection and severe disease, respectively.^{5, 8-11}

Individual durability studies for these vaccines have shown waning immunity over time, as measured by decreasing antibody titers and VE (breakthrough infections).^12-34 The literature for comparative vaccine antibody durability, however, is limited. One study showed marked decreases in antibodies over 6-8 months for mRNA-1273 and BNT162b2 which contrasted with relatively durable responses for Ad26.COV2.S; because responses to Ad26.COV2.S were lower early post-vaccination, antibody responses were comparable at 6-8 months.³ Another comparative antibody durability study showed 3-fold decreases in anti-S and anti-S-RBD antibody titers from peak post-vaccination (14-119 days) to >120 days with mRNA-1273 and BNT162b2, but mRNA-1273 titers were 3-fold higher at both time points.³⁵ Numerous comparative VE durability studies, occurring at different pandemic stages with differing prevalent strains, showed decreases to 80-95% at 3-7 months for mRNA-1273, 65-90% for BNT162b2, and 60-70% for Ad26.COV2.S.^{5, 10, 11, 35-38}

Due to the dearth of comparative vaccine durability studies, particularly for antibody responses, we sought to compare medium range durability (at 8.4 months) of antibody responses in a cohort of healthcare workers (HCWs) in the Bassett Healthcare Network (NY, USA) vaccinated with FDA-approved/authorized COVID-19 vaccines.

Our primary objective was to compare medium range (median 8.4 months) SARS-CoV-2 anti-S antibody titers in HCWs vaccinated with mRNA-1273, BNT162b2, or Ad26.COV2.S, or unvaccinated, or vaccine-boosted. Our secondary objective was a similar comparison for IgG-S-RBD competitive antibody blocking ELISA pseudoneutralization semiquantitative inhibition percentages against the USA-WA1/2020 strain, at a median of 8.4 months. Tertiary objectives included: comparisons of COVID-19 infection rates based on immunological parameters (IgG nucleocapsid (anti-N) seroconversion from an early pandemic baseline)) and clinical parameters (clinical infections based on self-reported survey and PCR swab positivity) and VE percentages (against COVID-19 infection); comparisons of antibody responses stratified by age, number of comorbidities, immunosuppression, time since primary vaccination, and COVID-19 infection; and assessments of correlation between anti-S and pseudoneutralization results.

Methods

Participants, Data Collection, and Interventions

HCWs were eligible for this follow-up study if they consented to storage of their plasma samples for future research and were available for blood drawing regardless of vaccination or infection status. Of 2056 HCWs, 1806 were invited to participate and 903 (50%) consented and completed the study’s survey. Among them, 653 (72%) had blood drawn September 23 – November 16, 2021; 238 did not have blood drawn because of scheduling/logistics challenges, 12 declined, and 6 were excluded because of incomplete primary vaccination, yielding 647 HCWs for inclusion (eFigure 1), stratified into 4 groups: (1) mRNA-1273-vaccinated; (2) BNT162b2-vaccinated; (3) Ad26.COV2.S-vaccinated; and (4) unvaccinated. A fifth group—boosted (regardless of vaccine type)—was added upon FDA/CDC recommendations for boosting of HCWs in October 2021. Primary vaccination was defined as two mRNA vaccine doses or one Ad26.COV2.S dose; “boosted” was defined as primary vaccination plus an additional vaccine.

• Download figure
• Open in new tab

Figure 1. Antibody Distribution by COVID-19 Vaccine Type

Boxplots showing antibody levels by study group at a median of 8.4 months since complete vaccination. The bar inside each box represents the median and the diamond represents the mean antibody level. A, Difference in unadjusted log-anti-S antibodies by vaccine type. B, Difference in unadjusted pseudoneutralization signal inhibition percentages by vaccine type. Note: Sixty percent (6/10) of the unvaccinated HCWs reported a COVID-19 infection prior to antibody testing.

The survey included questions on demographics, medical history (comorbidities predisposing to worse COVID-19 outcome, immunosuppression), COVID-19 infection history (timing of infection and swab testing, symptoms, and severity), and COVID-19 vaccine status (type, primary vaccination timing, and booster type/timing if done). To account for time lapse between survey completion and blood draw, COVID-19 vaccination/infection status was updated upon blood drawing.

Self-reported COVID-19 infection was based on survey response. Laboratory-confirmed infection was based on medical record evidence of a positive PCR or anti-N antibody test. COVID-19-related hospitalizations were confirmed by medical record review. Breakthrough infection rates were calculated from 14 days after primary vaccination through the day of blood drawing. Unvaccinated HCWs were assessed for infection any time from 241 days prior to their blood draw (the median time from vaccination in the overall cohort [255 days] minus 14 days to the day of blood drawing).

Anti-S antibody tests were performed by Bassett Medical Center’s Laboratory using the Abbott AdviseDx SARS-CoV-2 IgG II semi-quantitative assay with dilution if necessary (Abbott Laboratories, Abbott Park, IL). Anti-N antibody tests were also performed by Bassett’s laboratory using the SARS-CoV-2 IgG Abbott Architect assay (Abbott Laboratories, Abbott Park, IL).⁴⁰

Neutralization antibodies to spike protein RBD were measured using ELISA-based pseudoneutralization (competitive antibody blocking) COVID-19 assays (cPASS™, GenScript, Piscataway, NJ) using USA-WA1/2020 strain genetic background and were performed by Biodesix (Boulder, CO). The α-RBD nAb test measures a subset of antibodies that can block the interaction between the RBD on the SARS-CoV-2 spike protein and the human host receptor angiotensin-converting enzyme 2 (ACE2). Methods for the cPASS surrogate virus neutralization test (SVNT) were previously published, validated, and shown to be 100% sensitive and specific when compared to gold standard, plaque reduction neutralization test (PRNT), with qualitative analysis results 100% in agreement.^41-44 The nAb assay readout was percent signal inhibition by neutralizing antibodies, calculated as the optical density (OD) value of the sample relative to the OD of the negative control subtracted from one.⁴² The cPASS assay has an EUA for qualitative interpretation of results,⁴¹ but here we prioritize presenting numeric semi-quantitative percent signal inhibition to compare between-group percentages.

Outcome variables were compared by five prespecified characteristics thought to impact immune responses to COVID-19 vaccination: age, number of comorbidities, immunosuppression, time from completion of primary vaccination (days), and prior COVID-19 infection.

Results

Discussion

This COVID-19 vaccine comparative immune response durability cohort study showed that medium range antibody responses at a median of 8.4 months after primary vaccination were higher after vaccination with mRNA-1273 than BNT162b2. Ad26.COV2.S antibody responses were lower but not significantly different than with mRNA-1273 and BNT162b2. Specifically, both median IgG spike protein (anti-S) antibody titers and median anti-USA-WA1/2020 strain RBD pseudoneutralization inhibition percentages were significantly higher with mRNA-1273 than BNT162b2 (1539.5 AU/mL vs. 751.2 AU/mL, p<.001; and 90.9% vs. 77.2%, p<.001). Median anti-S and pseudoneutralization antibodies were dramatically higher in boosted HCWs (31898.8 AU/mL, p<.001 for both) and much lower in unvaccinated HCWs (in whom such antibodies were not nil (113.4 AU/mL and 40.1%, respectively, p<.03 for both), suggesting natural infections (“herd immunity”)). Anti-S titers and pseudoneutralization percentages had high between-assay correlation (ρ=.947) as previously reported.⁴⁴ Differences in antibody responses between mRNA-1273 and BNT162b2 persisted in adjusted analyses. Estimated vaccine effectiveness (VE) against COVID-19 infection was similarly high with both mRNA vaccines (87-89%), but not significantly different than Ad26.COV2.S (33%), although conclusions are limited by high statistical uncertainty due to low event rates, small sample sizes for Ad26.COV2.S and boosted groups, and insufficient post-boosting time to assess its VE.

Prespecified subgroup analyses, restricted to unadjusted comparisons between mRNA-1273 and BNT162b2 groups, found no significant differences in anti-S titers or pseudoneutralization percentages by number of comorbidities, immunosuppression status (in contrast with others’ findings for COVID-19 mRNA vaccines),^{17, 45} or time since primary vaccination.

Conversely, stratification by age showed lower anti-S antibody titers and pseudoneutralization percentages in older (≥ 50 years old) vs. younger (< 50 years old) HCWs with BNT162b2 (anti-S: 662.3 vs. 989.4 AU/mL, p=.002; pseudoneutralization: 68.8% vs. 84.7%, p<.001) but not mRNA-1273 (anti-S: 1414.8 vs. 1668.9 AU/mL, p=.40; pseudoneutralization: 90.3% vs. 92.0%, p=.19). Other studies have shown a similar age effect for the two mRNA vaccines^{7, 13} suggesting that primary vaccination with mRNA-1273 rather than BNT162b2 should be prioritized for older individuals. We did not complete subgroup analyses for VE precluding assessment of such an age effect.

Stratification by prior COVID-19 infection status also showed significant effects, with higher anti-S antibody titers and pseudoneutralization percentages in HCWs with previous COVID-19 infection with mRNA-1273 (anti-S: 3604.4 vs. 1487.4 AU/mL, p<.001; pseudoneutralization: 95.8% vs. 90.4%, p<.001) and BNT162b2 (anti-S: 5838.8 vs. 692.4 AU/mL, p<.001; pseudoneutralization: 96.0% vs. 75.8%, p<.001). Notably, the increased antibody responses seen with prior COVID-19 infection were more pronounced with BNT162b2 than mRNA-1273, particularly for anti-S antibodies (8.4-fold vs. 2.4-fold; and 1.27-fold vs. 1.06-fold, respectively). Other studies have also shown higher antibody responses in previously COVID-19 infected vs. naive individuals with both mRNA vaccines.^{4, 46, 47}

This is one of the first comparative COVID-19 vaccine immunologic durability studies; our results are consistent with results from two other studies (below).^{3, 35} A synopsis of the current literature shows many more studies reporting antibody and VE results for individual vaccines rather than comparative analyses.

Individual mRNA-1273 antibody studies show high anti-S, anti-S-RBD, and neutralizing antibody titers post-vaccination that were durable for 3-6 months earlier in the pandemic and increased in immunosuppressed adults upon boosting.^12-15 VE studies show high protection against infection and severe disease after vaccination (95%), but later in the pandemic when Delta predominated, decreased by 36-46% between 7-9 and 12 months.^{27, 28} A prison outbreak study when Delta predominated, found VE of 56.6% against infection (80.5% in those previously infected) and 84.2% against symptomatic infection.⁴⁸

BNT162b2 antibody studies show high antibody titers (same specific antibodies as above) post-vaccination that wane by 2-6 months (more rapidly in immunosuppressed adults), and increase with boosting.^16-24 VE studies show high protection against infection with durability for 6 months (90-95%) earlier in the pandemic when Alpha and Beta variants predominated, but with waning to as low as 20% later when Beta and Delta variants predominated; VE against severe disease (hospitalization and death) remained high (>90%) throughout the pandemic.^{29-31, 34} VE against infection, severe infection, and mortality in Israel when Delta predominated was much higher with boosting.^{49, 50}

Ad26.COV2.S antibody studies show high titers post-vaccination (albeit lower than with both mRNA vaccines) that are relatively durable for 8 months (only 1.8-fold decrease).^{25, 26} VE studies show protection against infection and severe disease of 64-85% after vaccination.^{33, 51}

Comparative antibody studies, soon after primary vaccination, show titers that are similar or higher for mRNA-1273 vs. BNT162b2 and higher with either vs. Ad26.COV2.S.^1-7 One of two comparative antibody durability studies showed marked decreases in anti-S-RBD and nAb titers for mRNA-1273 and BNT162b2 at 6-8 months. In that study, Ad26.COV2.S’s initially lower titers were more durable such that by 6-8 months titers were similar between the three vaccines.³ Another study showed about 3-fold higher peak (14-119 days) anti-S and anti-S-RBD titers with mRNA-1273 vs. BNT162b2 with similar 3-fold decreases after 4 months, such that after 4 months, both antibody levels remained 3-fold higher with mRNA-1273.³⁵ Likewise, our results show significantly higher anti-S titers and pseudoneutralization percentages at 8.4 months for mRNA-1273 vs. BNT162b2, and higher (but not significantly different) results for mRNA-1273 and BNT162b2 vs. Ad26.COV2.S. However, we cannot make conclusions about relative diminution of these titers as we did not complete sequential antibody assays.

Comparative VE studies show higher protection against infection and hospitalization for mRNA-1273 (90-95%) and BNT162b2 (90%) than for Ad26.COV2.S (70-85%) post-vaccination.^{5, 8-11} Comparative VE durability studies against infection and severe disease at different pandemic times show waning to 80-95% for mRNA-1273, 65-90% by 3-7 months for BNT162b2, and 60-70% for Ad26.COV2.S;^{5, 10, 11, 36-38} the higher ranges were for VE against more severe disease for all three vaccines. In contrast, our study showed equivalently high VE against infection at 8.4 months (87-89%) with mRNA-1273 and BNT162b2 despite significantly higher anti-S titers and pseudoneutralization percentages with mRNA-1273 than BNT162b2; VE was higher but not significantly different with both mRNA vaccines than with Ad26.COV2.S (33%).

Key study limitations included the following: (1) Our study was observational. We addressed predicted confounders by adjusting for covariates likely to impact results; however, unmeasured confounders (e.g., COVID-19 prevention behavior, vaccine preferences, varying infection rates/exposures, and prevalent variants) may have influenced results. Lack of randomization required standardization of the period during which infection rates and VE were quantified. (2) We did not complete peak antibody assays post-primary vaccination precluding comparisons over time. (3) Analyses of Groups 3-5 were limited by small sample sizes. (4) Significant baseline variable group differences included immunosuppression (higher in Group 5, possibly lowering antibody responses in that group), prior COVID-19 infection rates (higher in Groups 3 and 4, possibly increasing antibody responses in those groups), and time from primary vaccination (higher in Groups 1 and 2, possibly lowering antibody responses in those groups). (5) The study design (with addition of the boosted group) and statistical analysis plan (primary analysis focused on subpopulation of surveyed participants) were modified after initiation of enrollment; however, these changes were prespecified prior to seeing primary outcome data. (6) The boosted group included varying primary vaccination-boosting combinations although BNT162b2-BNT162b2 was predominant (67.9%); and insufficient follow-up time to assess boosting’s effects on VE. (7) Antibody inhibition pseudoneutralization assays were completed rather than gold standard PRNT; however, good correlation has been reported for these tests.⁴⁴

Conclusions

In this COVID-19 vaccines comparative immunology durability cohort study, at a median of 8.4 months after primary vaccination, IgG spike protein antibody titers and pseudoneutralization inhibition percentages against SARS-CoV-2 USA-WA1/2020 were significantly higher with mRNA-1273 than with BNT162b, both of which were not significantly higher than with Ad26.COV2.S. Vaccine effectiveness against infection remained high with both mRNA vaccines (87-89%), also being higher but not significantly different than with Ad26.COV2.S (33%). Due to statistical uncertainty in the small Ad26.COV2.S group, clinically meaningful outcomes differences cannot be excluded. Boosting led to dramatic increases in anti-S and pseudoneutralization responses.

Author Contributions

Concept and design: Brunner, Freilich, Victory, Lasher, Fink, Gadomski, Shah

Drafting of the manuscript: Brunner, Freilich, Victory, Lasher, Fink, Gadomski

Critical revision of the manuscript for important intellectual content: Brunner, Freilich, Victory, Lasher, Fink, Gadomski, Shah

Statistical analysis: Krupa, Scribani, Jenkins

Obtained funding: Freilich, Shah

Administrative, technical, or material support: Brunner, Freilich, Victory, Cross, Bush, Shah, Lasher, Fink, Peek

Supervision: Brunner, Freilich, Victory, Lasher, Fink, Gadomski

Conflict of Interest Disclosures

There are no conflicts of interest to disclose.

Funding/Support

Funding for this study was provided by The Stephen C. Clark Research Endowment, The E. Donnell Thomas Resident Research Endowment and The Marion Cooper Fund.

Role of Funder

None