key: cord-0844739-64if6424
authors: Lind, M.; Robertson, A.; Silva, J.; Warner, F.; Coppi, A. C.; Price, N.; Duckwall, C.; Sosensky, P.; Di Giuseppe, E. C.; Borg, R.; Fofana, M.; Ranzani, O. T.; Dean, N. E.; Andrews, J. R.; Croda, J.; Iwasaki, A.; Cummings, D. A. T.; Ko, A.; Hitchings, M.; Schulz, W. L.
title: Effectiveness of Primary and Booster COVID-19 mRNA Vaccination against Infection Caused by the SARS-CoV-2 Omicron Variant in People with a Prior SARS-CoV-2 Infection
date: 2022-04-20
journal: nan
DOI: 10.1101/2022.04.19.22274056
sha: a944af193a8ae867d73fd3b984b5ac39efe8264a
doc_id: 844739
cord_uid: 64if6424

Importance: The benefit of primary and booster vaccination in people who experienced prior SARS-CoV-2 infection remains unclear. Objective: To estimate the effectiveness of a primary (two-dose) and booster (third dose) vaccination against Omicron infection among previously infection people. Design: Test-negative case-control study. Setting: Yale New Haven Health System facilities serving southern Connecticut communities. Participants: Vaccine eligible people who received SARS-CoV-2 RT-PCR testing between November 1, 2021, and January 31, 2022. Exposure: COVID-19 mRNA primary and booster vaccination. Main Outcomes and Measures: We conducted two analyses, each with an outcome of Omicron BA.1 variant infection (S-gene target failure defined) and each stratified by prior SARS-CoV-2 infection status. We estimated the effectiveness of primary vaccination during the period before and during booster eligibility (14-149 and [≥]150 days, respectively, after 2nd dose) and of booster vaccination ([≥]14 days after booster dose). To test whether booster vaccination reduced the risk of infection beyond that of the primary series, we compared the odds among boosted and booster eligible people. Results: Overall, 10,676 cases and 119,397 controls were included (median age: cases: 35 years, controls: 39 years). Among cases and controls, 6.1% and 7.8% had a prior infection. The effectiveness of primary vaccination 14-149 days after 2nd dose was 36.1% (95% CI, 7.1-56.1%) and 28.5% (95% CI, 20.0-36.2%) for people with and without prior infection, respectively. The effectiveness of booster vaccination was 45.8% (95% CI, 20.0-63.2%) and 56.9% (95% CI, 52.1-61.2%) in people with and without prior infection, respectively. The odds ratio comparing boosted and booster eligible people with prior infection was 0.83 (95% CI, 0.56-1.23), whereas the odds ratio comparing boosted and booster eligible people without prior infection was 0.51 (95% CI, 0.46-0.56). Conclusions and Relevance: Primary vaccination provided significant but limited protection against Omicron BA.1 infection among people with and without prior infection. While booster vaccination was associated with additional protection in people without prior infection, it was not associated with additional protection among people with prior infection. These findings support primary vaccination in people regardless of prior infection status but suggest that infection history should be considered when evaluating the need for booster vaccination.

Results: Overall, 10,676 cases and 119,397 controls were included (median age: cases: 35 years, controls: 39 years). Among cases and controls, 6.1% and 7.8% had a prior infection. The effectiveness of primary vaccination 14-149 days after 2 nd dose was 36.1% (95% CI, 7.1-56.1%) and 28.5% (95% CI, 20.0-36.2%) for people with and without prior infection, respectively. The effectiveness of booster vaccination was 45.8% (95% CI, 20.0-63.2%) and 56.9% (95% CI, 52.1-61.2%) in people with and without prior infection, respectively. The odds ratio comparing boosted and booster eligible people with prior infection was 0.83 (95% CI, 0.56-1.23), whereas the odds ratio comparing boosted and booster eligible people without prior infection was 0.51 (95% CI, 0.46-0.56).

Conclusions and Relevance: Primary vaccination provided significant but limited protection against Omicron BA.1 infection among people with and without prior infection. While booster vaccination was associated with additional protection in people without prior infection, it was not associated with additional protection among people with prior infection. These findings support primary vaccination in people regardless of prior infection status but suggest that infection history should be considered when evaluating the need for booster vaccination.

Although COVID-19 vaccines provide lower levels of protection against the B.1.1.529 (Omicron) than the B.1.617.2 (Delta) variant of SARS-CoV-2, current evidence indicates that primary and booster (third) vaccination significantly reduces the risk of Omicron related outcomes in the general population. [1] [2] [3] [4] [5] However, the benefit of vaccination in people with a prior SARS-CoV-2 infection remains uncertain.

Previous studies, conducted prior to the Omicron epidemic wave, found that primary vaccination (two doses) afforded protection against reinfection beyond that provided by a prior infection [6] [7] [8] [9] and that a booster dose significantly increase such protection. 10 In contrast, Shrestha et al. found that primary vaccination did not provide additional protection (Hazard Ratio, 0.77; 95% CI, 0.53-1.12) against SARS-CoV-2 re-infection among previously infected people during the first month of the Omicron wave. 11 Furthermore, evidence is lacking for the additional benefit of booster vaccination against Omicron infection in individuals with a prior infection, which is needed to inform vaccination policies for this subpopulation.

In this study, we analyzed data from a large cohort of people from the Yale New Haven Health system who underwent molecular testing for S-gene target failure (SGTF) to evaluate the benefit of primary series and booster doses in the context of the Omicron wave. Specifically, we estimated the effectiveness of primary and booster vaccination against Omicron (lineage BA.1) infection among people with and without a documented prior SARS-CoV-2 infection. We also examined whether booster vaccination reduced the risk of Omicron infection beyond that afforded by primary vaccination among people with and without a prior documented infection.

We conducted a test-negative case-control (TNCC) analysis using data collected as part of the Studying

Haven Health System (YNHH). The YNHH is a large academic health system comprising five hospital delivery networks and associated outpatient clinics in Connecticut and Rhode Island. We chose the TNCC design because it has been shown to provide effectiveness estimates consistent with those from randomized control trials, has been widely applied to estimate real-world effectiveness for COVID-19 vaccines, and mitigates the risk of confounding introduced by care-seeking and testing access. 1, [12] [13] [14] [15] The study population comprised vaccine eligible (≥5 years of age) people who had at least one SARS-CoV-2 test or mRNA (mRNA-1273 [Moderna] or BNT162b2 [Pfizer]) vaccine dose in the medical records. We identified SARS-CoV-2 RT-PCR tests that were collected from the study population and performed with the TaqPath™ COVID 19 (Thermo Fisher Scientific) diagnostic assay between November 1, 2021 and January 31, 2022, the period prior to and during the Omicron epidemic wave in Connecticut ( Figure 1A ). At the beginning of the study, Delta was the predominant variant in Connecticut, accounting for 99.63% (3,808 of 3,822) of the sequenced samples deposited in the GISAID database that were collected between November 1 and November 28, 2021. 16 We used the TaqPath assay to select tests as cases and controls since its S-gene probe, which fails for Omicron (BA.1) but not for Delta, allows for prediction of an Omicron infection when the primary circulating variants are Omicron and Delta. 17 We excluded tests that were performed after receiving a heterologous primary vaccination (i.e. different first and second dose manufacturers) or an Ad26.COV2 vaccine dose. Additionally, we excluded tests that were performed among people who received booster doses prior to eligibility (defined as five months since 2 nd primary vaccine dose and after booster vaccination approval in the US [ September 22, 2021] ). 16 We excluded tests that were performed in the 90 days after a positive SARS-CoV-2 test (rapid antigen or RT-PCR), had a positive reflex result with an inconclusive S-gene target failure (SGTF) finding, or were obtained from people with more than one prior SARS-CoV-2 infection or with missing confounder data ( Figure 1 ).

The Yale Computational Health Platform was used to extract demographic, comorbidity, COVID-19 vaccination and SARS-CoV-2 testing data from electronic medical records. 18, 19 Additional COVID-19 vaccination records from the state vaccination registry were linked to the YNHH medical All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 20, 2022. ;  records and extracted through the same platform. This study was approved by the Yale Institutional Review Board (ID# 2000030222).

Our exposure of interest was time from completion of primary (two doses) and booster (third dose) vaccination with mRNA-1273 or BNT162b2. We stratified vaccination by time since dose (<14 days since 2 nd dose, 14-149 days since 2 nd dose, 17, 20 An eligible control was a negative SARS-CoV-2 RT-PCR test collected ≥ 7 days prior to a positive test or the end of the study period (to account for non-resulted cases). Our sample included all eligible cases and up to three negative tests (controls) per person during the study period. If an individual had more than one negative test within a seven-day period, one random test was selected during the period as a control.

We conducted two primary analyses, each stratified by prior SARS-CoV-2 infection status. First, we estimated vaccine effectiveness as one minus the odds ratio for infection comparing vaccinated to unvaccinated people. Second, we examined whether a booster dose was associated with increased protection beyond that afforded by the primary series by comparing the odds of infection among recently boosted people (14-59 days after booster dose) to the odds among booster eligible people. In alignment with CDC booster dose recommendations at the time of analysis, 21 we defined booster eligible people as those who completed their primary series ≥ 150 days (five months) prior to the included test and had not received a booster dose. For this analysis, we were interested in the level of protection associated with a All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. A mixed effects generalized additive logistic regression was used to evaluate associations. We included the following a priori selected covariates: date of test (continuous), age (continuous), sex, race/ethnicity, Charlson comorbidity score (categorized as 0, 1-2, 3-4, 5+) 23 , number of non-emergent YNHH encounters in the year prior to vaccine rollout in Connecticut (December 2020; categorized as 0, 1-2, 3-4, 5+), insurance group (uninsured, Medicaid, Medicare, other), social vulnerability index (SVI) of residential zip code (continuous) 22 and municipality. Continuous factors were modeled using a natural spline with 3 knots and we included a random intercept for municipality. 24, 25 To account for waning infection-mediated immunity, we included time since prior SARS-CoV-2 infection as a continuous factor in analyses limited to people with prior infection. All analyses were conducted in R, version 4.1.

We performed multiple sensitivity analyses to ensure our findings were robust to alternative study design, data cleaning and modeling assumptions. Specifically, we tested the robustness of our findings to the following scenarios: 1:1 matching with replacement, exclusion of heterologous booster doses, inclusion of tests among people with more than one prior infection, exclusion of discordant test results, inclusion of positive TaqPath results with inconclusive SGTF (included as negative tests), and inclusion of all All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 20, 2022. 

Between November 1, 2021, and January 31, 2022, we identified 155,827 SARS-CoV-2 tests that were performed with the TaqPath assay on samples obtained from 113,033 unique people in the YNHH system ( Figure 1B ). The first SGTF defined Omicron infection in the study population was identified on Cases and controls had similar characteristics with respect to age, gender, SVI of residential zip code and Charlson comorbidity score (Table 1) 

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (Table 2 ).

In the secondary analysis that was restricted to people without a prior infection, the odds of Omicron infection increased over time since booster vaccination and was significantly higher 90+ days after a booster dose relative to the period 14-59 days after the dose (OR: 1.6; 95% CI, 1.2-2.0, eTable1).

Yet, the odds of infection among boosted people 90+ days after the booster dose was lower than the odds among booster eligible people (OR: 0.72; 95% CI, 0.66-0.91, Table 3 ).

In sensitivity analyses, the effectiveness of booster vaccination against Omicron infection (≥14 days after the booster dose) ranged between 35.1-48.5% for people with a prior infection and 55.0-58.5% for people without a prior infection (eFigures1-8). Compared with the primary analysis, we observed lower precision from our matched (1:1 with replacement) analysis (eFigure1). Following the exclusion of tests collected among people whose prior infection occurred after their first vaccine dose, we observed non-significantly lower effectiveness estimates for primary (14-149 days after 2 nd dose: 33.5%; 95% CI, 2.7-54.6%) and booster (14+days after booster dose: 41.9%; 95% CI, 12.1-61.6%]) vaccination. Adjusting for time between testing and completion of primary vaccination did not significantly alter the estimated All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

Leveraging data from a large population of Connecticut residents, we examined the effectiveness of primary series and booster doses against Omicron infections among primary series (2-dose) mRNA recipients with and without a prior infection. We found primary vaccination was associated with significant but low levels of protection among people with and without a prior infection. While booster vaccination was associated with protection beyond that afforded by the primary series in people without a prior infection, we did not identify a significant increase in protection among people with a prior infection.

Contrary to the findings of Shrestha et al., 11 our analysis, which ascertained Omicron infection in cases by the presence of SGTF and had increased precision for vaccine effectiveness, found that primary vaccination was associated with a significant reduction in the risk (34.0%; 95% CI, 18.5-46.5%) of Omicron infection among people with a prior infection. Though we found the level of protection afforded by primary vaccination to be low, our findings suggest that primary vaccination may be warranted regardless of prior infection status.

Our estimate of booster vaccination effectiveness among previously infected people had reduced precision (45.8%; 95% CI, 20.0-63.2%) compared to the estimate for people without a prior infection (56.9%; 95% CI, 52.1-61.2%]). Because the differences in these estimates may be driven by differences in care-seeking behaviors, we caution their direct comparison. However, in a parallel analysis, the odds of infection did not differ significantly (OR: 0.83; 95% CI, 0.56-1.22) between boosted and booster eligible people with a prior SARS-CoV-2 infection. Together, these findings suggest that boosters may not confer additional protection beyond that afforded by primary vaccination among previously infected people and that, in vaccine limited scenarios, individuals who received a primary series and have been previously infected should not be prioritized for a booster dose. These findings provide supportive evidence for the inclusion of documented prior SARS-CoV-2 infections in addition to vaccinations for future vaccination All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 20, 2022. ; https://doi.org/10.1101/2022.04.19.22274056 doi: medRxiv preprint requirements and mandatory proof of immunity (such as vaccination cards). 11,27-32 However, given the reduced precision of these estimates, additional research from other regions should be conducted to provide additional clarity on the benefits of boosters within this sub-population.

In alignment with prior studies, 1,7,33 we found the risk of Omicron infection among boosted people without a prior infection increased significantly three months after booster dose administration.

However, the odds of infection among boosted people remained significantly lower than the odds among booster eligible people (OR: 0.7; 95% CI, 0.6-0.9). Thus, even with the decline in protection, booster vaccination appears to provide additional protection beyond that conferred by primary vaccination in people without a prior SARS-CoV-2 infection.

Weekly testing for certain unvaccinated professionals, such as employees of healthcare facilities that accept Medicare and/or Medicaid or Connecticut state employees, was required by the state and federal government during the study period. 34, 35 Because such requirements resulted in increased testing among unvaccinated but not vaccinated persons, our vaccine effectiveness estimates are likely conservative. However, the bias introduced by required testing does not extend to the comparisons among vaccinated groups and our findings comparing boosted to booster eligible people are likely to be unaffected by testing policies.

Our analysis was limited to a population of Connecticut residents and was reliant on medical record data that is subject to misclassification. In the place of whole genome sequence data, we used SGTF status to ascertain Omicron infections as cases. The use of SGTF as a proxy has been widely used during the Omicron epidemic wave and has been recommended as an indicator of Omicron lineage BA.1 infection by the WHO. 1, 19 While a new sub-lineage of the Omicron variant (BA.2) has been detected in the US without SGTF, sequencing data from YNHH showed 100% agreement between SGTF and whole genome sequence-defined Omicron through December 2021. 17 In January 2022, we observed a small number (184) of positive tests that did not have SGTF and were not included as cases in the analyses. Our sample is overly representative of mild cases of SARS-CoV-2 infection since TaqPath testing was primarily All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 20, 2022. employed in the YNHH outpatient setting. Additionally, our sample excluded cases with high Ct values since Ct values of at least 30 were required for SGTF calls.

We did not have adequate sample to evaluate the level of protection conferred by two or more prior infections (n=49). We expect a proportion of prior SARS-CoV-2 infections may have gone undetected and that ascertainment of prior infection history may be subject to misclassification.

Nevertheless, our analyses of people with documented prior infection are unaffected by such misclassification. Despite accounting for health seeking behavior in our study design and confounder selection, residual behavioral differences may exist. Finally, the analyses were not powered to test associations for severe COVID-19, we therefore cannot exclude that booster vaccination may increase protection against such outcomes in people with a prior SARS-CoV-2 infection.

Primary vaccination with two COVID-19 mRNA vaccine doses provided significant but limited protection among people with and without a prior SARS-CoV-2 infection. While booster vaccination resulted in additional protection beyond that afforded by the primary vaccination among people without a prior infection, it did not result in additional protection among previously infected people. These findings support primary vaccination regardless of prior infection history but suggest that a person's history of prior SARS-CoV-2 infection should be considered in subsequent vaccination decisions, such as booster vaccination.

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 20, 2022. ; https://doi.org/10.1101/2022.04.19.22274056 doi: medRxiv preprint Data Use Agreement: The data used in this study belongs to Yale University. Qualified researchers may submit a data share request for de-identified patient level data by contacting the corresponding author with a detailed description of the research question. Information for work to advance intelligent disease prevention and health promotion; collaborates with the National Center for Cardiovascular Diseases in Beijing; is a technical consultant to Hugo Health, a personal health information platform, and co-founder of Refactor Health, an AI-augmented data All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 20, 2022. ; .006 a Limited to people without prior documented SARS-CoV-2 infections b Adjusted for date of test, age (in years), sex, race/ethnicity, insurance, comorbidity (Charlson Score), SVI of zip code, presence of prior infection, municipality, and number of non-emergent visits during the year prior to vaccine rollout in Connecticut (December 2nd 2019 and December 1st 2020) c Booster eligible defined as primary series recipients who completed their primary series (2-doses) 150+ days prior to the test and were yet to receive a booster (third) dose; 150 days was selected as it reflects the CDC booster recommendations at the time of manuscript submission All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The first Omicron infection in the study population was identified on November 11, 2021. c Excluded tests that were performed after a person was given a booster dose before FDA authorization or that was given less than 150 days after primary vaccination completion. d There were 182 people with missing SVI data and 52 people with missing sex data. e People were allowed to contribute up to three negative tests to the control sample. In the event that they had more than three negative tests over the study period, three tests were randomly selected. If a person had more than one negative test within a 7 day period, a random test was selected.

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

Association Between 3 Doses of mRNA COVID-19 Vaccine and Symptomatic Infection Caused by the SARS-CoV-2 Omicron and Delta Variants

Effectiveness of COVID-19 vaccines against the Omicron (B.1.1.529) variant of concern

COVID-19 Incidence and Death Rates Among Unvaccinated and Fully Vaccinated Adults with and Without Booster Doses During Periods of Delta and Omicron Variant Emergence -25

Effectiveness of a Third Dose of mRNA Vaccines Against COVID-19-Associated Emergency Department and Urgent Care Encounters and Hospitalizations Among Adults During Periods of Delta and Omicron Variant Predominance -VISION Network, 10 States

SARS-CoV-2 Delta VOC in Scotland: demographics, risk of hospital admission, and vaccine effectiveness. The Lancet

Effectiveness and durability of protection against future SARS-CoV-2 infection conferred by COVID-19 vaccination and previous infection; findings from the UK SIREN prospective cohort study of healthcare workers

Protection against SARS-CoV-2 after Covid-19 Vaccination and Previous Infection

Risk of SARS-CoV-2 reinfection and COVID-19 hospitalisation in individuals with natural and hybrid immunity: a retrospective, total population cohort study in Sweden

COVID-19 Cases and Hospitalizations by COVID-19 Vaccination Status and Previous COVID-19 Diagnosis -California

Effectiveness of CoronaVac, ChAdOx1, BNT162b2 and Ad26.COV2.S among individuals with prior SARS-CoV-2 infection in Brazil

Necessity of COVID-19 Vaccination in Persons Who Have Already Had COVID-19

The test-negative design: validity, accuracy and precision of vaccine efficacy estimates compared to the gold standard of randomised placebocontrolled clinical trials

Lecture 9: Study designs for evaluating vaccine efficacy

Effectiveness of ChAdOx1 vaccine in older adults during SARS-CoV-2 Gamma variant circulation in São Paulo

Effectiveness of the CoronaVac vaccine in older adults during a gamma variant associated epidemic of covid-19 in Brazil: test negative case-control study

disease and diplomacy: GISAID's innovative contribution to global health

Rapid emergence of SARS-CoV-2 Omicron variant is associated with an infection advantage over Delta in vaccinated persons

Health Care and Precision Medicine Research: Analysis of a Scalable Data Science Platform

Agile Health Care Analytics: Enabling Real-Time Disease Surveillance With a Computational Health Platform

European Centre for Disease Prevention and Control, Europe WHORO for. Methods for the Detection and Identification of SARS-CoV-2 Variants

CDC. COVID-19 Booster Shot

Depletion-of-susceptibles bias in influenza vaccine waning studies: how to ensure robust results

A new method of classifying prognostic comorbidity in longitudinal studies: development and validation

A comparison of methods to adjust for continuous covariates in the analysis of randomised trials

Adjustment for continuous confounders: an example of how to prevent residual confounding

R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing

Protective immunity after recovery from SARS-CoV-2 infection

Health-care workers recovered from natural SARS-CoV-2 infection should be exempt from mandatory vaccination edicts

The health pass becomes the vaccine pass

The Government of the Hong Kong Special Adminstrative Region. Government adjusts vaccination requirements of Vaccine Pass

Coronavirus (COVID-19)

Waning 2-Dose and 3-Dose Effectiveness of mRNA Vaccines Against COVID-19-Associated Emergency Department and Urgent Care Encounters and Hospitalizations Among Adults During Periods of Delta and Omicron Variant Predominance -VISION Network, 10 States

Workplaces Subject to COVID-19 Vaccine Requirements. CT.gov -Connecticut's Official State Website

Prior infection defined as a positive RT-PCR or rapid antigen test at least 90 days prior to included test c Booster eligible defined as primary series recipients who completed their primary series 150+ days prior to the test and were yet to receive a booster (third) dose; 150 days was selected as it reflects the CDC booster recommendations at the time of manuscript submission d Six tests were collected among people with a prior infection that occurred following their booster dose and were removed from this analysis