key: cord-0975281-288y1wvi
authors: Tenforde, Mark W; Patel, Manish M; Ginde, Adit A; Douin, David J; Talbot, H Keipp; Casey, Jonathan D; Mohr, Nicholas M; Zepeski, Anne; Gaglani, Manjusha; McNeal, Tresa; Ghamande, Shekhar; Shapiro, Nathan I; Gibbs, Kevin W; Files, D Clark; Hager, David N; Shehu, Arber; Prekker, Matthew E; Erickson, Heidi L; Exline, Matthew C; Gong, Michelle N; Mohamed, Amira; Henning, Daniel J; Peltan, Ithan D; Brown, Samuel M; Martin, Emily T; Monto, Arnold S; Khan, Akram; Hough, C Terri; Busse, Laurence; ten Lohuis, Caitlin C; Duggal, Abhijit; Wilson, Jennifer G; Gordon, Alexandra June; Qadir, Nida; Chang, Steven Y; Mallow, Christopher; Gershengorn, Hayley B; Babcock, Hilary M; Kwon, Jennie H; Halasa, Natasha; Chappell, James D; Lauring, Adam S; Grijalva, Carlos G; Rice, Todd W; Jones, Ian D; Stubblefield, William B; Baughman, Adrienne; Womack, Kelsey N; Lindsell, Christopher J; Hart, Kimberly W; Zhu, Yuwei; Olson, Samantha M; Stephenson, Meagan; Schrag, Stephanie J; Kobayashi, Miwako; Verani, Jennifer R; Self, Wesley H
title: Effectiveness of SARS-CoV-2 mRNA Vaccines for Preventing Covid-19 Hospitalizations in the United States
date: 2021-08-06
journal: Clin Infect Dis
DOI: 10.1093/cid/ciab687
sha: 5c36725753119a54e2ba9c87b34ae7cc96b0c59e
doc_id: 975281
cord_uid: 288y1wvi

BACKGROUND: As SARS-CoV-2 vaccination coverage increases in the United States (US), there is a need to understand the real-world effectiveness against severe Covid-19 and among people at increased risk for poor outcomes. METHODS: In a multicenter case-control analysis of US adults hospitalized March 11-May 5, 2021, we evaluated vaccine effectiveness to prevent Covid-19 hospitalizations by comparing odds of prior vaccination with an mRNA vaccine (Pfizer-BioNTech or Moderna) between cases hospitalized with Covid-19 and hospital-based controls who tested negative for SARS-CoV-2. RESULTS: Among 1212 participants, including 593 cases and 619 controls, median age was 58 years, 22.8% were Black, 13.9% were Hispanic, and 21.0% had immunosuppression. SARS-CoV-2 lineage B.1.1.7 (Alpha) was the most common variant (67.9% of viruses with lineage determined). Full vaccination (receipt of two vaccine doses ≥14 days before illness onset) had been received by 8.2% of cases and 36.4% of controls. Overall vaccine effectiveness was 87.1% (95% CI: 80.7 to 91.3%). Vaccine effectiveness was similar for Pfizer-BioNTech and Moderna vaccines, and highest in adults aged 18-49 years (97.4%; 95% CI: 79.3 to 99.7%). Among 45 patients with vaccine-breakthrough Covid hospitalizations, 44 (97.8%) were ≥50 years old and 20 (44.4%) had immunosuppression. Vaccine effectiveness was lower among patients with immunosuppression (62.9%; 95% CI: 20.8 to 82.6%) than without immunosuppression (91.3%; 95% CI: 85.6 to 94.8%). CONCLUSION: During March–May 2021, SARS-CoV-2 mRNA vaccines were highly effective for preventing Covid-19 hospitalizations among US adults. SARS-CoV-2 vaccination was beneficial for patients with immunosuppression, but effectiveness was lower in the immunosuppressed population.

preventing Covid-19 hospitalizations by vaccine product, by age group, and by underlying medical conditions. [5] A c c e p t e d M a n u s c r i p t 7

We conducted a prospective observational case-control evaluation of vaccine effectiveness by comparing the odds of antecedent SARS-CoV-2 vaccination in hospitalized case-patients with Covid-19 versus control-patients without Covid-19. We included two control groups: 1) "test-negative" controls were hospitalized with signs or symptoms of an acute respiratory illness but tested negative for SARS-CoV-2; and 2) "syndrome-negative" controls were hospitalized without signs or symptoms of an acute respiratory illness and tested negative for SARS-CoV-2. Test-negative controls are commonly used in hospital-based vaccine effectiveness evaluations; [6] [7] [8] [9] in the test-negative design, utilizing a comparison group with the same clinical syndrome and similar level of acuity as cases reduces bias due to differential healthcare seeking behavior. Because of the potential for misclassification of true cases as test-negative controls due to false-negative tests, particularly for those presenting late in the course of illness, we included the second control group of hospitalized patients without an acute respiratory illness. [10] [11]

This surveillance activity included patients hospitalized from March 11 through May 5, 2021 at 18 US hospitals within the IVY Network. [6, 12] This activity was conducted consistent with applicable federal law and CDC policy (Supplementary Appendix B).

Sites screened hospitalized adults ≥18 years old for potential eligibility through daily review of hospital admission logs and electronic medical records. Detailed eligibility criteria are shown in Supplementary Appendix B. Covid-19 cases included patients hospitalized with a clinical syndrome consistent with acute Covid-19 (≥1 of the following: fever; cough; shortness of breath; loss of taste; loss of smell; use of respiratory support for the acute illness; or new pulmonary findings on chest A c c e p t e d M a n u s c r i p t 8 imaging consistent with pneumonia) and a positive test for SARS-CoV-2 within 10 days following symptom onset. [13] [14] [15] Test-negative controls were hospitalized with a clinical syndrome consistent   with acute Covid-19 and tested negative for SARS-CoV-2. Syndrome-negative controls were hospitalized without a clinical syndrome consistent with Covid-19 and tested negative for SARS-CoV-2. Individual matching between cases and controls was not performed. Sites attempted to capture all cases admitted to the hospital during the surveillance period and targeted a case: control ratio of approximately 1:1. Information on vaccination status was not collected until after patients were included.

Participants (or their proxies) were interviewed by trained personnel to collect data on demographics, medical conditions, SARS-CoV-2 vaccination, and other patient characteristics.

Additional information on underlying medical conditions and SARS-CoV-2 clinical testing was obtained through standardized medical record review.

Upper respiratory specimens (nasal swabs or saliva) were collected, frozen, and shipped to a central laboratory at Vanderbilt University Medical Center (Nashville, Tennessee). Specimens underwent reverse transcription polymerase chain reaction (RT-PCR) testing for the SARS-CoV-2 nucleocapsid gene using standardized methods and interpretive criteria. [16] Specimens positive for SARS-CoV-2 with a cycle threshold <32 were shipped to the University of Michigan (Ann Arbor, Michigan) for viral whole genome sequencing using the ARTIC Network version 3 protocol on an Oxford Nanopore Technologies instrument (Supplementary Appendix B) . [17] SARS-CoV-2 lineages were assigned with >80% coverage using Pangolin genomes. [18] A c c e p t e d M a n u s c r i p t 9

Final classification of case-control status was determined with consideration of both clinical SARS-CoV-2 testing completed at local hospital laboratories and RT-PCR testing completed at the central laboratory. Cases tested positive for SARS-CoV-2 by a clinical test or central laboratory RT-PCR test.

Cases with SARS-CoV-2 detected by RT-PCR with a cycle threshold >32 were included in the analysis, but viral sequencing information was not available for these cases. Test-negative and syndrome negative controls tested negative for SARS-CoV-2 by all clinical and central laboratory testing.

Details of SARS-CoV-2 vaccination, including dates and location of vaccination, vaccine product, and lot number, were ascertained through a systematic process including patient or proxy interview and The SARS-CoV-2 mRNA vaccines are administered as a two-dose series; participants were considered fully vaccinated 14 days after receipt of the second vaccine dose. [19] Vaccination status was classified based on the number of mRNA vaccine doses received before a reference date, which was the date of symptom onset for cases and test-negative controls and date of hospital admission for syndrome-negative controls. Participants were classified as: unvaccinated if they had received no vaccine doses prior to the reference date; partially vaccinated if they received one dose ≥14 days before the reference date; and fully vaccinated if they received both doses ≥14 days before the reference date. As protective immunity from SARS-CoV-2 vaccines is not expected immediately after the first dose, [12] patients who received a first dose <14 days before the reference date were A c c e p t e d M a n u s c r i p t 10 excluded from the analysis. Patients who received a SARS-CoV-2 vaccine that had not been authorized in the US were excluded. Due to recent introduction of the Janssen (Johnson & Johnson) SARS-CoV-2 vaccine following its EUA in February 2021, [2] patients who received this vaccine were also excluded.

Vaccine effectiveness and 95% confidence intervals (95% CI) were determined by comparing the odds of prior SARS-CoV-2 vaccination in case-patients and control-patients, calculated as: vaccine effectiveness = (1 -odds ratio) × 100%. [20] Primary vaccine effectiveness estimates were calculated in adults of all ages for full vaccination versus unvaccinated and for partial vaccination versus unvaccinated. Unadjusted odds ratios were calculated with simple logistic regression and then a model building approach was applied to estimate adjusted vaccine effectiveness accounting for potential confounders.

Prespecified covariates in base multivariable logistic regression models included calendar time in biweekly intervals, US Department of Health and Human Services region, age, sex, and self-reported race and Hispanic ethnicity. We repeated the regression by adding health status indicators (such as number of chronic conditions and prior hospitalizations in the past year) and SARS-CoV-2 exposure variables (such as mask use and attending large gatherings) potentially associated with the likelihood of vaccination and risk of Covid-19 hospitalization (detailed in Supplementary Appendix B). An absolute change in the odds ratio of vaccination of more than 5% in either direction was used as a pre-specified cutoff for inclusion of additional variables to the base model. Potential effect modification of prior SARS-CoV-2 infection (at least 14 days prior to the current illness) was assessed using a likelihood ratio test (with a P-value <0.15 suggestive of effect modification). [21] Separate assessments were initially performed using the test-negative control and the syndrome-negative control groups to assess comparability of estimates. Effectiveness estimates A c c e p t e d M a n u s c r i p t 11 were very similar using the test-negative and syndrome-negative control groups. Therefore, control groups were combined to improve precision.

Vaccine effectiveness estimates were stratified by age group (18- 

We included 1212 patients (593 cases, 334 test-negative controls, and 285 syndrome-negative controls) ( Figure S1 ) enrolled from 18 clinical sites ( Figure 1 ) over the course of 51 days ( Figure S2 ).

Overall, median age was 58 years, 276 (22.8%) were non-Hispanic Black, 168 (13.9%) were Hispanic, and 254 (21.0%) had an immunocompromising condition (Table 1, Table S1 ). Among the 593 case patients, 32.2% were admitted to an intensive care unit (ICU) and 8.3% died during the hospitalization. Table 2) .

The base vaccine effectiveness model was used as additional variables that were considered did not change the odds ratio of vaccination by the pre-specified cutoff of more than 5%. Vaccine effectiveness results for full vaccination were very similar using the test-negative control group (86.6%, 95% CI: 79.0-91.4%) and syndrome-negative control group (87.0%, 95% CI: 79.0-91.9%) ( Figure S3 and Figure S4) (Table   S2 ).

Forty-five Covid-19 case patients were fully vaccinated before symptom onset (Table S3) This analysis adds to early real-world evaluations that demonstrated high vaccine effectiveness against Covid-19 in groups prioritized for early vaccination, such as healthcare workers. [22, 23] This analysis had certain limitations. While we included control groups that were likely to reduce bias from differential healthcare seeking behavior, there was potential for residual confounding. People who chose to be vaccinated may have been more likely to engage in other behaviors to reduce their risk for Covid-19, such as mask use and avoiding large crowds. However, adjusting for self-reported variables on non-vaccine preventive measures did not substantively change vaccine effectiveness estimates suggesting this was not a major confounder. Race and Hispanic ethnicity differed between case and control groups; this likely represented underlying differences in the incidence of SARS-CoV-2 infection by race and ethnicity in the US and models were adjusted for race and ethnicity. [28] Different immunocompromising conditions are likely associated with varying severity of immunosuppression; more severe immunosuppression may be associated with lower vaccine effectiveness, but this analysis was not powered to look at vaccine effectiveness among subgroups of immunocompromising conditions. In an effort to capture all COVID-19 cases admitted to participating hospitals during a period of high community transmission, enrollment of cases and controls was not matched on a day-to-day basis; however, all cases and controls were A c c e p t e d M a n u s c r i p t 16 enrolled within a 51-day period and vaccine effectiveness models were adjusted for calendar time.

As hospitalized adults frequently had multiple chronic medical conditions that may impact effectiveness of vaccines, findings from this analysis may not be broadly generalizable to populations with lower burden of chronic medical conditions. Lastly, most sequenced viruses in this analysis were B.1.1.7 (Alpha) variants, which represented the majority of circulating viruses in the US during this time period;[1] vaccine effectiveness against other emerging variants will require additional study.

In conclusion, the SARS-CoV-2 mRNA vaccines were highly effective for preventing Covid- reports grants from CDC, during the conduct of the study; grants from AbbVie, grants from Faron Pharmaceuticals, outside the submitted work. Dr. Gong reports grants from CDC, during the conduct of the study; grants from NIH, grants from AHRQ, fees for participating on DSMB not related to this work from Regeneron, personal fees from Philips Healthcare for participation on scientific advisory panel not related to this work, outside the submitted work. Dr. Grijalva M a n u s c r i p t 27 A c c e p t e d M a n u s c r i p t 32 Figure 3 

for Disease Control and Prevention. COVID data tracker

Food & Drug Administration. COVID-19 vaccines

Efficacy and Safety of the mRNA-1273 SARS-CoV-2

Safety and Efficacy of the BNT162b2 mRNA Covid-19

Hospitalization Rates and Characteristics of Patients Hospitalized with Laboratory-Confirmed Coronavirus Disease 2019 -COVID-NET, 14 States

Influenza Vaccine Effectiveness for Prevention of Severe Influenza-Associated Illness among Adults in the United States, 2019-2020: A testnegative study

Influenza vaccine effectiveness against hospitalization in the United States

Results From the US Hospitalized Adult Influenza Vaccine Effectiveness Network (HAIVEN)

The case test-negative design for studies of the effectiveness of influenza vaccine

Theoretical framework for retrospective studies of the effectiveness of SARS-CoV-2 vaccines

Evaluation of COVID-19 vaccine effectiveness: Interim guidance

Effectiveness of Pfizer-BioNTech and Moderna Vaccines Against COVID-19 Among Hospitalized Adults Aged ≥65 Years -United States

The Incubation Period of Coronavirus Disease

From Publicly Reported Confirmed Cases: Estimation and Application

Modeling the impacts of clinical influenza testing on influenza vaccine effectiveness estimates

Incorporating Real-time Influenza Detection Into the Test-negative Design for Estimating Influenza Vaccine Effectiveness: The Real-time Testnegative Design (rtTND)

Food & Drug Administration. CDC 2019-Novel Coronavirus (2019-nCoV) Real-Time RT-PCR Diagnostic Panel

nCoV2-2019 sequencing protocol v3 (LoCost) V.3. Available at

A dynamic nomenclature proposal for SARS-CoV-2 lineages to assist genomic epidemiology

Interim clinical considerations for use of COVID-19 vaccines currently authorized in the United States

Influenza Vaccine Effectiveness in the United States during the 2015-2016 Season

Does raising type 1 error rate improve power to detect interactions in linear regression models? A simulation study

Interim Estimates of Vaccine Effectiveness of BNT162b2 and mRNA-1273 COVID-19 Vaccines in Preventing SARS-CoV-2 Infection Among Health Care Personnel, First Responders, and Other Essential and Frontline Workers -Eight

Interim Estimates of Vaccine Effectiveness of Pfizer-BioNTech and Moderna COVID-19 Vaccines Among Health Care Personnel -33

Antibody Response to 2-Dose SARS-CoV-2 mRNA Vaccine Series in Solid Organ Transplant Recipients

High antibody response to two-dose SARS-CoV-2 messenger RNA vaccination in patients with rheumatic and musculoskeletal diseases

Analysis of MarketScan Data for Immunosuppressive Conditions and Hospitalizations for Acute Respiratory Illness, United States

BNT162b2 mRNA Covid-19 Vaccine in a Nationwide Mass Vaccination Setting

Racial and Ethnic Disparities in Incidence of SARS

US States and DC

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Trends in Patient Characteristics and COVID-19 In-Hospital Mortality in the United States During the COVID-19 Pandemic

COVID-19 reported patient impact and hospital capacity by state

Capa/6xf2-c3ie

Annual estimates of the resident population

A c c e p t e d M a n u s c r i p t 20 A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t 26 Among fully vaccinated patients, days between second vaccine dose and symptom onset (or hospital admission for syndrome-negative control group) -median (IQR) 44 (25 -54) 43 (26 -