Effectiveness of a Third Dose of Pfizer-BioNTech and Moderna Vaccines in Preventing COVID-19 Hospitalization Among Immunocompetent and Immunocompromised Adults — United States, August–December 2021

COVID-19 mRNA vaccines (BNT162b2 [Pfizer-BioNTech] and mRNA-1273 [Moderna]) provide protection against infection with SARS-CoV-2, the virus that causes COVID-19, and are highly effective against COVID-19–associated hospitalization among eligible persons who receive 2 doses (1,2). However, vaccine effectiveness (VE) among persons with immunocompromising conditions* is lower than that among immunocompetent persons (2), and VE declines after several months among all persons (3). On August 12, 2021, the Food and Drug Administration (FDA) issued an emergency use authorization (EUA) for a third mRNA vaccine dose as part of a primary series ≥28 days after dose 2 for persons aged ≥12 years with immunocompromising conditions, and, on November 19, 2021, as a booster dose for all adults aged ≥18 years at least 6 months after dose 2, changed to ≥5 months after dose 2 on January 3, 2022 (4,5,6). Among 2,952 adults (including 1,385 COVID-19 case-patients and 1,567 COVID-19–negative controls) hospitalized at 21 U.S. hospitals during August 19–December 15, 2021, effectiveness of mRNA vaccines against COVID-19–associated hospitalization was compared between adults eligible for but who had not received a third vaccine dose (1,251) and vaccine-eligible adults who received a third dose ≥7 days before illness onset (312). Among 1,875 adults without immunocompromising conditions (including 1,065 [57%] unvaccinated, 679 [36%] 2-dose recipients, and 131 [7%] 3-dose [booster] recipients), VE against COVID-19 hospitalization was higher among those who received a booster dose (97%; 95% CI = 95%–99%) compared with that among 2-dose recipients (82%; 95% CI = 77%–86%) (p <0.001). Among 1,077 adults with immunocompromising conditions (including 324 [30%] unvaccinated, 572 [53%] 2-dose recipients, and 181 [17%] 3-dose recipients), VE was higher among those who received a third dose to complete a primary series (88%; 95% CI = 81%–93%) compared with 2-dose recipients (69%; 95% CI = 57%–78%) (p <0.001). Administration of a third COVID-19 mRNA vaccine dose as part of a primary series among immunocompromised adults, or as a booster dose among immunocompetent adults, provides improved protection against COVID-19–associated hospitalization.

During August 19–December 15, 2021, a period in which the SARS-CoV-2 B.1.617.2 (Delta) variant was predominant, adults admitted to 21 hospitals in 18 states within the Influenza or Other Viruses in the Acutely Ill Network (IVY Network) who received testing for SARS-CoV-2 were included in a VE analysis. The analysis start date of August 19, 2021, was one week after the EUA of a third dose for persons with immunocompromising conditions. VE against COVID-19 hospitalization was compared between two groups: 1) adults who had completed a 2-dose mRNA vaccination series and were eligible for but had not received a third dose^†; and 2) adults who were eligible for and had received a third dose ≥7 days before illness onset. VE was calculated for 2-dose and 3-dose recipients by comparing odds of antecedent vaccination between COVID-19 case-patients and control patients who did not have COVID-19. Case-patients had COVID-19–like illness^§ and received positive SARS-CoV-2 test results by a nucleic acid amplification test (NAAT) or antigen test within 10 days of illness onset. Control patients were hospitalized with or without COVID-19–like illness and received negative SARS-CoV-2 test results by NAAT.

Patients or their proxies were interviewed regarding patient demographic and clinical characteristics, and medical record searches were completed to collect information about chronic medical conditions. Information about receipt of prior COVID-19 vaccination doses, including dates, locations, and vaccine product received, was obtained through self-report and review of source documentation (including state vaccination registries, medical records, and vaccination cards). A patient was considered to be vaccinated if vaccination could be verified through source documentation or by self-report, including dates and location. Three vaccination groups were considered: 1) unvaccinated patients, who had received no COVID-19 vaccine doses before illness onset; 2) 2-dose mRNA vaccine recipients, who were eligible for but had not received a third vaccine dose or had received a third dose <7 days before illness onset; and 3) 3-dose mRNA vaccine recipients, who received a third dose ≥7 days before illness onset. Patients were excluded if they were admitted to the hospital ≤7 days after EUA authorization of a third dose, received one or more vaccine doses but did not qualify for inclusion in the 2-dose or 3-dose vaccine group,^¶ received a non-mRNA vaccine (e.g., Ad26.COV2.S [Janssen (Johnson & Johnson)]), or if verification of vaccination was pending.

VE against COVID-19–associated hospitalization was estimated using logistic regression, comparing the odds of being vaccinated versus being unvaccinated among case-patients and controls using the equation VE = 100 × (1 − adjusted odds ratio). The regression model included case-patient or control status as the outcome variable and an indicator variable for vaccination group (unvaccinated, 2-dose recipient, or 3-dose recipient), and was adjusted for admission date, region of hospital (U.S. Department of Health and Human Services or U.S. Census Bureau), age group (18–49, 50–64, or ≥65 years), sex, and self-reported race and ethnicity. Separate models were generated for immunocompetent adults and adults with immunocompromising conditions. To compare VE among 2-dose versus 3-dose mRNA vaccine recipients, post hoc comparisons were performed with the pwcompare function in Stata with a two-sided significance threshold of p<0.05. Analyses were conducted using Stata software (version 16.0; StataCorp). This activity was determined to be public health surveillance by each participating site and CDC and was conducted consistent with applicable federal law and CDC policy.**

During August 19–December 15, 2021, the IVY Network enrolled 4,094 adults aged ≥18 years. After excluding 1,142 patients (619 because they did not belong to an included vaccination group, 386 because the patients had received a non-mRNA vaccine or vaccination verification was incomplete, and 137 because they met other exclusion criteria), 2,952 hospitalized patients were included (1,385 case-patients and 1,567 non–COVID-19 controls). Among all participants, median age was 62 years, 49% of patients were female, 58% were non-Hispanic White, and 36% had an immunocompromising condition. Among the 1,385 case-patients, 931 (67%), 408 (29%), and 46 (3%) had received 0, 2, and 3 mRNA vaccine doses, respectively. Among 1,567 non–COVID-19 controls, 458 (29%), 843 (54%), and 266 (17%) had received 0, 2, and 3 mRNA vaccine doses, respectively. Among patients without immunocompromising conditions (Table 1), 2- and 3-dose recipients were similar in terms of age (median = 69 and 72 years, respectively) but differed in self-reported race/ethnicity distribution (a higher percentage of non-Hispanic White persons were among 3-dose recipients; p = 0.008) and U.S. Census Bureau region of the admitting hospital (p = 0.030). Two-dose recipients were less likely to report working in health care settings (5%) than were 3-dose recipients (10%) (p = 0.023) and were more likely to be enrolled as a COVID-19 case-patient (31%) than were 3-dose recipients (8%) (p<0.001). Among patients with immunocompromising conditions (Table 2), 2-dose recipients were more likely to be enrolled as a case-patient (34%) than were 3-dose recipients (20%) (p<0.001). VE against COVID-19 hospitalization among adults without immunocompromising conditions was 82% (95% CI = 77%–86%) for 2 doses and 97% (95% CI = 95%–99%) for 3 doses (p<0.001) (Table 3). VE against COVID-19 hospitalization among adults with immunocompromising conditions was 69% (95% CI = 57%–78%) for 2 doses and 88% (95% CI = 81%–93%) for 3 doses (p<0.001). In a sensitivity analysis among patients with moderately to severely immunocompromising conditions defined by CDC^†† (7), VE was 65% (95% CI = 49%–76%) for receipt of 2 doses, and 87% (95% CI = 78%–92%) for receipt of 3 doses (p<0.001).

Corresponding author: Mark W. Tenforde, [email protected].

¹CDC COVID-19 Emergency Response Team; ²Baylor Scott & White Health, Temple, Texas; ³Texas A&M University College of Medicine, Temple, Texas; ⁴University of Colorado School of Medicine, Aurora, Colorado; ⁵Vanderbilt University Medical Center, Nashville, Tennessee; ⁶University of Iowa, Iowa City, Iowa; ⁷Wake Forest University Baptist Medical Center, Winston-Salem, North Carolina; ⁸The Johns Hopkins Hospital, Baltimore, Maryland; ⁹Hennepin County Medical Center, Minneapolis, Minnesota; ¹⁰Montefiore Medical Center, Albert Einstein College of Medicine, New York City borough of the Bronx, New York; ¹¹University of Washington School of Medicine, Seattle, Washington; ¹²Baystate Medical Center, Springfield, Massachusetts; ¹³Intermountain Medical Center and University of Utah, Salt Lake City, Utah; ¹⁴University of Michigan School of Public Health, Ann Arbor, Michigan; ¹⁵Oregon Health & Science University Hospital, Portland, Oregon; ¹⁶Emory University School of Medicine, Atlanta, Georgia; ¹⁷Cleveland Clinic, Cleveland, Ohio; ¹⁸Stanford University School of Medicine, Palo Alto, California; ¹⁹Ronald Reagan-UCLA Medical Center, Los Angeles, California; ²⁰University of Miami, Miami, Florida; ²¹Washington University, St. Louis, Missouri; ²²Ohio State University Wexner Medical Center, Columbus, Ohio; ²³University of Michigan School of Medicine, Ann Arbor, Michigan; ²⁴Beth Israel Deaconess Medical Center, Boston, Massachusetts.

References

1. Thompson MG, Stenehjem E, Grannis S, et al. Effectiveness of COVID-19 vaccines in ambulatory and inpatient care settings. N Engl J Med 2021;385:1355–71. Epub September 8, 2021. https://doi.org/10.1056/NEJMoa2110362 PMID:34496194
2. Tenforde MW, Self WH, Adams K, et al.; Influenza and Other Viruses in the Acutely Ill (IVY) Network. Association between mRNA vaccination and COVID-19 hospitalization and disease severity. JAMA 2021;326:2043–54. https://doi.org/10.1001/jama.2021.19499 PMID:34734975
3. Tartof SY, Slezak JM, Fischer H, et al. Effectiveness of mRNA BNT162b2 COVID-19 vaccine up to 6 months in a large integrated health system in the USA: a retrospective cohort study. Lancet 2021;398:1407–16. https://doi.org/10.1016/S0140-6736(21)02183-8 PMID:34619098
4. Food and Drug Administration. Coronavirus (COVID-19) update: FDA takes multiple actions to expand use of Pfizer-BioNtech COVID-19 vaccine. Silver Spring, MD: US Department of Health and Human Services, Food and Drug Administration; 2022. https://www.fda.gov/news-events/press-announcements/coronavirus-covid-19-update-fda-takes-multiple-actions-expand-use-pfizer-biontech-covid-19-vaccine
5. Food and Drug Administration. Coronavirus (COVID-19) update: FDA authorizes additional vaccine dose for certain immunocompromised individuals. Silver Spring, MD: US Department of Health and Human Services, Food and Drug Administration; 2021. https://www.fda.gov/news-events/press-announcements/coronavirus-covid-19-update-fda-authorizes-additional-vaccine-dose-certain-immunocompromised.
6. Food and Drug Administration. Coronavirus (COVID-19) update: FDA expands eligibility for COVID-19 vaccine boosters. Silver Spring, MD: US Department of Health and Human Services, Food and Drug Administration; 2021. https://www.fda.gov/news-events/press-announcements/coronavirus-covid-19-update-fda-expands-eligibility-covid-19-vaccine-boosters.
7. CDC. COVID-19: COVID-19 vaccines for moderately or severely immunocompromised people. Atlanta, GA: US Department of Health and Human Services, CDC; 2021. Accessed January 6, 2021.https://www.cdc.gov/coronavirus/2019-ncov/vaccines/recommendations/immuno.html
8. CDC. CDC COVID data tracker. Atlanta, GA: US Department of Health and Human Services, CDC; 2021. Accessed December 14, 2021. https://covid.cdc.gov/covid-data-tracker/#cases_casesper100klast7days
9. Nemet I, Kliker L, Lustig Y, et al. Third BNT162b2 vaccination neutralization of SARS-CoV-2 Omicron infection. N Engl J Med 2021;NEJMc2119358. https://doi.org/10.1056/NEJMc2119358 PMID:34965337
10. Thompson MG, Natarajan K, Irving SA, et al. 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, August 2021–January 2022. MMWR Morb Mortal Wkly Rep Epub January 21, 2022. https://doi.org/10.1038/s41591-021-01540-1 PMID:34588689

Abbreviations: LTCF = long-term care facility; NA = not applicable.
* Hospitals by U.S. Census region included Northeast: Baystate Medical Center (Springfield, Massachusetts), Beth Israel Deaconess Medical Center (Boston, Massachusetts), Montefiore Medical Center (New York City borough of the Bronx, New York); Midwest: University of Iowa Hospitals & Clinics (Iowa City, Iowa), University of Michigan Hospital (Ann Arbor, Michigan), Hennepin County Medical Center (Minneapolis, Minnesota), Barnes-Jewish Hospital (St. Louis, Missouri), Cleveland Clinic (Cleveland, Ohio), The Ohio State University Wexner Medical Center (Columbus, Ohio); South: Vanderbilt University Medical Center (Nashville, Tennessee), University of Miami Medical Center (Miami, Florida), Emory University Hospital (Atlanta, Georgia), The Johns Hopkins Hospital (Baltimore, Maryland), Atrium Health Wake Forest Baptist Medical Center (Winston-Salem, North Carolina), Baylor Scott & White Medical Center (Temple, Texas); West: Stanford University Medical Center (Palo Alto, California), Ronald Reagan UCLA Medical Center (Los Angeles, California), UCHealth University of Colorado Hospital (Aurora, Colorado), Oregon Health & Science University Hospital (Portland, Oregon), Intermountain Medical Center (Murray, Utah), University of Washington Medical Center (Seattle, Washington).
† Three vaccination groups were included; unvaccinated patients received no COVID-19 vaccine doses before the date of current illness onset; 2-dose recipients received 2 doses of an mRNA COVID-19 vaccine with the second dose received ≥180 days before the date of illness onset; 3-dose recipients received a third (booster) dose of vaccine with ≥180 days between the second and third dose and ≥7 days between the date of dose 3 receipt and illness onset. Among those who received a third dose with mRNA-1273 from Moderna, one half a dose (0.25 mL) was recommended for the third dose.
^§ Dose 3 received ≥6 months after dose 2.
^¶ Comparisons were made between 2-dose (primary series only) and 3-dose recipients (booster) groups using Pearson’s chi-square testing for categorical variables and the non-parametric Wilcoxon rank-sum test for continuous variables.
** Race and ethnic groups self-reported.
^†† LTCFs included residence in a nursing home, assisted living home, or rehab hospital or other subacute or chronic facility before the hospital admission.

Abbreviations: LTCF = long-term care facility; NA = not applicable.
* Immunocompromising conditions included having one or more of the following: active solid organ cancer (active cancer defined as treatment for the cancer or newly diagnosed cancer in the past 6 months); active hematologic cancer (such as leukemia, lymphoma, or myeloma); HIV infection without AIDS; AIDS; congenital immunodeficiency syndrome; prior splenectomy; prior solid organ, stem cell, or bone marrow transplant; immunosuppressive medication; systemic lupus erythematosus; rheumatoid arthritis; psoriasis; scleroderma; or inflammatory bowel disease, including Crohn’s disease or ulcerative colitis. Among 2-dose vaccine recipients with immunocompromising conditions (572), 274 (48%) had active cancer, 156 (27%) were on immunosuppressive medications, 118 (21%) had a history of solid organ transplant, and 146 (26%) had another immunocompromising condition; among 3-dose vaccine recipients with immunocompromising conditions (181), 88 (49%) had active cancer, 74 (41%) were on immunosuppressive medications, 65 (36%) had a history of solid organ transplant, and 20 (11%) had another immunocompromising condition.
^† Hospitals by U.S. Census region included Northeast: Baystate Medical Center (Springfield, Massachusetts), Beth Israel Deaconess Medical Center (Boston, Massachusetts), Montefiore Medical Center (New York City borough of the Bronx, New York); Midwest: University of Iowa Hospitals & Clinics (Iowa City, Iowa), University of Michigan Hospital (Ann Arbor, Michigan), Hennepin County Medical Center (Minneapolis, Minnesota), Barnes-Jewish Hospital (St. Louis, Missouri), Cleveland Clinic (Cleveland, Ohio), The Ohio State University Wexner Medical Center (Columbus, Ohio); South: Vanderbilt University Medical Center (Nashville, Tennessee), University of Miami Medical Center (Miami, Florida), Emory University Hospital (Atlanta, Georgia), The Johns Hopkins Hospital (Baltimore, Maryland), Atrium Health Wake Forest Baptist Medical Center (Winston-Salem, North Carolina), Baylor Scott & White Medical Center (Temple, Texas); West: Stanford University Medical Center (Palo Alto, California), Ronald Reagan UCLA Medical Center (Los Angeles, California), UCHealth University of Colorado Hospital (Aurora, Colorado), Oregon Health & Science University Hospital (Portland, Oregon), Intermountain Medical Center (Murray, Utah), University of Washington Medical Center (Seattle, Washington).
^§ Three vaccination groups were included; unvaccinated patients received no COVID-19 vaccine doses before the date of current illness onset; 2-dose recipients received 2 doses of an mRNA COVID-19 vaccine with the second dose received ≥28 days before the date of illness onset; 3-dose recipients received 3 doses of vaccine to complete a primary vaccine series with ≥28 days between the second and third dose and ≥7 days between the date of dose 3 receipt and illness onset.
^¶ Dose 3 received ≥28 days after dose 2.
** Comparisons were made between 2-dose and 3-dose recipient groups using Pearson’s chi-square testing for categorical variables and the nonparametric Wilcoxon rank-sum test for continuous variables.
^†† Race and ethnic groups self-reported.
^§§ LTCFs included residence in a nursing home, assisted living home, or rehab hospital or other subacute or chronic facility before the hospital admission.

Abbreviation: VE = vaccine effectiveness.
* VE was estimated using logistic regression, comparing the odds of being vaccinated with the Moderna or Pfizer-BioNTech COVID-19 vaccine product versus being unvaccinated among case-patients and controls using the equation VE = 100 × (1 – adjusted odds ratio). The regression model included three categories for vaccination status: unvaccinated, vaccinated with 2 doses of an mRNA vaccine, or vaccinated with 3 doses of an mRNA vaccine. VE was calculated separately comparing 2-dose recipients to unvaccinated controls and 3-dose recipients to unvaccinated controls. Models were adjusted for date of hospital admission (biweekly intervals), U.S. Department of Health and Human Services region of hospital, age group (18–49, 50–64, or ≥65 years), sex, and race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic of any race, non-Hispanic Other, or unknown).
† Hospitals by U.S. Department of Health and Human Services region included: Region 1: Baystate Medical Center (Springfield, Massachusetts), Beth Israel Deaconess Medical Center (Boston, Massachusetts); Region 2: Montefiore Medical Center (New York City borough of the Bronx, New York); Region 3: The Johns Hopkins Hospital (Baltimore, Maryland); Region 4: Vanderbilt University Medical Center (Nashville, Tennessee), University of Miami Medical Center (Miami, Florida), Emory University Hospital (Atlanta, Georgia), Atrium Health Wake Forest Baptist Medical Center (Winston-Salem, North Carolina); Region 5: University of Michigan Hospital (Ann Arbor, Michigan), Hennepin County Medical Center (Minneapolis, Minnesota), Cleveland Clinic (Cleveland, Ohio), The Ohio State University Wexner Medical Center (Columbus, Ohio); Region 6: Baylor Scott & White Medical Center (Temple, Texas); Region 7: University of Iowa Hospitals & Clinics (Iowa City, Iowa), Barnes Jewish Hospital (St. Louis, Missouri); Region 8: UCHealth University of Colorado Hospital (Aurora, Colorado), Intermountain Medical Center (Murray, Utah); Region 9: Stanford University Medical Center (Palo Alto, California), Ronald Reagan UCLA Medical Center (Los Angeles, California); Region 10: Oregon Health & Science University Hospital (Portland, Oregon), University of Washington Medical Center (Seattle, Washington).
^§ Post-hoc pairwise comparisons of VE for 2-dose versus 3-dose vaccine recipients was evaluated using the pwcompare function in Stata software (version 16; StataCorp); a two sided p-value <0.05 indicated a significant difference in VE between groups.