On March 18, 2022, this report was posted online as an MMWR Early Release.

resize icon View Larger Close

COVID-19 mRNA vaccines (BNT162b2 [Pfizer-BioNTech] and mRNA-1273 [Moderna]) are effective at preventing COVID-19–associated hospitalization (1–3). However, how well mRNA vaccines protect against the most severe outcomes of these hospitalizations, including invasive mechanical ventilation (IMV) or death is uncertain. Using a case-control design, mRNA vaccine effectiveness (VE) against COVID-19–associated IMV and in-hospital death was evaluated among adults aged ≥18 years hospitalized at 21 U.S. medical centers during March 11, 2021–January 24, 2022. During this period, the most commonly circulating variants of SARS-CoV-2, the virus that causes COVID-19, were B.1.1.7 (Alpha), B.1.617.2 (Delta), and B.1.1.529 (Omicron). Previous vaccination (2 or 3 versus 0 vaccine doses before illness onset) in prospectively enrolled COVID-19 case-patients who received IMV or died within 28 days of hospitalization was compared with that among hospitalized control patients without COVID-19. Among 1,440 COVID-19 case-patients who received IMV or died, 307 (21%) had received 2 or 3 vaccine doses before illness onset. Among 6,104 control-patients, 4,020 (66%) had received 2 or 3 vaccine doses. Among the 1,440 case-patients who received IMV or died, those who were vaccinated were older (median age = 69 years), more likely to be immunocompromised* (40%), and had more chronic medical conditions compared with unvaccinated case-patients (median age = 55 years; immunocompromised = 10%; p<0.001 for both). VE against IMV or in-hospital death was 90% (95% CI = 88%–91%) overall, including 88% (95% CI = 86%–90%) for 2 doses and 94% (95% CI = 91%–96%) for 3 doses, and 94% (95% CI = 88%–97%) for 3 doses during the Omicron-predominant period. COVID-19 mRNA vaccines are highly effective in preventing COVID-19–associated death and respiratory failure treated with IMV. CDC recommends that all persons eligible for vaccination get vaccinated and stay up to date with COVID-19 vaccination (4).

Using surveillance data from the Influenza and Other Viruses in the Acutely Ill (IVY) Network, a case-control analysis was conducted to evaluate effectiveness of mRNA COVID-19 vaccines against COVID-19–associated IMV or death. During March 11, 2021–January 24, 2022, adults aged ≥18 years hospitalized at 21 medical centers in 18 states† who received testing for SARS-CoV-2 were enrolled. Case-patients were adults who were hospitalized with COVID-19–like illness§ and who received positive SARS-CoV-2 nucleic acid amplification test (NAAT) or antigen test results within 10 days of illness onset. Case-patients in this analysis were limited to those who received IMV or died in the hospital within 28 days of admission. Control-patients were hospitalized adults with or without COVID-19–like illness who received a negative NAAT test result for SARS-CoV-2 within 10 days of illness onset. Individual matching was not performed, but sites attempted 1:1 enrollment of case-patients and controls, with controls enrolled within 2 weeks of case-patients. Patients or their proxies were interviewed about demographic and clinical characteristics and COVID-19 vaccination history. COVID-19 mRNA vaccination status (i.e., receipt of Pfizer-BioNTech or Moderna vaccine products) was ascertained from state registry data, hospital electronic medical records, vaccination record cards, and self-report. For this analysis, patients were included if they 1) received 2 doses of an mRNA vaccine, with the second dose administered ≥14 days before illness onset, 2) received 3 doses of an mRNA vaccine following authorization¶ with the third dose administered ≥7 days before illness onset, or 3) received no COVID-19 mRNA vaccine doses before illness onset. Information about chronic medical conditions and in-hospital outcomes, including IMV or death within 28 days of admission, were collected through structured chart reviews.

Differences in demographic and clinical characteristics between COVID-19 case-patients who were vaccinated with 2 or 3 vaccine doses versus unvaccinated were compared using Pearson’s chi-square for categorical variables or Wilcoxon rank-sum tests for continuous variables. VE was calculated using unconditional logistic regression by comparing the odds for previous mRNA vaccination (2 or 3 doses) among COVID-19 case-patients who received IMV or experienced in-hospital death versus control-patients. VE was calculated as (1 − odds ratio) × 100, and estimates were adjusted for U.S. Health and Human Services region, calendar time in biweekly intervals, age, sex, and self-reported race and Hispanic ethnicity as prespecified covariates. Results were stratified by age, immunocompromising conditions, number of categories of chronic medical conditions,** number of COVID-19 mRNA vaccine doses received, and variant-predominant period when admitted to hospital. Variant-predominant periods were defined as pre-Delta (March 11–July 3, 2021), Delta (July 4–December 25, 2021), or Omicron (December 26, 2021–January 24, 2022), based on when a variant accounted for >50% of sequenced SARS-CoV2 viruses using on whole-genome sequencing of specimens collected in the IVY network. An additional sensitivity analysis was conducted by restricting COVID-19–negative controls to those known to have received IMV or to have died in the hospital within 28 days of admission. Analyses were conducted using STATA software (version 16.0; StataCorp); p-values <0.05 were considered statistically significant. This activity was determined to be public health surveillance by each participating site and CDC and was conducted in a manner consistent with applicable federal law and CDC policy.††

Among 9,211 COVID-19 case-patients with IMV or in-hospital death and COVID-19–negative controls enrolled during March 11, 2021–January 24, 2022, 1,667 (18%) were excluded from the analysis. The most common reasons for exclusion included receiving a licensed mRNA COVID-19 vaccine but not being in a vaccination group considered in this analysis (638), receiving a non-mRNA COVID-19 vaccine product (445), inability to determine vaccination status (279), COVID-19–like illness onset after hospital admission (119), and receiving a third vaccine dose before authorization (96); 90 patients were excluded for other reasons. Among 7,544 included patients, 1,440 (19%) were COVID-19 case-patients with IMV, death, or both, and 6,104 (81%) were COVID-19–negative controls. Compared with unvaccinated case-patients with IMV or in-hospital death, those who were vaccinated (2 or 3 doses) were older (median age 69 versus 55 years; p<0.001), more likely to live in a long-term care facility (11% versus 2%; p<0.001), more likely to have been hospitalized previously in the past year (44% versus 22%; p<0.001), more likely to have immunocompromising conditions (40% versus 10%; p<0.001), and had more chronic medical conditions (Table 1).

Overall VE against COVID-19–associated IMV or death across the surveillance period was 90% (95% CI = 88%–91%) (Table 2), similar to that for IMV only (91%; 95% CI = 89%–92%) and in-hospital death only (88%; 95% CI = 85%–90%), and similar in a sensitivity analysis restricting COVID-19 test-negative control-patients to those who also received IMV or died in the hospital (86%; 95% CI = 82%–89%). Among recipients of 2 vaccine doses, VE over the entire study period was 92% (95% CI = 90%–94%) at 14–150 days after receipt of the second dose versus 84% (95% CI = 80%–87%) at >150 days postvaccination. VE was 94% (95% CI = 91%–96%) among recipients of 3 vaccine doses. Among immunocompetent adults with no chronic medical conditions, VE for 2 or 3 vaccine doses was 98% (95% CI = 97%–99%). VE was lowest among adults with immunocompromising conditions (74%; 95% CI = 64%–81%). However, among 123 vaccinated COVID-19 case-patients with immunocompromising conditions, only 17 (14%) had received 3 vaccine doses and were considered fully vaccinated.§§ During the Omicron period, VE against IMV or in-hospital death was 79% (95% CI = 66%–87%) for recipients of 2 doses and 94% (95% CI = 88%–97%) for recipients of 3 doses.