Clin Infect Dis. 2021 Jul 28:ciab654. doi: 10.1093/cid/ciab654. Online ahead of print.
BACKGROUND: Evidence for vaccine effectiveness (VE) against influenza-associated pneumonia has varied by season, location, and strain. We estimate VE against hospitalization for radiographically identified influenza-associated pneumonia during 2015-2016 to 2017-2018 seasons in the US Hospitalized Adult Influenza Vaccine Effectiveness Network (HAIVEN).
METHODS: Among adults aged ≥18 years admitted to 10 US hospitals for acute respiratory illness (ARI), clinician-investigators used keywords from reports of chest imaging performed during 3 days around hospital admission to assign a diagnosis of 'definite/probable pneumonia'. We used a test-negative design to estimate VE against hospitalization for radiographically identified laboratory-confirmed influenza-associated pneumonia, comparing RT-PCR confirmed influenza cases with test-negative subjects. Influenza vaccination status was documented in immunization records or self-reported, including date and location. Multivariable logistic regression models were used to adjust for age, site, season, calendar-time, and other factors.
RESULTS: Of 4,843 adults hospitalized with ARI included in the primary analysis, 266 (5.5%) had 'definite/probable pneumonia' and confirmed influenza. Adjusted VE against hospitalization for any radiographically confirmed influenza-associated pneumonia was 38% (95% confidence interval [CI]): 17%-53%); by type/subtype, it was 74% (95% CI: 52%-87%), influenza A (H1N1)pdm09; 25% (-15% to 50%), A (H3N2); and 23% (95% CI: -32% to 54%), influenza B. Adjusted VE against intensive care for any influenza was 57% (95% CI, 19%-77%).
CONCLUSIONS: Influenza vaccination was modestly effective among adults in preventing hospitalizations and the need for intensive care associated with influenza pneumonia. VE was significantly higher against A (H1N1)pdm09 and was low against A (H3N2) and B.