BACKGROUND: Little is known about inactivated influenza vaccine effectiveness (IVE) in preventing very severe disease, including influenza-associated intensive care unit (ICU) admissions.

METHODS: The Southern Hemisphere Influenza and Vaccine Effectiveness Research and Surveillance (SHIVERS) project enrolled adults (aged?≥?18?years) with acute respiratory illness (ARI) in general ward (GW) hospital settings (n?=?3034) and ICUs (n?=?101) during 2012-2015. IVE was assessed using a test-negative design comparing the odds of influenza vaccination among influenza positives vs. negatives (confirmed by real-time reverse transcription polymerase chain reaction). All models were adjusted for season, weeks from season peak, and a vaccination propensity score.

RESULTS: Influenza virus infection was confirmed in 28% of GW hospital and 41% of ICU patients; influenza vaccination was documented for 56% and 41%, respectively. Across seasons, IVE was 37% (95% confidence intervals [CI]?=?23-48%) among GW patients and 82% (95% CI?=?45-94%) among ICU patients. IVE point estimates were?>?70% against ICU influenza and consistently higher than IVE against GW influenza when stratified by season, by virus (sub)types, and for adults with or without chronic medical conditions and for both adults aged <65 and ≥65?years old. Among hospitalized influenza positives, influenza vaccination was associated with a 59% reduction in the odds of ICU admission (aOR?=?0.41, 95% CI?=?0.18-0.96) and with shorter ICU lengths of stay (LOS), but not with radiograph-confirmed pneumonia or GW hospital LOS.

CONCLUSION: Inactivated influenza vaccines prevented influenza-associated ICU admissions, may have higher effectiveness in ICU than GW hospital settings, and appeared to reduce the risk of severe disease among those who are infected despite vaccination.