Abstract
BACKGROUNDPortugal is establishing a surveillance system for severe acute respiratory infection (SARI). Data from its existing influenza surveillance system in intensive care units (ICU), operating since 2012, has not yet been analysed.AIMWe aimed to identify individual and seasonal determinants of death from influenza in ICUs to inform ICU capacity planning, triage and SARI surveillance.METHODSWe conducted a retrospective cohort study of laboratory-confirmed influenza cases admitted to 27 ICUs between 2012 and 2024. Covariates included demographics, comorbidities, season, influenza type, seasonal and weekly influenza caseload in the ICU and weekly ICU occupancy. We calculated case fatality rates and adjusted risk ratios (aRR) for death during ICU admission using log-binomial regression. Directed acyclic graphs informed model-specific adjustments, including age, sex, influenza type and comorbidities.RESULTSOf 1,071 cases with known outcome, 262 (24%) died. Case fatality rates were higher among patients with chronic liver disease (aRR: 2.0; 95% CI: 1.5-2.6), cancer (aRR: 1.6; 95% CI: 1.1-2.1) and during a high caseload season (aRR: 1.52; 95% CI: 1.16-2.05). Case fatality rates increased with age and were highest for those aged ≥80 years (aRR: 11; 95% CI: 2.4-184), compared with 0-19-year-olds.CONCLUSIONLiver disease, cancer and older age were associated with increased fatality. Case fatality was higher in seasons with higher caseloads but showed no significant variation within seasons and did not increase during influenza peaks. These findings inform ICU triage and capacity planning for future seasons and support the implementation of broader SARI surveillance.