Background. Annual influenza vaccine is recommended for those at greatest risk of severe influenza infection. Recent reports of a negative impact of serial influenza vaccination on vaccine effectiveness (VE) raises concerns about the recommendation for annual influenza vaccines, particularly in persons at greatest risk.
Methods. The Influenza Complications Alert Network (FluCAN) is an Australian hospital-based sentinel surveillance program. In this observational study, cases were defined as subjects aged >9 years admitted with influenza confirmed by polymerase chain reaction. Controls were subjects with acute respiratory illness testing negative for influenza. Propensity scores were used to adjust for the likelihood of being vaccinated. VE was calculated as 1 - adjusted odds ratio of vaccination in cases compared with test-negative controls.
Results. Over 2010-2015, 6223 cases and 6505 controls were hospitalized with confirmed influenza and influenza test-negative acute respiratory illness, respectively. Following stratification by quintile of propensity score, site, and year, VE was estimated to be 43% (95% confidence interval [CI], 37%-49%) overall. VE was estimated to be 51% (95% CI, 45%-57%) in those vaccinated in both the current and previous season, compared with 33% (95% CI, 17%-47%) vaccinated in the current season only and 35% (95% CI, 21%-46%) in the previous season only. Similar results were observed for influenza A/H1N1, influenza A/H3N2, and influenza B strains.
Conclusions. Vaccination in both the current and previous seasons was associated with a higher VE against hospitalization with influenza than vaccination in either single season. These findings reinforce current recommendations for annual influenza vaccination, particularly those at greatest risk of influenza disease.