Dose-Dependent Negative Effects of Prior Multiple Vaccinations Against Influenza A and Influenza B Among Schoolchildren: A Study of Kamigoto Island in Japan During the 2011–2012, 2012–2013, and 2013–2014 Influenza Seasons

Background: We investigated the negative effects of prior multiple vaccinations on influenza vaccine effectiveness (VE) and analyzed the association of VE with prior vaccine doses. Methods: Patients aged 9-18 years presenting with influenza-like illness at a community hospital on a remote Japanese island during the 2011-2012, 2012-2013, and 2013-2014 influenza seasons were tested for influenza using a rapid diagnostic test (RDT). A test-negative, case-control study design was used to estimate the VEs of trivalent inactivated influenza vaccine. Histories of vaccination and medically attended influenza (MA-flu) A and B during 3 previous seasons were collected from registry systems. VE was calculated using multilevel mixed-effects logistic regression models adjusted for the history of RDT-confirmed MA-flu. Results: During 3 influenza seasons, 1668 influenza-like illness episodes were analyzed, including 421 and 358 episodes of MA-fluA and MA-fluB, respectively. The adjusted VE (95% confidence interval) yielded significant dose-dependent attenuations by prior vaccinations against both MA-fluA (0 doses during previous 3 seasons: 96% [69%-100%], 1 dose: 48% [-7% to 74%], 2 doses: 52% [11%-74%], 3 doses: 21% [-25% to 51%]; P for trend < .05) and MA-fluB (0 doses: 66% [-5% to 89%], 1 dose: 48% [-14% to 76%], 2 doses: 34% [-33% to 67%], 3 doses: -7% [-83% to 37%]; P for trend < .05). After excluding episodes of MA-flu during prior 3 seasons, similar trends were observed. Conclusions: Repeated previous vaccinations over multiple seasons had significant dose-dependent negative impacts on VE against both MA-fluA and MA-fluB. Further studies to confirm this finding are necessary.