Duration of Influenza Vaccine Effectiveness: A Systematic Review, Meta-analysis, and Meta-regression of Test-Negative Design Case-Control Studies

Abstract

BACKGROUND: Whether influenza vaccination offers protection for the duration of an influenza season was called into question recently after analysis of data from test-negative design (TND) case-control studies. METHOD: The published literature was systematically reviewed to identify TND studies that estimated the change in vaccine effectiveness (VE) with respect to time since vaccination. RESULTS: Fourteen studies were identified through the literature search as meeting eligibility criteria. Meta-analyses were performed to compare VE 15-90 days after vaccination to VE 91-180 days after vaccination. A significant decline in VE was observed for influenza virus subtype A/H3 (change in VE, -33; 95% confidence interval [CI], -57 to -12) and type B (change in VE, -19; 95% CI, -33 to -6). VE declined for influenza virus subtype A/H1, but this difference was not statistically significant (change in VE -8; 95% CI, -27 to 21). A multivariable mixed-effects meta-regression model indicated that the change VE was associated with the proportion of study participants who were cases and the proportion who were vaccinated controls (P < .05). This could reflect biological effects such as (1) mismatch between the vaccine received and the circulating strains (among cases), (2) herd immunity (among controls), or (3) the reduced power of individual TND studies in the later parts of an influenza outbreak. CONCLUSIONS: Exploration of new influenza vaccination strategies must be a priority for influenza control, particularly in tropical countries with year-round influenza virus activity.