Safety, efficacy, and immunogenicity of an inactivated influenza vaccine in healthy adults: a randomized, placebo-controlled trial over two influenza seasons

BACKGROUND Seasonal influenza imposes a substantial personal morbidity and societal cost burden. Vaccination is the major strategy for influenza prevention; however, because antigenically drifted influenza A and B viruses circulate annually, influenza vaccines must be updated to provide protection against the predicted prevalent strains for the next influenza season. The aim of this study was to assess the efficacy, safety, reactogenicity, and immunogenicity of a trivalent inactivated split virion influenza vaccine (TIV) in healthy adults over two influenza seasons in the US. METHODS The primary endpoint of this double-blind, randomized study was the average efficacy of TIV versus placebo for the prevention of vaccine-matched, culture-confirmed influenza (VMCCI) across the 2005-2006 and 2006-2007 influenza seasons. Secondary endpoints included the prevention of laboratory-confirmed (defined by culture and/or serology) influenza, as well as safety, reactogenicity, immunogenicity, and consistency between three consecutive vaccine lots. Participants were assessed actively during both influenza seasons, and nasopharyngeal swabs were collected for viral culture from individuals with influenza-like illness. Blood specimens were obtained for serology one month after vaccination and at the end of each influenza season's surveillance period. RESULTS Although the point estimate for efficacy in the prevention of all laboratory-confirmed influenza was 63.2% (97.5% confidence interval [CI] lower bound of 48.2%), the point estimate for the primary endpoint, efficacy of TIV against VMCCI across both influenza seasons, was 46.3% with a 97.5% CI lower bound of 9.8%. This did not satisfy the pre-specified success criterion of a one-sided 97.5% CI lower bound of >35% for vaccine efficacy. The VMCCI attack rates were very low overall at 0.6% and 1.2% in the TIV and placebo groups, respectively. Apart from a mismatch for influenza B virus lineage in 2005-2006, there was a good match between TIV and the circulating strains. TIV was highly immunogenic, and immune responses were consistent between three different TIV lots. The most common reactogenicity events and spontaneous adverse events were associated with the injection site, and were mild in severity. CONCLUSIONS Despite a good immune response, and an average efficacy over two influenza seasons against laboratory-confirmed influenza of 63.2%, the pre-specified target (lower one-sided 97.5% confidence bound for efficacy > 35%) for the primary efficacy endpoint, the prevention of VMCCI, was not met. However, the results should be interpreted with caution in view of the very low attack rates we observed at the study sites in the 2005-2006 and 2006-2007, which corresponded to relatively mild influenza seasons in the US. Overall, the results showed that TIV has an acceptable safety profile and offered clinical benefit that exceeded risk. TRIAL REGISTRATION NCT00216242.