Thirty studies (88,468 participants) were included in the review. Five studies were of low quality, seven moderate quality and 18 studies were high quality.
One hundred and one analyses were performed, with 93 evaluating heterogeneity. Significant statistical heterogeneity was identified for 46 of 93 analyses (full details reported in the review).
The overall efficacy of influenza vaccine was 65% compared to controls (RR 0.35, 95% CI 0.29 to 0.43; Ι²=75%; 23 studies).
Efficacy of vaccines was highest against infections that matched the vaccine formulation (RR 0.22, 95% CI 0.16 to 0.30; Ι²=72%; 10 studies) and lowest for infections that did not match the vaccine (RR 0.45, 95% CI 0.31 to 0.65; Ι²=69%; nine studies). Similarly, vaccine efficacy was highest compared to controls when there was good vaccine match to the circulating seasonal strains (RR 0.32, 95% CI 0.22 to 0.46; Ι²=82%; 11 studies) and lowest when vaccines were poorly matched to seasonal strains (RR 0.41, 95% CI 0.28 to 0.61; Ι²=65%; eight studies).
There was evidence of effect modification according to vaccine type and age group, influenza type and type of case ascertainment method (results fully reported in the review).
Sensitivity analyses in only high quality studies did not statistically significantly alter the findings.
There was no evidence of publication bias.