AB004. All-cause mortality of seasonal influenza vaccination among the elderly: a systematic review and meta-analysis

Background: Despite the recommendation of seasonal influenza vaccination (SIV) for the elderly, vaccine effectiveness (VE) remains doubtful among this age group. Evidence regarding VE is needed for justification and evaluation. We sought to develop pooled VE estimates using all-cause mortality as an outcome measure to provide better insights for VE evaluation.

Methods: A systematic review was performed to identify observational studies reporting all-cause mortality as an outcome measure for the VE of SIV. We searched MEDLINE (Ovid) and EMBASE from inception to June 2020. Meta-analysis with random effects model for all-cause mortality was conducted. Heterogeneity and publication bias were assessed.

Results: We identified 8 observational studies for systematic review and meta-analysis from a total of 1,782 studies in the initial search. The pooled VE estimate of SIV among the elderly considering all-cause mortality as the outcome was 22% [95% confidence interval (CI): 12% to 31%, P<0.05]. Significant heterogeneity (I²=95.7%) was demonstrated. A large P value (0.969) in Egger’s test and symmetrical funnel plot were observed.

Conclusions: Pooled VE estimate of SIV considering all-cause mortality as the outcome was derived. The heterogeneity of our study could be due to differences among study characteristics and designs, vaccine strain match, underlying conditions and previous vaccination status. The funnel plot suggests no significant publication biases, while this study may still be prone to other possible biases such as reporting bias. Overall, a review of current evidence demonstrated significant VE of SIV among the elderly by reducing all-cause mortality, which could be utilised in the encouragement of SIV uptake by public health officials in the future.

Keywords: Influenza; vaccination; systematic review; meta-analysis