Duration of the serological response and effectiveness of the inactivated influenza vaccine in healthy adults aged 18-65 years: a systematic review and meta-analysis.

Background: The protection provided by the inactivated influenza vaccine (IIV) in adults can wane during a single influenza season. We aimed to assess temporal changes in haemagglutinin inhibition assay (HAI) titres and vaccine effectiveness in healthy adults after IIV.

Methods: In this systematic review and meta-analysis, we searched MEDLINE, Embase, and Cochrane CENTRAL and screened citations for studies published from database inception to July 6, 2022. We included studies that reported two or more vaccine effectiveness or HAI estimates in healthy adults (aged 18-65 years) after a single standard dose IIV with at least one effect estimate being 2 months after vaccination or longer. Unpublished studies, studies not published in English, cross-sectional studies, case reports, and modelling studies that did not report original data were excluded. We applied random-effects meta-analysis and metaregression to aggregate mean differences in log₂ HAI geometric mean titres (GMT) and odds ratio (OR) for influenza infection over time with inverse-variance weighting. We completed quality assessments using Joanna Briggs Institute critical appraisal checklists, assessed between-study variation using τ², and assessed publication bias using funnel plots. This study is registered with PROSPERO (CRD42021233774).

Findings: Of 1387 records identified, we included 35 studies in the systematic review, of which 23 studies reported vaccine-effectiveness estimates, 11 studies reported HAI titres, and one study reported both. HAI titres were significantly higher 5-7 months after IIV than before vaccination (influenza A/H1: log₂ GMT 1·57 [95% CI 1·06-2·09], τ²=0·70; influenza A/H3: 1·16 [0·70-1·62], τ²=0·46; influenza B: 0·78 [0·08-1·49], τ²=0·41). Three studies reported titres 11-13 months after vaccination with no significant difference compared with pre-vaccination titres for influenza A/H1, A/H3, and B. For influenza A/H1, the odds of influenza infection favoured vaccination more than 5 months after IIV (aOR 0·66 [95% CI 0·51-0·87], τ²=not calculable), whereas vaccination made no significant difference on influenza A/H3 or influenza B infection more than 5 months after vaccination (influenza A/H3: 0·96 [0·78-1·18], τ²=not calculable; influenza B: 0·75 [0·49-1·16], τ²=0·10). An assessment of publication bias was restricted due to the small number of studies included in each analysis; when there were sufficient studies, no publication bias was observed through visual inspection of the funnel plots.

Interpretation: IIV-induced HAI titres in healthy adults aged 18-65 years persisted 5-7 months after vaccination, but might not be significantly sustained at 11-13 months. For influenza A/H3 and influenza B, there was a significant decline in vaccine effectiveness beyond 5 months after vaccination. Evaluating the temporal changes in vaccine effectiveness remains challenging due to methodological biases; a greater understanding is needed to optimise vaccine design and timing to maximise effectiveness throughout the influenza season.

Funding: None.