- Diefenbach-Elstob, Tanya;
- Chilver, Monique;
- Spirkoska, Violeta;
- Carville, Kylie;
- Dapat, Clyde;
- Turra, Mark;
- Tran, Thomas;
- Deng, Yi-Mo;
- Peck, Heidi;
- Barr, Ian;
- Stocks, Nigel;
- Sullivan, Sheena
Background
Vaccine effectiveness (VE) estimates provide important post-marketing assessment of how well seasonal influenza vaccines prevent medically attended influenza disease. We present VE estimates for primary care in Australia for the 2017-2019 seasons.Methods
The study used a test-negative design. Influenza VE was estimated from adjusted logistic regression models comparing the odds of vaccination among influenza-test-positive cases and test-negative non-cases. Estimates were made overall and separately by influenza type, subtype, lineage and clade and stratified by age group. Antigenic similarity of influenza viruses to vaccine strains was assessed using the haemagglutination inhibition assay, and phylogenetic analysis was performed on sequenced viruses.Results
The study included 2879, 1973 and 3371 general practice patients with swabs collected during 2017, 2018 and 2019 respectively. Influenza A(H3N2) was predominant in 2017 and 2019, while influenza A(H1N1)pdm09 predominated in 2018. VE was estimated at 37% (95% CI 22, 48) for the 2017 season, 53% (95% CI 33, 67) for 2018 and 50% (95% CI 40, 58) for 2019. In general, estimates were higher against A(H1N1)pdm09 and influenza B viruses and lower against A(H3N2) viruses. Across the three seasons, antigenic data identified a greater proportion of A(H1N1)pdm09 and influenza B viruses than A(H3N2) viruses as antigenically similar to the cell-propagated reference viruses. VE estimates by clade generally indicated higher VE among viruses in the same clade as the vaccine viruses.Conclusions
Influenza VE varied across influenza seasons and by influenza type/subtype. Given the ongoing evolution of circulating influenza viruses, vaccine improvements are needed, especially for influenza A(H3N2).