Intraseason decline in influenza vaccine effectiveness during the 2016 southern hemisphere influenza season: A test-negative design study and phylogenetic assessment. Issue 19 (1st May 2019)

• Record Type: Journal Article
• Authors: Regan, Annette K.; Fielding, James E.; Chilver, Monique B.; Carville, Kylie S.; Minney-Smith, Cara A.; Grant, Kristina A.; Thomson, Chloe; Hahesy, Trish; Deng, Yi-Mo; Stocks, Nigel; Sullivan, Sheena G.
• Is Part Of: Vaccine. Volume 37:Issue 19(2019)
• Page Start: 2634
• Page End: 2641
• Publication Date: 2019-05-01
• Subjects: Influenza -- Influenza vaccines -- Test-negative design -- Vaccine effectiveness -- Case-control study; ASPREN Australian Sentinel Practices Research Network -- HA haemagglutinin -- ILI influenza-like illness -- NA neuraminidase -- RT-PCR reverse transcriptase-polymerase chain reaction -- SPNWA Sentinel Practitioners Network of Western Australia -- VE vaccine effectiveness -- VicSPIN Victoria Sentinel Practice Influenza Network; Vaccines -- Periodicals; 615.372
• DOI: 10.1016/j.vaccine.2019.02.027
• Languages: English
• ISSNs: 0264-410X

Abstract

Highlights: A test-negative design was used to estimate effectiveness of the 2016 influenza vaccine. Among 1085 general practice patients, influenza vaccine was 40% effective against infection. As time between vaccination and symptom onset increased, vaccine effectiveness (VE) decreased. The VE point estimate was highest (60%) when vaccination occurred <3 months prior to symptom onset. Vaccination close to the influenza season may provide optimal protection. Abstract: Background: We estimated the effectiveness of seasonal inactivated influenza vaccine and the potential influence of timing of immunization on vaccine effectiveness (VE) using data from the 2016 southern hemisphere influenza season. Methods: Data were pooled from three routine syndromic sentinel surveillance systems in general practices in Australia. Each system routinely collected specimens for influenza testing from patients presenting with influenza-like illness. Next generation sequencing was used to characterize viruses. Using a test-negative design, VE was estimated based on the odds of vaccination among influenza-positive cases as compared to influenza-negative controls. Subgroup analyses were used to estimate VE by type, subtype and lineage, as well as age group and time between vaccination and symptom onset. Results: A total of 1085 patients tested for influenza in 2016 were included in the analysis, of whom 447 (41%) tested positive for influenza. The majority of detections were influenza A/H3N2 (74%). One-third (31%) of patients received the 2016 southern hemisphere formulation influenza vaccine. Overall, VE was estimated at 40% (95% CI: 18–56%). VE estimates were highest for patients immunized within two months prior to symptom onset (VE: 60%; 95% CI: 26–78%) and lowest for patients immunized >4 months prior to symptom onset (VE: 19%; 95% CI: −73–62%). Discussion: Overall, the 2016 influenza vaccine showed good protection against laboratory-confirmed infection among general practice patients. Results by duration of vaccination suggest a significant decline in effectiveness during the 2016 influenza season, indicating immunization close to influenza season offered optimal protection.