Early estimates of seasonal influenza vaccine effectiveness in Europe: results from the i-movE multicentre case–control study, 2012/13

We conducted a test-negative case–control study based in five European sentinel surveillance networks. The early 2012/13 adjusted influenza vaccine effectiveness was 78.2% (95% CI: 18.0 to 94.2) against influenza B, 62.1% (95% CI: −22.9 to 88.3%) against A(H1) pdm09, 41.9 (95% CI: −67.1 to 79.8) against A(H3N2) and 50.4% (95% CI: −20.7 to 79.6) against all influenza types in the target groups for vaccination. Efforts to improve influenza vaccines should continue to better protect those at risk of severe illness or complications.


Background
Since 2008/9, the Influenza Monitoring Vaccine Effectiveness (I-MOVE) network has estimated the effectiveness of seasonal influenza vaccine to prevent medically attended influenza-like illness (ILI) laboratory confirmed as influenza [1][2][3][4][5][6][7].One of the components of I-MOVE is a multicentre case-control study based on practitioners participating in the European Influenza Surveillance Network (EISN) [8].This season, six study sites are participating in the multicentre study: Germany, Ireland, Poland, Portugal, Romania and Spain.
Here we provide early season estimates of the 2012/13 influenza vaccine based on data collected from week 43 2012 to week 3 2013.Poland is not included in the early season estimates as there were no vaccinated individuals among the patients recruited.
Participating practitioners swabbed and interviewed all or a systematic sample of patients consulting for ILI.The common variables collected in all study sites were symptoms, date of onset and swabbing, 2012/13 seasonal vaccination status and date of vaccination, sex, age, presence of chronic conditions and number of hospitalisations for the chronic condition in the past 12 months.Four of the five study sites included a question on belonging to the target groups for vaccination.In Portugal, this information was gathered using information on age, chronic conditions, pregnancy, whether the patient was a health professional or carer and a household member or carer of an at-risk patient aged less than six months.
In the pooled analysis, we included patients with a nasopharyngeal swab taken less than eight days after symptom onset and meeting the European Union case definition for ILI: sudden onset of symptoms, at least one of four systemic symptoms (fever or feverishness, malaise, headache, myalgia) and at least one of three respiratory symptoms (cough, sore throat, shortness of breath) [10].A case of confirmed influenza was an ILI patient who was swabbed and tested positive for influenza virus using real-time reverse-transcription polymerase chain reaction or culture.Controls were ILI patients who tested negative for any influenza virus.
We defined a person as vaccinated if he or she had received at least one dose of a 2012/13 seasonal influenza vaccine more than 14 days before ILI symptom onset.All the others were classified as unvaccinated.
For each study site we included ILI patients with symptom onset up to the end of week 3 2013 and more than 14 days after the start of national or regional influenza vaccination campaigns.
We conducted a complete case analysis excluding individuals with missing information on key confounders.We estimated the pooled seasonal influenza vaccine effectiveness (VE) as 1 minus the odds ratio (OR) expressed as a percentage, using a one-stage method with the study site as fixed effect in the model.
We used a logistic regression model to adjust for presence of at least one chronic disease, sex, age group and month of symptom onset.
We estimated VE against all laboratory-confirmed influenza, and individually against influenza A(H3N2), A(H1N1)pdm09 and influenza B. For each study site, we excluded controls with symptom onset in the weeks before symptom onset of the first influenza type/ subtype case depending on the outcome used.We also estimated VE restricted to the target groups for vaccination.

Early estimates of 2012/13 influenza vaccine effectiveness
In the five countries, the influenza season, as defined by national thresholds, started at different times -in week 50 2012 in Germany and Ireland, week 3 2013 in Spain and Romania (Table 1).In week 3, the ILI incidence in Portugal was still below the threshold defining the start of the season.The first study site to recruit cases was Germany (in week 43 2012) and the last Romania (in week 2 2013).
Of the 1,161 ILI patients enrolled, 271 tested positive for influenza A (23.3%), 162 for influenza B (14.0%) and one tested positive for both influenza B and influenza A(H1)pdm09.Among the 269 influenza A viruses that could be subtyped, 146 (54.3%) were A(H1)pdm09 and 123 (45.7%) were A(H3N2).Influenza A virus was predominant in Germany, and influenza B in Ireland and Spain (Table 2).

Discussion
These early estimates suggest a moderate VE against all influenza viruses.By type and subtype, the highest VE was against influenza B and the lowest against influenza A(H3N2).a Controls used to compare with all influenza cases.b One influenza case positive for influenza B and for influenza A(H1)pdm09 was included in both analyses.c The virus from three influenza A cases could not be subtyped: these cases are not included in the descriptive analysis.d Vaccination more than 14 days before onset of influenza-like illness symptoms.Three records were excluded for October.
During the study period, as of week 3-2013, two of the participating countries reported a low ILI/ARI activity (Portugal, Romania) and three medium activity (Germany, Ireland, Spain) [11].The sample size varied by study site.As most patients (863/1,161, 74.3%) were enrolled in Germany, the pooled estimates are highly influenced by the German data.When restricting the analysis to Ireland, Portugal, Romania and Spain (n=296, four vaccinated cases), the adjusted point VE against all influenza was higher (73.5%; 95% CI: 8.7 to 92.3) than the adjusted point VE including the five study sites.
As in 2011/12, the results suggests a low-to-moderate VE for influenza A(H3N2) [9].Our point estimate is lower than that reported by Canada [12] and the United States [13], countries with a predominance of this subtype in the early phase of the 2012/13 season.Most of the influenza A(H3N2) cases included in our study were German patients.In Germany, most of

Number of records received for pooled analysis:
1,622 Excluding records: 1,161 Excluding records with missing data from covariates for complete case analysis: • Persons with missing information on age (n=2) • Person with missing information on sex (n=11) • Persons with missing information on chronic conditions (n=62) the 103 influenza A(H3N2) viruses characterised at the national reference centre for influenza were similar to the A/Victoria/361/2011(H3N2) vaccine virus for the 2012/13 season [14].This similarity was also reported by the Community Network of Reference Laboratories for Human Influenza in Europe (CNRL) [15].This discrepancy between the apparently well-matched vaccine strain/circulating strains and low VE has also been noted this season in Canada [12].
Our 2013 early point VE estimates for influenza B are similar to estimates from the United States [13] and higher than those reported in the United Kingdom [16].
All the influenza B virus isolates genetically characterised from influenza cases enrolled in the Irish (three isolates), Portuguese (three isolates) and Spanish (four isolates) study sites were B/Yamagata, the lineage included in the 2012/13 vaccine.Data from the German national reference centre for influenza indicated that among 75 influenza B strains characterised, 68 were Yamagata and 7 Victoria.
The sample size did not allow VE estimation by type and subtype among the target population for vaccination.The low VE against all influenza types in this population is similar to the estimates the I-MOVE multicentre case-control study provided last season against influenza A(H3) [3].
This season, in which different influenza viruses are co-circulating in Europe, the I-MOVE multicentre casecontrol study provided early adjusted VE estimates for influenza B, A(H3N2) and A(H1N1)pdm09 viruses.However, due to small sample size, the precision around these estimates is low and should be taken into account when interpreting these preliminary results.
The results underscore the importance of providing early VE estimates against virus subtype regardless of the reported relatedness between circulating viruses and administered vaccines.The early VE estimates could be useful when defining the recommendations for next season's vaccine composition.
In conclusion, our early season estimates suggest that the 2012/13 influenza vaccine is effective in preventing medically attended laboratory-confirmed influenza, with a higher VE against influenza B than against influenza A subtypes.The lower VE among the target groups and against influenza A(H3N2) underlines that efforts to improve the influenza vaccine should continue in order to better protect those at risk of severe illness or complications.

Members of the I-MOVE case-control study team
-like illness; ISO: International Organization for Standardization.a According to the thresholds used to define the start of the influenza season in each of the countries.b ILI patients meeting the European Union case definition, swabbed less than eight days after onset of symptoms within the study period.cFrom 15 days after the start of the vaccination campaign to week 3/2013.We excluded controls with an onset of symptoms in the weeks before the first influenza case in the study site.d ILI patients included in the study, after excluding those with missing information on laboratory results, vaccination status or date of vaccination.

Figure
FigureFlowchart of data exclusion for pooled analysis, I-MOVE multicentre case-control study, influenza season 2012/13

Table 1
Study details, multicentre case-control study, study sites in five European Union countries, week 43/2012-week 3/2013

Table 3
Pooled crude and adjusted 2012/13 seasonal vaccine effectiveness against laboratory confirmed influenza by influenza type/ suptype, overall and among target groups for vaccination.Multicentre case-control study in five European Union study sites, week 43 (2012)-week 3 (2013), influenza season 2012/13 a Study site included in the model as fixed effect.b Model adjusted for presence of at least one chronic disease, sex, 10-year age group and month of symptom onset.c Model adjusted for presence of at least one chronic disease, sex, 10-year age group until age 60, where age is coded as ≥60 years and month of symptom onset.d Model adjusted for presence of at least one chronic disease, sex, age group (0-14; 15-59 and ≥60 years) and month of symptom onset.

Excluding records with missing 2012/13 influenza vaccination status (n=40):
Persons not meeting the EU ILI case definition (n=296)• Persons with interval between onset of symptoms and swabbing >7 days or exact date of onset not known (n=75) • Persons who are controls presenting before ISO week of first influenza case and after ISO week of last influenza case (weeks of symptom onset) (n=88)