Early estimates of seasonal influenza vaccine effectiveness in Europe , 2010 / 11 : I-MOVE , a multicentre case – control study

We present early estimates (up to week 4 of 2011) of the 2010/11 seasonal influenza vaccine effectiveness in preventing medically attended influenza-like illness (ILI) laboratory confirmed as influenza. Practitioners from seven European sentinel networks systematically swabbed ILI patients. We included patients meeting the European Union ILI case definition and swabbed less than eight days after symptom onset. Laboratoryconfirmed influenza cases were compared with negative controls. The adjusted vaccine effectiveness was 42.3% (95% CI: –7.3 to 69.0%), suggesting moderate protection of the seasonal vaccine.


Background
The Influenza Monitoring Vaccine Effectiveness in Europe (I-MOVE) network was established in 2007 by the European Centre for Disease Prevention and Control (ECDC) to monitor seasonal and pandemic influenza vaccine effectiveness [1][2][3].In the 2010/11 season, to estimate the effectiveness of the seasonal vaccine in preventing medically attended influenzalike illness (ILI) laboratory confirmed as influenza we undertook a multicentre case-control study based on sentinel practitioner surveillance networks from eight study sites (France, Hungary, Ireland, Italy, Romania, Poland, Portugal and Spain).We report the preliminary results from seven study sites (data from France are not included in this preliminary analysis as data collection is ongoing).

Data collection and analysis
We used similar methods to those used in the first two seasons of I-MOVE [1,3].The studies were conducted within the context of the existing European Influenza Surveillance Network (EISN) [4].
The study population consisted of patients consulting a participating practitioner for ILI within eight days after symptom onset.Practitioners systematically selected ILI patients to swab.
A case of confirmed influenza was an ILI patient (defined according to the European Union case definition [5]) who was swabbed and tested positive for influenza using real-time polymerase chain reaction (PCR) or culture.Controls were ILI patients who were swabbed and tested negative for any influenza virus.Individuals were considered vaccinated if they had received a dose of the seasonal vaccine more than 14 days before the date of onset of ILI symptoms.Participating sentinel practitioners interviewed ILI patients to collect information on ILI signs and symptoms, date of onset of symptoms, current vaccination status (including date of vaccination), prior seasonal and pandemic influenza vaccination status and a list of potential confounding factors: age, sex, presence of chronic condition(s), severity of chronic disease(s) using the number of hospitalisations for the chronic disease(s) in the previous 12 months as a proxy, smoking history (non-smoker, past, current smoker), number of practitioner visits in the previous 12 months.We included in the study patients recruited up to the end of week 4 of 2011, meeting the European ILI case definition with onset of symptoms more than 14 days after the start of national 2010/11 influenza vaccination campaigns.In each study, we excluded controls with symptom onset in the weeks before the week of symptom onset of the first confirmed influenza case of the season and individuals with missing information on laboratory results.In addition, for effectiveness of the vaccine in preventing influenza A(H1N1)2009 virus infection, we excluded any individual positive for other influenza virus types and excluded controls with symptom onset in the weeks before the week of symptom onset of the first case of influenza A(H1N1)2009 virus infection recruited in the 2010/11 season.
We estimated the pooled seasonal influenza vaccine effectiveness as one minus the odds ratio (OR) (expressed as a percentage) using a one-stage method with the study site as fixed effect in the model.To estimate adjusted vaccine effectiveness, we used logistic regression models including all potential confounding factors.
We first conducted the analysis excluding all individuals with at least one missing value (complete case analysis).We then estimated missing data for vaccination status and covariates using the multiple multivariate imputation by chained equations procedure in Stata [6].We used missing at random assumptions.We used all predictors together to impute the missing values and independently analysed 20 copies of the data using 30 cycles of regression.

Estimates of seasonal influenza vaccine effectiveness
A total of 585 practitioners agreed to participate in the study; 352 of them (60%) recruited at least one ILI patient (Table 1).After excluding 71 individuals with missing information on laboratory results, a total of 1,671 ILI patients were included in the analysis: 846 cases and 825 controls (Figure 1).Among the cases, 649 (76.7%) were positive for influenza A(H1N1)2009 virus, nine (1.1%) for influenza A(H3N2) virus, 15 (1.8%) were positive for influenza A virus that could not be subtyped and 173 (20.5%) were positive for influenza B virus.
Among 1,658 individuals with information on vaccination status and vaccination date for seasonal vaccination in 2010/11, 116 (7.0%) were vaccinated (ranging from 2.2% in Poland and Ireland to 19.9% in Italy).
The median age was lower in cases (29 years, standard deviation (SD): 18 years) than in controls (34 years, SD: 21 years) (Table 2).The delay between onset of symptoms and swabbing was slightly shorter in cases (mean: 1.8 days, range: 0-7 days) than in controls (mean: 1.9 days, range: 0-7 days).The proportion of individuals presenting with fever, malaise, headache, myalgia or cough was higher among cases than among controls (Table 2).Compared with cases, a higher proportion of controls had diabetes, heart disease or were hospitalised at least once for their chronic disease in the previous 12 months.A higher proportion of controls were current or past smokers, vaccinated with the 2009/10 seasonal influenza vaccine, and vaccinated with the 2009/10 pandemic influenza vaccine.The median number of practitioner visits in the previous 12 months was two for cases (ranging from 0 to 26) and three for controls (ranging from 0 to 60) (Table 2).
A total of 34 cases were vaccinated with the 2010/11 seasonal vaccine.In two of the seven studies there were no vaccinated individuals among the recruited cases.

Discussion
Our early pooled estimates suggest that the 2010/11 seasonal vaccine conferred moderate protection against medically attended laboratory-confirmed influenza.These results should be interpreted with caution, however, for reasons including low vaccine coverage and potential biases due to the test-negative design, confounding factors, missing values and small sample size due to the early estimation in the season.Those biases have been described elsewhere in detail [3,7].
Our estimates of the 2010/11 seasonal vaccine effectiveness apply to the study period (until the end of week 4 of 2011).They are based on data from seven European study sites sharing the same protocol and definition of variables.The pooled point estimates of vaccine effectiveness were between 35% (adjusted) and 61% (crude).We adjusted for most of the confounding factors described in the literature (see, for example, [7]).The adjusted vaccine effectiveness was lower than the crude vaccine effectiveness (absolute differences ranging from 16.2% to 24.7%), suggesting some positive confounding.The pooled early estimates are similar to those observed in the United Kingdom [9], the Navarre region in Spain [8] and the cycEVA study in Spain [10].Later in the season, the larger sample size per country will allow us to conduct precise pooled and stratified analyses and to further explore the difference in effectiveness of the seasonal vaccine with that of the 2009/10 pandemic vaccine.In addition, the use of validation subsets in France, in which we collect more accurate and additional information in a subsample of the ILI patients, will enable to base our estimates on data from eight countries.
I-MOVE is a unique network in Europe able to measure seasonal and pandemic vaccine effectiveness.The early estimates presented here suggest that the seasonal vaccine has a lower effectiveness than that observed with the monovalent pandemic vaccine [3].

Table 1
Practitioners' participation, influenza-like illness (ILI) patients recruited by case-control status, vaccination status and study site, multicentre case-control study, seven European Union country study sites, week 45 (2010)-week4 (2011) ISO : International Organization for Standardization.a ILI patients meeting the European Union case definition, swabbed less than eight days after onset of symptoms within the study period.b From 15 days after the start of the seasonal influenza vaccination campaign to the week of symptom onset of the last case recruited.Controls with an onset of symptoms in the weeks before the first case were excluded.c ILI patients in the study after excluding those with missing information on laboratory results, vaccination status or date of vaccination.ILI: influenza-like illness.a International Organization for Standardization (ISO) definition of a week.

Number of days between symptom onset and swabbing
GP: general practitioner.a Unless otherwise indicated.b Non-parametric test of the median.c Two-sided Fisher's exact test.
The main confounders identified were seasonal influenza vaccination in the previous season and age group.] or by a different study population.The ILI cases included in the 2009/10 I-MOVE multicentre case-control study were younger (mean age: 12 years for cases and 24 for controls) than those included in this 2010/11 early analysis.