Initial surveillance of 2009 influenza A(H1N1) pandemic in the European Union and European Economic Area, April – September 2009

European Union (EU) and European Economic Area (EEA) countries reported surveillance data on 2009 pandemic influenza A(H1N1) cases to the European Centre for Disease Prevention and Control (ECDC) through the Early Warning and Response System (EWRS) during the early phase of the 2009 pandemic. We describe the main epidemiological findings and their implications in respect to the second wave of the 2009 influenza pandemic. Two reporting systems were in place (aggregate and case-based) from June to September 2009 to monitor the evolution of the pandemic. The notification rate was assessed through aggregate reports. Individual data were analysed retrospectively to describe the population affected. The reporting peak of the first wave of the 2009 pandemic influenza was reached in the first week of August. Transmission was travel-related in the early stage and community transmission within EU/EEA countries was reported from June 2009. Seventy eight per cent of affected individuals were less than 30 years old. The proportions of cases with complications and underlying conditions were 3% and 7%, respectively. The most frequent underlying medical conditions were chronic lung (37%) and cardio-vascular diseases (15%). Complication and hospitalisation were both associated with underlying conditions regardless of age. The information from the first wave of the pandemic produced a basis to determine risk groups and vaccination strategies before the start of the winter wave. Public health recommendations should be guided by early capture of profiles of affected populations through monitoring of infectious diseases.


Introduction
When the 2009 influenza A (H1N1) pandemic started in April 2009 and first cases appeared in Europe, aggregated (number of cases) and case-based (patient-based records) reporting systems were rapidly implemented by the European Centre for Disease Prevention and Control (ECDC), the European Union (EU) and the European Economic Area (EEA) countries to fulfil the reporting requirements of the World Health Organization (WHO) and the EU [1].The Early Warning and Response System (EWRS) was used to confidentially report aggregated and case-based data [2].The EWRS was primarily designed as a communication platform and not as surveillance application.However, one of the main advantages of the system at the beginning of the pandemic was that it relies more on a human driven approach to reporting and this allowed timely (daily) reporting of aggregated data by the EWRS focal points in the EU/EEA countries to ECDC.The European data was then rapidly published in the ECDC's daily situation reports [3] to guide and support the response of the countries and the European Commission.Laboratory-confirmed cases of pandemic influenza were reported according to the EU case definition [4] which includes laboratory confirmation by PCR, antigen detection and a four-fold rise in influenza specific antibodies.A preliminary communication in this journal in June 2009, and the 2009 pandemic influenza A(H1N1) individual case reports from 2 June to 10 August 2009 [5,6], showed that community transmission had developed in several of the EU/EEA countries since the beginning of the epidemic.A large proportion (77%) of cases was reported in children and young adults less than 30 years of age.The frequency of reported symptoms was 89% for respiratory and 14% for gastro-intestinal symptoms and for 10% of pandemic influenza cases at least one underlying medical condition was reported.A number of reports from individual countries show similar data [7][8][9][10][11][12][13][14][15].
The objective of this article is to describe the main characteristics and risk factors of pandemic influenza cases reported by EU/EEA countries during the first pandemic wave from April to September 2009.

Methods
The investigators extracted two datasets from the EWRS to provide numbers and characteristics of the populations infected by the pandemic influenza virus.Aggregated numbers of 2009 pandemic influenza A(H1N1) virus infections were reported by 30 EU/EEA countries by notification date from 27 April to 22 September 2009.Characteristics of cases were described on a weekly basis using case-based data reported from 5 May to 22 September 2009 (Figure 1).Adoption of a mitigation strategy was defined as the point when a country was no longer recommending laboratory tests for all suspected cases and therefore not all pandemic influenza cases were reported to national public health authorities.

Aggregated data
Weekly notification rates were calculated by dividing the weekly aggregated number of cases reported by EU/EEA countries by their respective population extracted from the Eurostat website in August 2009 [16].The weekly denominator only included the population of countries for as long as they reported cases to ECDC.

Individual, case-based data
The set of variables reported in the case-based system were compiled using the WHO guidance for surveillance of human infection with the 2009 pandemic influenza A(H1N1) virus [17].The variables for the characterisation of the cases were: age, sex, travel-association, vital status (alive or dead), dates (notification, onset of symptoms, treatment started and death), clinical presentation, underlying conditions, complications, antiviral treatment and prophylaxis, seasonal influenza vaccination status, and hospitalisation.Trends over time were analysed by calendar weeks (week starting on Monday).
For cases reported from 5 May to 22 September 2009, the proportion of hospitalised cases was calculated using a weekly median (by country with an interquartile range (IQR) and the 95th percentile), the distribution of travel and non travel-associated cases was described by week of onset over 22 weeks and geographic area visited, age-specific notification rates were calculated over the 20 weeks reporting period.Completeness of reporting was calculated for sex, travel-association, antiviral treatment and prophylaxis, seasonal influenza vaccination and complication.If no data was missing, completeness equalled 100%.It was not possible to calculate completeness of reporting for underlying condition as there was no option for 'none' or 'unknown' underlying condition (see list below).
Age distributions were compared between groups of persons for the variables, sex, travel-association, antiviral treatment or prophylaxis, vaccination status, underlying conditions and complications, by using two-sample Wilcoxon rank-sum (Mann-Whitney) tests.
Underlying conditions were reported according to the following pre-defined categories: cancer, diabetes mellitus, human immunodeficiency virus (HIV) infection and other immune deficiencies, heart disease, seizure disorder, lung disease, pregnancy and malnutrition.Underlying conditions could also be reported in a free-text field.When conditions reported in the free-text fields matched one of the pre-defined categories mentioned above, they were re-classified into this category.
Associations between outcomes of pandemic influenza, hospitalisation or complications, and the variables sex age, fever, respiratory/gastro-intestinal symptoms, antiviral treatment or prophylaxis, seasonal influenza vaccination status, underlying conditions, were analysed by unadjusted and adjusted (for other variables) logistic regression models using STATA software.Interactions between variables were tested by using the likelihood ratio test to assess the significance of each variable in the model.Poland, Portugal, Romania) where hospitalisation was performed mainly for isolation purposes, leading to an over-representation of mild cases among hospitalised cases, were not included in risk factor analyses (n=1,748).

Aggregated data -weekly notification rates
In total, 51,768 confirmed cases of pandemic influenza were reported as aggregated case reports by all EU/ EEA countries.The weekly notification rate was calculated for the 51,575 cases reported from 27 April to 20 September 2009 (Figure 1).It increased from week 18 to week 27 (end of June) where it peaked with eight cases per million population.A second peak in the weekly notification rate was observed in week 32, in early August, with 13.6 cases per million population, and was followed by a decrease from week 33, when countries progressively adopted mitigation strategies (Table 1, Figure 2).
The population used as a denominator for the weekly notification rate decreased after week 29, when countries stopped reporting pandemic influenza cases to ECDC.
The average weekly notification rate over the period described above was greater than 10 per million population in Cyprus, Germany, Iceland, Luxembourg, Malta, Norway, Portugal and the UK.  3).

Hospitalised cases
The median of the weekly percentage of hospitalised cases by country was 21% with an IQR of 13 to 29% and a 95th percentile of 40% in 25 EU/EEA countries.Information on hospitalisation was not reported by Iceland, Spain and Slovenia (Table 1).Seven countries were identified with a median proportion of hospitalised cases greater than 40 % (95th percentile): Austria, Bulgaria, France, Latvia, Poland, Portugal and Romania.These countries had similarly high hospitalisation rates during their containment phase of the pandemic which decreased when hospitalisation was no longer recommended for isolation purposes in these countries.4).Two peaks were observed in those under 30 years of age: the first peak, in 10 to 14 year-olds, corresponded to a series of school outbreaks reported for example in the UK and Germany [7,8].The second peak was Age group

Age, sex and antiviral treatment
Cases per 100,000 population a 212 cases are reported below 1-year-old with an age-specific rate of 4 per 100,000 population.

Symptoms and underlying conditions
Frequencies of symptoms were calculated based on 4,452 cases, after exclusion of 753 (14%) cases reported without any symptom.Fever was reported in 87%, respiratory symptoms were reported in 85%, gastro-intestinal symptoms in 18%, and for 27% of cases other symptoms, mainly fatigue or asthenia, chill, loss of appetite were noted.The proportion of gastro-enteritis was 26 % among children aged less than 10 years.
Three hundred and forty-three of 5,205 (7%) pandemic influenza cases were reported with at least one underlying condition.Underlying conditions were specified in 331 (96%) of them.They were described as free text for 137 (41%) cases.The most common underlying conditions were unspecified chronic lung diseases, including asthma (124 cases, 37%).Other underlying conditions reported and associated or not with other conditions, were cardiovascular-diseases, diabetes, gastro-intestinal diseases, allergy, liver or kidney related conditions, neurological disorders, cancer, HIV.Pregnancy was reported in 14 women (4%) (Figure 5).

Discussion
The pandemic influenza cases reported in this article characterise the first wave of the 2009 pandemic in EU/EEA countries.They include a large proportion of travel-related cases that are not necessarily representative of the population affected by the pandemic during the following winter wave.Also representativeness of data varied between countries.The weekly notification rate calculated for aggregated data is a proxy for the notification rate of pandemic influenza over the summer months of 2009.Two peaks were observed: one in week 26 and one in week 31.The first is probably due to a reporting artefact in week 26, when a large number of cases from previous weeks were reported by the UK.The second peak marks the maximum number of cases reported during the first pandemic wave in EU/ EEA countries.The sentinel surveillance of influenzalike illness (ILI) and acute respiratory infections (ARI) also showed two peaks at a time similar to that of the reporting data: one in week 25 and one in week 31 [18].
High notification rates in specific countries like Cyprus and Malta can probably be explained by an increase of their population during the summer holiday season that could not be taken into account in the denominator.
The reported percentage of hospitalised patients in (21%) seems extremely high.At the beginning of the pandemic, hospitalisation was used for isolation purposes in some countries and this may have inflated the percentage rather than a high number of severe cases.In the Netherlands, a country that did not recommend hospitalisation for isolation purposes, a hospitalisation rate of only 2.2% (35 of 1,622 patients with confirmed pandemic influenza) was reported until 14 August 2009, when a change in notification criteria to only hospitalised patients was implemented [19].
Case-based data was available for merely 21% of the reported aggregated cases.However, this was expected because the purpose of the case-based system was to capture the first few hundred cases of pandemic influenza reported in all Member States, while case-based reporting was still feasible.This purpose was achieved in most countries that have reported more than 100 cases in the aggregated reports.The completeness of data for prophylaxis (28%) and complication (26%) was low.This can be interpreted in two different ways: either the missing information corresponds to 'no prophylaxis' or 'no complication', or to unknown information.As we chose to remove missing values from the denominator, proportions of persons who have received prophylaxis or with complication(s) may be over-estimated in our analysis.Clinical presentations of patients reported in our system are similar to those listed in a review (WHO consultation) of clinical aspects of 2009 pandemic influenza [20].In September 2009, the number of cases reported without any symptom was considered as quite high (14%) as information about the proportion of asymptomatic cases was still scares at that time.Asymptomatic cases when reported in the context of tracing contacts during the containment phase could have been underestimated if contact tracing was not systematically performed.
However, it is not known if these cases were really asymptomatic or if symptoms were not reported.In the latter case, 14% would be an over-estimation of the proportion of asymptomatic cases.Serological surveys are the only way to estimate the proportion of asymptomatic 2009 pandemic influenza cases.In the meanwhile, results from such studies suggest that a considerable number of those infected with pandemic influenza A(H1N1) virus may have been asymptomatic [21,22].
The overall proportion of underlying conditions (7%) reported in our dataset is similar to the information reported by WHO for Ontario, Canada in June 2009 [23].
We compared proportions of underlying conditions with results from other studies among hospitalised patients with pandemic influenza in the United States [24,25], Canada [26], New Zealand [27] and Mexico [28] (Table 4).Although not necessarily all cases reported with underlying conditions in our dataset were hospitalised, the proportion of chronic lung diseases (including asthma) and cardio-vascular diseases among hospitalised patients were similar to those reported elsewhere [24][25][26][27].However, the proportions of cases reported with metabolic conditions (diabetes and obesity) and pregnancy are lower in EU/EEA countries than those reported in hospitalised patients in the countries mentioned above.In our dataset, patients with underlying conditions were more likely to be hospitalised and underlying conditions were associated with complications regardless of age.
The fact that 45% of our cases did not receive any treatment may either indicate that they did not have a severe condition or it may reflect the treatment policies in the countries who may have only recommended treatment for severely ill.
Most cases were found in younger or middle-aged age groups.Above the age of 60, there was a steep decline in the number of pandemic influenza A(H1N1) cases.
This could be related to previous exposure of individuals over 60 years to influenza A(H1N1) viral strains circulating after the 1918 pandemic until the 1950s [29].
Recent sero-surveys conducted in the UK [30] and in Finland [31] support this hypothesis.
Only three deaths were reported in the individual case data, this contrasts with the 159 deaths reported in EU/EEA countries in the ECDC situation report of 22 September 2009 [3].Information about deaths is essential to assess severity of the disease appropriately.Additional monitoring systems are needed to collect this type of information in a timely manner.

Conclusion
The primary focus of this article was to present the case-based data collected during the first phase of the pandemic in EU/EEA countries and their implications for rapid public health action.The case-based reporting system was stopped in September 2010, due to the associated heavy work load and the high numbers of affected people.Case-based data were not collected in the population-based system during the second phase of the pandemic and thus our data cannot be used for comparison between the two waves.Overall, our results are in line with other observations that the early phase of the pandemic mainly affected children and young adults in European countries [7][8][9][10][11][12][13][14][15].Individuals infected with 2009 pandemic influenza A(H1N1) and with underlying condition(s) were more likely to be hospitalised or to develop (severe) complications regardless of their age, particularly those with underlying respiratory diseases.The epidemiological information collected during the first wave of the pandemic provided some initial indication to determine risk groups and vaccination strategies.In the early phase of the pandemic, results from serological studies would have been helpful to determine if and to what extent individuals over 60 years have pre-existing immunity against pandemic 2009 pandemic influenza A(H1N1) from H1N1 strains circulating after the 1918 pandemic up until the 1950s.Our reporting system provided baseline data and helped to guide initial public health recommendations, however, as the profile of the affected population may have changed over time it is important to continue monitoring.The initial surveillance system was followed by a case-based reporting system of severe acute respiratory infections among influenza cases.Both systems provided timely information of public health relevance about profiles of populations affected by 2009 pandemic influenza.

Figure 4
Figure 4Age-specific notification rate of 2009 pandemic influenza A(H1N1) cases reported by28 European Union and European Economic Area countries, individual case reports, 5 May -22 September 2009, (n=10,846)

Figure 5
Figure 5 Underlying conditions of 2009 pandemic influenza A(H1N1) cases reported in 26 European Union and European Economic Area countries, 5 May -22 September 2009 (n=331)

Figure 1
Data for analyses of 2009 pandemic influenza A(H1N1) cases reported through the Early Warning and Response System to the European Centre for Disease Prevention and Control by European Union and European Economic Area countries,

Risk factor analysis (hospitalisation and complication) c
cCases for 18 countries, cases excluded from Austria, Bulgaria, France, Latvia, Poland, Portugal, Romania.
cases of 2009 pandemic influenza A(H1N1) virus infection in European Union and European Economic Area countries by week of onset and continent of travel, 19 April (week 16) -20 September (week 38) 2009

Table 2
Characteristics of 2009 pandemic influenza A(H1N1) cases reported in 28 European Union and European Economic Area countries (n=11,037, except for underlying conditions, n=5,205), 5 May -22 September 2009 F: female; M: male; N: no; Y: yes a It was not possible to calculate the proportion of completeness for underlying condition as the category 'none' did not exist for this variable.

Table 3
Univariate and multivariate analysis for factors influencing hospitalisation and complications of 2009 pandemic influenza A(H1N1) cases in 18 European Union and European Economic Area countries, 5 May -22 September 2009

Table 4
Percentage of underlying and co-morbid conditions reported in studies performed among patients hospitalised with 2009 pandemic influenza A(H1N1) a Patients hospitalised in critical care units.b Asthma and/or chronic obstructive pulmonary disease.c Only chronic heart failure.d Arrhythmia and valvular heart diseases.