Introduction

Influenza is an important public health problem in the industrialised world. It is associated with higher general practice consultation rates (1), increased hospital admissions and excess deaths (2, 3). In England, the average number of excess deaths during influenza epidemic periods (1989-98) was estimated to be 12 554 each year (range 0-27,587) (3). In the United States (1972-92), this figure was estimated to be 5 700 (range 0-11,800) for influenza and pneumonia and 21 300 for all-causes of death (range 0-47,200) (2). The impact of Influenza must also be considered in terms of health care planning, increased days lost due to absence from work, and influenza pandemic planning (4).

The European Influenza Surveillance Scheme (EISS) is a collaborative project aimed at monitoring influenza activity in Europe (5-8). The project has received funding from the European Commission since November 1999 and has the following objectives: 1) to facilitate the rapid exchange of information on influenza activity in Europe 2) to combine clinical and virological data in the same population 3) to provide standardised information of high quality and 4) to identify which viruses circulate in the community, thus enabling comparison with the current vaccine composition.

All countries in Europe are welcome to join EISS. Full members meet the following criteria: 1) the surveillance network (consisting of sentinel physicians providing clinical information and national reference laboratories providing virological data) is nationally or regionally representative 2) the authority of the network is recognised by the national or regional health authority in the country or region; 3) clinical surveillance and virological surveillance are integrated in the same population (community) 4) the network has functioned successfully for two years and 5) the network can deliver data on a weekly basis.

The EISS project began in 1996 with seven member countries (Belgium, France, Germany, the Netherlands, Portugal, Spain and the United Kingdom), and this number has increased to 18 countries covering 20 surveillance networks during the 2001-02 influenza season. During the 2001-02 influenza season, six members (Ireland, Norway, Poland, Romania, the Slovak Republic and Sweden) were ‘associate members’, as they did not completely fit the membership criteria (Ireland because its surveillance network hadn’t been in operation for two years, and the others because they did not yet combine clinical and virological data in the same population). With all members, the EISS project now includes 25 national influenza reference laboratories, at least 10 500 sentinel physicians and covers a total population in Europe of 438 million inhabitants.

EISS uses two indicators to assess influenza activity in a country: the geographical spread (a WHO indicator) and the intensity of clinical activity. The geographical spread has the following levels:

• No activity: no evidence of influenza virus activity (clinical activity remains at baseline levels);

• Sporadic: isolated cases of laboratory confirmed influenza infection;

• Local outbreak: increased influenza activity in local areas (e.g. a city) within a region, or outbreaks in two or more institutions (e.g. schools) within a region. Laboratory confirmed;

• Regional activity: influenza activity above baseline levels in one or more regions with a population comprising less than 50% of the country’s total population. Laboratory confirmed;

• Widespread: influenza activity above baseline levels in one or more regions with a population comprising 50% or more of the country’s population. Laboratory confirmed.

The intensity of clinical activity compares the weekly clinical morbidity rate with historical data: Low = no influenza activity or influenza activity at baseline level; Medium = usual levels of influenza activity; High = higher than usual levels of influenza activity; Very high = particularly severe levels of influenza activity (less than once every 10 years).

Methods

From week 40/2001 (02–08/10/2001) to week 15/2002 (09– 15/4/2002), 20 networks in 18 countries actively participated in EISS (table 1). In each of these countries, one or several networks of sentinel physicians collected clinical activity indicators (the number of cases of influenza-like illness (ILI) or cases of acute respiratory infection (ARI)) and obtained nasal, pharyngeal, or nasopharyngeal specimens that were sent to the national reference laboratory(ies) for virological monitoring. Combining clinical and virological data in the same population allows the validation of clinical reports made by the sentinel physicians, and provides virological data in a clearly defined population (the general population that visits their general practitioner or paediatrician with an influenza-like illness or an acute respiratory infection) (9).

Table 1: Summary characteristics of the sentinel surveillance systems in EISS

1. De nombreux réseaux/pays étaient membres de projets ayant précédé EISS (créé en 1996)– Eurosentinel (1987-91) et Système d’alerte précoce ENS-CARE Influenza (1992-95) / Many of the networks/countries were members of pre-EISS surveillance projects in Europe – the Eurosentinel (1987-91) and ENS-CARE Influenza Early Warning System (1992-95) projects.
2. Nombre de praticiens au cours de la saison 2000-01/ Number of physicians during the 2000-01 influenza season
3. Praticiens travaillant au sein d'écoles (enfants) et en services de soins pour jeunes / Physicians working in community schools (children) and youth health services
4. Praticiens spécialisés en médecine interne/ Physicians specialised in internal medicine
5. ARI: infection respiratoire aiguë/ acute respiratory infection; ILI: syndrome grippal/ influenza-like illness
* Un ou plusieurs médecin par cabinet / One or more GP(s) per practice; NK: inconnu / not known.

The virological data collected results from rapid diagnostic tests (immuno-enzymological or immunofluorescence) and from cell cultures with specific identification. Some laboratories also use reverse transcription polymerase chain reaction (RT-PCR) routinely as a rapid test. In addition to specimens obtained from practitioners in the sentinel surveillance systems, the laboratories also collect and report to EISS test results on specimens obtained from other sources (e.g. from hospitals or non-sentinel physicians). These data are collected as a range of indices used to monitor influenza activity, and allow the validation of virological data obtained from the sentinel sources.

During the influenza season, weekly data on influenza activity is centralised nationally. The analysis of the epidemiological situation is mainly based on clinical morbidity rates (ILI or ARI/100 000 population) and on virological results (number of isolates/influenza virus detections and percentage of influenza positive sentinel specimens). After processing and analysis by national centres, the data gathered are entered into the EISS database (by the Thursday of the following week) that is accessible by internet (the EISS information technology organisation has been described previously (10)). This allows the publication of a report on influenza activity in Europe (the Weekly Electronic Bulletin) each Friday on the EISS website (www.eiss.org) during the influenza season (week 41/2001 to week 15/2002 during the 2001-02 season).

Results

From October to December (week 40/2001 to 52/2001), influenza viruses were circulating sporadically in some countries within EISS (Belgium, the Czech Republic, France, Italy, the Netherlands, Slovenia and Romania). In January 2002 (week 01/2002 to 05/2002), influenza activity increased in Europe (figure 1). During this time, widespread activity was observed in Belgium, France, Portugal and Spain. At the beginning of February (week 06/2002 and 07/2002), the clinical morbidity rates increased in Germany, Italy, the Netherlands, Norway and Sweden. At the end of February (week 08/2002 and 09/2002), the geographical spread of influenza was reported as widespread in Germany, Italy, the Netherlands, Norway, Romania and Switzerland, while in the other countries influenza activity declined smoothly. From the end of March to mid-April (week 12/2002 to 15/2002), widespread activity continued to be reported in Norway and Germany, while the activity in the other countries was stable or decreasing. In week 16/2002, influenza activity was back to pre-epidemic levels in all EISS networks.

Table 2 presents an overview of the influenza activity in the EISS networks during the 2001-02 winter season. The peak levels of clinical morbidity varied considerably from one network to another, with networks reporting ILI rates ranging from 9 to 1 201 cases per 100 000 population. Clinical morbidity rates of ARI were generally higher and also varied considerably, from 375 cases per 100 000 population in Romania to 3 320 per 100 000 population in France. Peak clinical morbidity rates were generally reached around week 04/2002 to 06/2002, with some networks in the north and east of Europe reaching these levels later in the winter (around week 11/2002). The geographical spread of influenza ranged from sporadic (in six networks) to widespread (in eleven networks), whilst the intensity of the activity ranged from low (eight networks) to high (in Spain). Eleven networks reported that the dominant subtype was influenza A(H3N2), three networks reported that it was influenza B and one network (Germany) reported that it was influenza A(H3N2) and influenza B. In some networks (e.g. France) the dominant subtype was A(H3N2), but influenza B viruses co-circulated during all of the winter season.

Table 2. Overview of influenza activity in the EISS networks during the 2001-02 season

^* Le pic du taux de morbidité peut être sous-évalué car les médecins sentinelle de la République Slovaque ont commencé à déclarer les cas de syndromes grippaux au lieu des IRA au cours de la saison 2001-02, mais certains ont pu continuer à déclarer les cas de IRA. / The peak in morbidity rate may be biased upwards as sentinel physicians in the Slovak Republic began reporting cases of ILI instead of ARI during the 2001-02 season; some of the sentinel physicians may have continued reporting cases of ARI.
n.a: non applicable en l’absence de données sentinelles / not applicable as no sentinel data is available.

During the 2001-02 influenza season, a total of 14 151 respiratory specimens were tested for influenza from the sentinel surveillance systems (see figure 2), of which 3689 (26%) were positive for influenza. The median number of specimens collected per network was 402 (range 0 to 3657). Of the positive specimens, 2512 (68%) were typed as influenza A and 1177 (32%) as influenza B. Of the 1762 subtyped influenza A viruses, 1600 (91%) were of the H3N2 subtype, 77 (4%) of the H1N1 and 85 (5%) of the H1N2 subtype.

At least 25 071 respiratory specimens were tested for influenza from non-sentinel sources during the 2001-02 influenza season (the total number is actually higher as some countries did not know the total number of respiratory samples tested and only reported positive test results). A total of 4 806 (19%) non-sentinel respiratory specimens were positive for influenza (figure 2). The median number of non-sentinel specimens tested for each country was 84 (range 0 to 19 430). Of the positive specimens, 4 162 (87%) were typed as influenza A and 644 (13%) as influenza B. Of the 576 subtyped influenza A viruses, 535 (93%) were of the H3N2 subtype, 31 (5%) of the H1N1 subtype and 10 (2%) of the H1N2 subtype. The total number of non-sentinel influenza A(H1N2) detections is underestimated, as an additional 217 isolates from non-sentinel sources in England were subsequently characterised by the PHLS virus reference laboratory, but these data were not yet entered into the EISS database at the time of analysis.

During the winter of 2001–02, influenza A (H1N2) viruses were reported from 12 networks participating in EISS (Belgium, England, France, Germany, Ireland, the Netherlands, Portugal, Romania, Scotland, Sweden, Switzerland and Wales). Belgium, Germany, the Netherlands, Scotland, Sweden, Switzerland and Wales reported less than five isolates each. Ireland reported 12 isolates, Romania eight isolates from the Constanta region, and France reported seven isolates derived from all regions of the country during the course of the season. In England, however, 56 influenza A(H1N2) viruses were reported to EISS from all regions, and for the duration of the season, suggesting a wider circulation of the new subtype (table 3). Outbreaks occurred that lasted between one and five weeks, with high attack rates generally among school-aged children (range 3–43%) (PHLS, N. Goddard, personal communication). Unpublished data from a retrospective analysis undertaken in England suggests that the new subtype was not circulating widely prior to September 2001.

Table 3. Influenza A(H1N2) viruses reported during the 2001-02 season

NK: inconnu / not known.

During the 2001–02 influenza season (weeks 40/2001 to 15/2002), five networks (France, Germany, Italy, the Netherlands and Norway) reported 118 detections of influenza B/Victoria/2/87-lineage viruses to EISS (table 4). Germany reported 86% of these detections, with 83 of the 102 reports coming from non-sentinel sources. The first B/Victoria/2/87-like viruses in Germany were isolated in week 10 and continued to circulate until the end of the season. Viruses from this lineage were detected all over the country and were mainly isolated during the month of March, when influenza activity in Germany reached its peak. France reported 8 isolations (4 sentinel and 4 non-sentinel) of influenza B/Victoria/2/87-lineage viruses from week 11–15. The Netherlands reported the detection of three viruses, all from non-sentinel sources during week 01/2002, and Italy reported one virus derived from a sentinel specimen during week 51/2001. Norway reported four viruses between weeks 07/2002 and 13/2002, with an additional three after the EISS reporting period had ended (after 15/2002), all of which were from non-sentinel sources. The Victoria-like viruses seen in Norway were similar to the 1999-2000 Asian vaccine strain B/Shandong/7/97, which is a B/HongKong/330/01-like virus (11).

Table 4. Influenza B/Victoria/2/87-like virus strains reported during the 2001-02 season

¹ En Allemagne, 19 déclarations de virus B/ Victoria/2/87 (4 sentinelles, 15 non-sentinelles) ont été faites après la semaine 15. / In Germany 19 more reports of B/ Victoria/2/87-like viruses (4 sentinel, 15 non-sentinel) were made after week 15.
² En Norvège, 3 déclarations de virus B/Victoria ont été faites après la semaine 15. Les cas étaient âgés de 12 mois, 5 ans et 13 ans. Le cas âgé de 12 mois est décédé après avoir présenté des symptômes d’infections des voies respiratoires supérieures pendant une semaine. / In Norway 3 more reports of B/Victoria-like viruses were made after week 15. The ages of the cases were 12 months, 5 yr and 13 yr. The 12 month old child died unexpectedly after one week of symptoms of upper respiratory tract infection.

Figure 3 shows the age distribution of the influenza A(H1N2) isolates and influenza B/Victoria/2/87-lineage viruses reported to EISS. While the influenza A(H1N2) and the influenza B/Victoria/2/87-lineage viruses were found in people of all ages, the younger age groups were predominantly affected. Influenza A(H1N2) viruses were mainly isolated in adolescents aged 10–19, and influenza B/Victoria-like viruses were found mainly in younger children aged 5–14 years.

Discussion

In most EISS countries, the influenza A virus (H3N2) was the dominant influenza virus circulating in the population during the 2001-02 season. In a number of coun- tries this virus co-circulated with influenza B (e.g. in France and Germany). The influenza A(H1N1) subtype was detected sporadically. In most of the EISS participating countries, the morbidity rates recorded during the 2001–02 winter was low to medium, illustrating mild to moderate influenza activity.

Table 2 highlights the difficulty of obtaining easily comparable clinical morbidity data for Europe. Morbidity rates peaked very differently, even between close neighbours like Belgium (951 cases of ILI per 100 000 population) and the Netherlands (136 cases of ILI per 100 000 population). These differences can probably be explained by a multitude of factors, including: different case definitions (12) differences in the health care systems (e.g. the density of general practitioners), some networks having sentinel paediatricians; the methods used to estimate the population denominators (13) and different consultation rates for influenza associated with the need to have a medical certificate for sick leave (13). In Belgium, a medical certificate is required from a physician for a single day of absence due to illness (probably leading to a higher consultation rate), whilst in the Netherlands there is no determined time limit for which a medical certificate is required to be absent from work (probably leading to a lower consultation rate).

The virological results (figure 2) reveal that there are large variations in the total number of sentinel samples collected by the networks each season. In terms of sentinel specimens collected per 100,000 population, Slovenia, Belgium, Switzerland and the Czech Republic collected the most sentinel specimens during the 2001-02 season (data not shown). The EISS group does not have any recommendation on the number of sentinel specimens to be collected each week. An inventory carried out in October 2000 found a wide variety of sentinel specimen collection protocols ranging from the absence of protocol to recommendations on the total number of specimens each sentinel physician should collect each week (12).

The virological results (figure 2) also reveal that there were significantly more cases of influenza B in sentinel specimens than in non-sentinel specimens (32% versus 13%, p < 0.001), and that sentinel respiratory specimens were significantly more frequently subtyped than non-sentinel respiratory specimens (70% versus 14%, p < 0.001). The former could be related to the generally less severe clinical outcome of infections of influenza B viruses in comparison with those of influenza A: more often, patients with influenza B would be seen by their physician rather than in a hospital (resulting in an overrepresentation of influenza A viruses from non-sentinel sources) (14). The finding that sentinel specimens were significantly more frequently subtyped is probably related to the fact that the national reference laboratories are responsible for providing the sentinel practitioners with feedback on the test results. It may also be an indication that the national reference laboratories have decided to pay more attention to these specimens, as they come from a clearly defined population (community-based data) and are a better indicator of virological activity in the general population.

The WHO announced the isolation of a new strain of the influenza A virus – A(H1N2) in early February 2002–based on information from the WHO global influenza surveillance programme, and the PHLS surveillance of ‰ ‰ influenza in England and Wales. The new strain appears to have arisen by reassortment of the two currently circulating human viruses (H1N1 and H3N2) to produce the H1N2 virus (15). Because the new strain is a combination of the two influenza A components contained in the current season’s vaccine (H1N1 and H3N2), people who received the vaccine should have developed a good level of immunity to the new strain. This is reflected in the age distribution of cases with younger age groups predominantly affected by the new subtype (see Figure 2). The younger age groups in the population are more susceptible to primary infection and also less likely to have received the vaccine. The latter point is illustrated by the fact that where reported, none of the cases reported to EISS had been vaccinated.

In all countries, except England and Ireland, reports of influenza A(H1N2) made to EISS represented sporadic isolates. Overall, the data reported to the EISS surveillance scheme suggest that the new subtype (H1N2) was not circulating widely across Europe during the 2001-02 season. It remains to be seen whether this will continue to be the case during the forthcoming 2002-03 influenza season. The reassortment that led to the new virus strain may have been a sporadic event, which will not persist, or alternatively this virus may replace the influenza A (H1N1) strain that is circulating across Europe.

There were generally sporadic isolations of influenza B/Victoria/2/87-lineage viruses in Europe during the 2001-02 influenza season, and these mainly occured at the end of the season. Germany reported a large number of isolations (102) from week 10/2002 onwards, mainly from non-sentinel sources. These reports should be interpreted with care as many of them may have come from investigations of influenza outbreaks (which would lead to many isolations in non-sentinel specimens). Overall, these viruses represented 20% (10% in sentinel specimens and 23% in non-sentinel specimens) of influenza B viruses circulating in Germany during the 2001-02 season.

Identification of circulating viruses within the population, and the recognition of virological changes are important EISS’ objectives. The emergence of two novel viruses during the 2001-02 season exemplified the benefit of having a system to facilitate the rapid exchange of information across Europe. Whilst the impact of the new H1N2 subtype and the B/Victoria/2/87-lineage viruses was not substantial in terms of either morbidity or mortality, it highlights the need for continual surveillance of circulating viruses to detect the emergence or re-emergence of viruses with pandemic potential. In addition, detection of new strains has implications on the diagnostic methodology currently undertaken by countries participating in EISS. Until now, the subtyping of influenza A viruses has been based on the characterisation of the haemagglutinin (HA) component of the influenza virus. Detection of the influenza A(H1N2) subtype means that it will now be necessary for laboratories to additionally undertake characterisation of the neuraminidase (NA) component of all influenza A viruses isolated.

Prior to the 2001-2002 season, the influenza B viruses circulating globally were antigenically closely related to B/Sichuan/379/99, with influenza B viruses of the B/Victoria-lineage remaining confined to regions in East Asia (China, Japan) since the early 1990s. During the 2001- 02 season, however, influenza viruses of the B/Victoria/2/87-lineage – similar to B/HongKong/330/2001-like virus – spread in an increasing number of countries in Europe and North America (16). It is anticipated this strain will become more widespread and may even replace the B/Sichuan/379/99-like viruses in due course. In light of this fact, the vaccine composition for the coming season, 2002-03, incorporates a B/HongKong/330/2001-like virus. The composition of the influenza vaccine for the 2002-03 season (Northern hemisphere winter) was announced by the World Health Organisation in Geneva. The vaccine will contain:

• an A/New Caledonia/20/99 (H1N1)-like virus

• an A/Moscow/10/99 (H3N2)-like virus

(the widely used vaccine strain is A/Panama/2007/99)

• a B/HongKong/330/2001-like virus.

This article was written on behalf of all EISS members: Alexandrescu V (RO), Aymard M (FR), Bartelds AIM (NL), Buchholz U (DE), Brydak L (PO), Christie P (UK), Cohen JM (FR), Domegan L (IRL), Falcao I (PT), Fleming DM (UK), Grauballe P (DK), Haas, W (DE), Havlickova M (CZ), Heckler R (DE), Heijnen M-L (NL), Hungnes O (N), Iversen B (N), de Jong JC (NL), Kaiser L (CH), Kramer M (DE), Kristufkova Z (SK), Lina B (FR), Linde A (SW), Lupulescu E (RO), Machala M (PO), Manuguerra J.-C. (FR), de Mateo S (ES), Meerhoff T (NL), Mosnier A (FR), Müller D (CH), Nolan D (IRL), O’Flanagan D (IRL), Paget WJ (NL), Perez-Brena P (ES), Penttinen P (SW), Pierquin F (B), Pregliasco F (IT), Prosenc K (SL), Rebelo de Andrade H (PT), Samuelsson S (DK), Schweiger B (DE), Socan M (SL), Thomas D (UK), Thomas Y (CH), Tumova B (CZ), Uphoff H (DE), Valette M (FR), Vega T (ES), van der Velden K (NL), van der Werf S (FR), Watson J (UK), Yane F (B) et Zambon M (UK).

Acknowledgements

EISS would not exist without the regular participation of sentinel physicians across Europe. We would like to thank them for making this surveillance scheme possible.

Participants :

Allemagne / Germany : ArbeitsGemeinschaft Influenza, Marburg; Robert Koch Institute, Berlin; Niedersächsisches Landesgesundheitsamt, Hannover
Belgique / Belgium : Scientific Institute of Public Health - Louis Pasteur, Bruxelles
Danemark / Denmark : Statens Serum Institut, Copenhagen
Espagne / Spain : Instituto de Salud Carlos III, Madrid; Sentinel Networks of Madrid, Castilla y Leon, Valencia, Pais Vasco, Guadalajara and Andalucia
France : Groupes Régionaux d’Observation de la Grippe, Open Rome, Paris; Institut Pasteur, Paris; Centre Hospitalo-Universitaire, Lyon
Grande-Bretagne / Great Britain : Royal College of General Practitioneers, Birmingham; PHLS Communicable Disease Surveillance Centre, London; PHLS Central Public Health Laboratory, London; Scottish Centre for Infection and Environnemental Health, Glasgow; Communicable Disease Surveillance Centre, Cardiff
Irlande / Ireland : Irish College of General Practitioners, Dublin; National Disease Surveillance Centre, Dublin
Italie / Italy : Istituto di Virologia, Milano; Dipartimento di Scienze della Salute, Genova; Istituto Superiore di Sanita, Roma
Norvege / Norway : National Institute of Public Health, Oslo
Pays-Bas / Netherlands : Netherlands Institute for Health Services Research, Utrecht; National Institute for Public Health and the Environment, Bilthoven; Erasmus University, Rotterdam
Pologne / Poland : National Institute of Hygiene, Warsaw
Portugal : Instituto Nacional de Saude, Lisboa
République Tchèque / Czech Republic : National Institute of Public Health, Praha; National Influenza Center, Praha
République Slovaque / Slovak Republic : State Health Institute of the Slovak Republic, Bratislava
Roumanie / Romania : Cantacuzino Institute, Bucharest
Slovénie / Slovenia : Institute of Public Health, Ljubljana
Suède / Sweden : Swedish Institute for Infectious Disease Control, Solna
Suisse / Switzerland : Swiss Federal Office of Public Health, Bern; National Centre for Influenza, Hôpital Cantonal Universitaire, Geneva