Travel associated legionnaires ’ disease in Europe : 1997 and 1998

Introduction Le Réseau de Surveillance Européen de la maladie des légionnaires associée aux voyages a été mis en place en 1987 par le Groupe de travail européen sur les infections à Legionella (European Working Group on Legionella Infections, EWGLI). Son objectif est d’identifier les cas d’infections à légionelles survenant chez les personnes de retour de voyages, et de détecter les épidémies et les foyers de maladie des légionnaires. A l’origine, cette action était coordonnée à Stockholm par l’Institut Suédois pour le Contrôle des Maladies Infectieuses (Swedish Institute for Infectious Disease Control, SIIDC) et financée par l’Organisation Mondiale de la Santé (OMS). En 1993, le Centre de Surveillance des Maladies Transmissibles du Service de Santé Publique, à Londres (Communicable Disease Surveillance Centre, Public Health Laboratory Service (CDSC, PHLS)) a pris le relais. Ce changement a coïncidé avec le financement de ce réseau par la DirecTravel associated legionnaires’ disease in Europe: 1997 and 1998


Introduction
The European Surveillance Scheme for Travel Associated Legionnaires' Disease was set up by the European Working Group on Legionella Infections (EWGLI) in 1987 to identify cases of legionella infection in returning travellers and to detect outbreaks and clusters of legionnaires' disease.The scheme was initially run from the Swedish Institute for Infectious Disease Control (SIIDC) in Stockholm where it was funded by the World Health Organization (WHO).In 1993 it moved to its present location in the Public Health Laboratory Service (PHLS) Communicable Disease Surveillance Centre (CDSC) in London, a move which coincided with the beginning of funding by Directorate General V of the European Commission.

Methods
The surveillance scheme methods have been described previously (1).

Pays participant au programme de surveillance Ewgli en 1998.
Countries participating in the Ewgli surveillance scheme in 1998.
• Un antécédent de voyage dans les dix jours ayant précédé l'apparition de la maladie.Un voyage est défini comme un séjour en dehors de son domicile pendant au moins une nuit.Ne sont pas pris en compte les séjours d'une nuit chez des particuliers.
Tous les pays participants conservent une copie des données d'EWGLI.La ® Twenty-four countries took part in the scheme in 1997 and seven joined in 1998.There are now 36 collaborating centres in 31 countries (figure 1).
A case of travel associated legionnaires' disease is defined as follows: • Clinical or radiographic evidence of pneumonia accompanied by appropriate laboratory diagnosis.
• A history of travel in the ten days before the onset of illness.Travel is defined as staying away from home for one night or more.Overnight stays in private accommodation are not included.
Cases diagnosed by the detection of specific legionella antigen in urine using validated reagents have been regarded as confirmed since 1 January 1998.Details of cases ascertained by national or regional surveillance systems of participating countries are sent by fax or email to the coordinating centre at CDSC.The information collected includes the full travel itinerary and clinical and microbiological data.When a new case is added to the database at CDSC, the database is searched to see if previous cases have been reported at the same accommodation site.The collaborator, or ministry of health, in the presumed country of infection is immediately informed by fax of all cases associated with their country.All collaborators and WHO are informed immediately if the case is part of a cluster.Clusters are defined as two or more cases associated with the same accommodation site who became ill within six months of each other.If the case stayed at accommodation associated with previous cases, but became ill over six months later, then the cases are said to be linked.Collaborators and WHO are informed of linked groups at the end of each month.Some countries choose to inform representatives of their national organisations of tour operators about cases that arise in tourist accommodation.
All collaborating countries maintain a copy of the EWGLI data set.Most countries have a copy of the EWGLI database which is updated by email at the end of each month.

Results
The number of cases reported has increased from three cases in 1987 to 242 cases in 1997 and 232 in 1998 ( ® previous years.The average size of a cluster in both 1997 and 1998 was three cases.(figure 2) The most commonly used methods of diagnosis have changed during 1997/8.Forty-one per cent (99 in 1997, and 96 in 1998) of cases are now being diagnosed by detection of urinary antigen.The increase has been at the expense of serological methods; the proportion of cases confirmed by culture and by other methods remain similar (figure 3).
The age and sex profile of cases is similar to previous years (3).In 1997 and 1998 there were more than twice as many men as women (table 1).The age of cases was normally distributed around a mean of 57 in 1997 and 56 in 1998.
The outcome of illness was reported for most cases, but if the case was still ill when reported the information was often not updated and the outcome remained unknown.Fifty-six per cent of cases in 1997 are known to have recovered and 57% in 1998.Twenty-six deaths were reported in 1997 and 25 in 1998, a provisional case fatality rate of 11% in both years.Cases whose outcome was unknown at report accounted for 5% of reports in 1997 and 10% in 1998.
The seasonal pattern, based on dates of onset, has not changed significantly from previous years.Two peaks were seen in both 1997 and 1998, the first in June/July and the second in September/October.The timing of these peaks varies slightly from year to year (figure 4).A small peak seen around Easter time in 1998 had not been observed in previous years.
Most cases are reported from countries in northern Europe -in particular from England and Wales, Scotland, the Netherlands, Sweden, France, and Denmark -but 12 cases (almost 3%) were reported from Italy in 1997/8.Infections are usually diagnosed after return to the country of residence.Twelve countries reported cases in 1997 and 16 in 1998.

Travel
The 474 cases with onset in 1997/8 had made 697 visits to 51 countries.The Mediterranean region was the most popular destination.In both 1997 and 1998 more than 20% of cases had visited Spain.Italy, Greece, and Turkey combined accounted for 30% of cases.France and Germany had 15% of the visitors in 1997 and 14% in 1998 (figure 5).Cases who took their holidays in northern Europe stayed in a wider variety of places than those who went to Mediterranean coastal resorts.5).Les cas ayant passé leurs vacances en Europe du Nord ont séjourné dans un plus grand nombre d'endroits que ceux qui sont allés dans des stations balnéaires sur la côte méditerranéenne.
Le nombre de cas associé à un pays est généralement proportionnel au nombre total de personnes l'ayant visité.Cependant, en Espagne où le nombre de cas est important, le taux par million de voyageurs du Royaume-Uni (seul pays pour lequel l'information sur le nombre de voyageurs est disponible) n'y est pas plus important que dans les pays moins touristiques (Office National des Statistiques, données non publiées) (tableau 2).

Bateau de croisière 2 :
Trois cas de maladie des légionnaires, et un cas de légionellose non pulmonaire sont survenus en mai et juin 1998 à bord d'un bateau de croisière britannique.Le bateau était précédemment immatriculé en Italie et les nouveaux propriétaires n'étaient pas au courant que deux cas précédents de maladie des légionnaires, dont l'un fatal, y avait été détectés.Après inspection du bateau, de sérieux vices de forme ont été mis en évidence dans la régulation thermique du système d'eau chaude et froide, ainsi que dans le système électrique.La croisière ® The number of cases associated with a country is usually proportionate to the total number of people who visit the country.Therefore, although Spain has a high number of cases, the rate per million travellers from the United Kingdom (the only country for which information on the number of travellers is available) is no higher than in countries that receive fewer visitors (Office for National Statistics, unpublished data) (table 2).
Twenty-five clusters were detected in 1997 and 19 in 1998.Six of the clusters in 1997 and 10 in 1998 would not have been detected without the surveillance scheme since each included only one national from several countries.Most of the clusters detected occurred in the most visited countries, but there were some exceptions.For instance, the number of clusters on cruise ships was higher than would be expected given the numbers of people who take cruise holidays.

Outbreaks and clusters
Cruise ship 1: An outbreak of six cases (one fatal) in English and Scottish residents occurred on a Rhine cruise ship in 1997.The cases arose between July and October 1997 and had travelled on four separate cruises.The Dutch owned ship was taken out of operation when the outbreak was detected.The temperature of the hot water system was found to be inadequate and the whirlpool spa had been improperly maintained.Legionella pneumophila serogroup (sg) 4 was isolated from this pool but although there was strong epidemiological evidence that this was the source (4,8,9) evidence of L. pneumophila sg 4 infection was found in none of the patients.
Turkey: Sixteen cases and one suspected case of legionnaires' disease were identified in an outbreak at a hotel in Istanbul in September and October 1997.Four people died.Sixteen of the cases were French and one was Belgian.Isolates were obtained from six patients; typing showed that all were infected with the same strain of L. pneumophila sg 1 of a distinct, and previously unknown, type.There was no opportunity for environmental investigation of the hotel and the source of infection was never found.The epidemiology strongly suggested an extended point source.Two tour companies used the hotel and reported that the hotel was closed for renovation after the outbreak was detected (B Decludt, personal communication).
Cruise ship 2: Three cases of legionnaires' disease and one case of nonpneumonic legionellosis arose on a British ship in May and June 1998.The ship had previously been registered in Italy and the new owners were unaware that it had been associated with two previous cases of legionnaires' disease, one of which had been fatal.The ship was inspected and serious flaws were found in the temperature regulation of the hot and cold water system and in the electrical system.The ship's itinerary was disrupted while these faults were rectified.L. pneumophila sg 1 was isolated from the ship's water supply but no clinical isolates were available for comparison (5).
France: An increase in the number of cases reporting travel to Paris was observed in June 1998.None of the cases was associated with the same buildings but several were visiting France for the football world cup.Investigation by the French authorities and case searching through EWGLI resulted in the detection of nine travel associated cases: four English, three Scottish, one Swedish, and one Danish.Eleven cases were French residents.A case control ® EUROSURVEILLANCE VOL. 4
Spain: A outbreak of 11 cases occurred at a hotel in Benidorm between August and December 1998.Two cases had previously been associated with the hotel, one in 1990 and one in 1996.The first case in the cluster was reported to EWGLI in September 1998.The second and third cases were reported on 21 and 24 December, and a cluster alert was issued.The tour operators using the hotel withdrew their clients on 24 December.Inspections of the hotel and sampling of the water systems were carried out by the local health authorities and a private company.The water system had been chlorinated before samples were taken and no legionella were isolated.Over the next few weeks, as more cases were reported to CDSC, it transpired that eight cases had occurred by the time the cluster alert was issued but that they had not been reported (7).

Discussion
The surveillance scheme has continued to expand during 1997 and 1998.The gradual increase in the numbers of cases reported since the start of the scheme in 1987 is thought to be due to improved detection and reporting, rather than increased incidence.
The characteristics of the cases reported to EWGLI have remained consistent over the past few years, although the methods used to diagnose patients with legionnaires' disease changed during 1997/8 with more widespread use of urinary antigen detection techniques.The proportions of single, linked, and clustered cases have changed very little as have the seasonal distribution and the countries visited by cases.The importance of the scheme, in promptly detecting clusters and outbreaks, is illustrated by the fact that 24% of the clusters in 1997 and 53% in 1998 would not have been detected without EWGLI.Collaboration with tour operators is proving valuable in the prevention of further cases, by ensuring prompt action after clusters are reported at tourist accommodation sites.