Introduction
Every year, it is estimated that 77 million tourists visit France. Most
of them (90%) come from other European countries, mainly from the United
Kingdom (20%), Germany (19%) and the Netherlands (16%). Clusters of travel
associated legionnaires' disease are a matter of concern and many outbreaks
have been already published (1-3). In 1997, nationalsurveillance of legionnaires'
disease was heightened in France and in 2002, 1021 cases were notified
to the French national surveillance scheme with an incidence of 1.7 cases
per 100 000 population (4). In 1998, the French ministry of health published
national recommendations on the prevention of exposure to Legionella in
public places and accommodation (5,6).
In July 2002, the European Guidelines for control and prevention of travel
associated Legionnaires' disease were adopted at the voluntary level by
most European surveillance of travel associated legionnaires' disease
scheme (EWGLINET) participant countries (1). It should be noted that France
started following them as from July 2001, one year before their final
approval. This paper describes investigated clusters linked to travel
in France during a two year period from September 2001 to August 2003
are described.
Methods
EWGLINET and the definitions and procedures for responding to cases
of travel associated legionnaires' diseases are described elsewhere
(7,8). The French national institute of public health surveillance (Institut
de Veille Sanitaire) notifies the coordinating centre in London of all
the French cases of legionellosis in patients who had been travelling
during the incubation period in France or in other countries, and receives
notifications of foreign cases travelling in France.
When a cluster is detected by the French national surveillance system
or through EWGLINET, the local health authorities are immediately informed
and an environmental investigation is conducted. A preliminary report
stating whether control measures are in progress and if the hotel remains
open or not should be sent to the co-ordinating centre within two weeks
(Form A) from the cluster alert, while a full report (Form B) should
be sent within six weeks from the cluster alert. Local health authorities
are responsible for filling these forms, which are available in French,
to notify surveillance networks of the conclusion of their investigation.
Results
Between September 2001 and August 2003, EWGLINET and the French national
surveillance system identified 37 clusters located at various French
tourist accommodation sites. Thirteen clusters occurred between September
2001 and August 2002 and 24 between September 2002 and August 2003.
These clusters occurred in 27 hotels and 10 campsites.
The number of clinical cases per cluster was as follows: 30 clusters
of 2 cases (81%), 6 clusters of 3 cases (16%) and 4 consisting of a
single case (3%) giving a total of 82 cases. The mean age of the cases
was 60 years (range 27-96 years). Most of the cases were male (80%)
and the sex ratio male/female was 4.
According to the European case definition (4), 75 (91%) cases were confirmed
and 7 (9%) were probable. Diagnosis was by detection of urinary antigen
for 67 cases (82%), by culture for 6 (7.3%), and by a four fold rise
in specific serum antibody titre for 2 (2.4%). Five patients had a single
high titre and 2 were diagnosed by PCR.
The mean length of patients' stay at the accommodation sites was 5 days
(range 1-27 days). Most of the accommodation sites were located in southern
France (Figure 1). Twenty four of the patients who stayed at cluster
sites also stayed at other sites in France, whilst 4 also stayed at
sites in other European countries.
French citizens were involved in 9 (24.3%) clusters together with other
European citizens whereas in 16 (43.2%) clusters, patients were exclusively
French. In 12 (32.5%) other clusters, only other European citizens were
affected.
The mean time interval between the first and second case was 94 days
(range 0-626 days) and in 13 (35%) clusters, the interval was less than
one month. In 13 (35%) clusters, the second case occurred more than
6 months after the first case notified was notified.
For 36 of the 37 clusters, the local health authorities performed environmental
investigations. One campsite was closed during the winter season at
the date of notification and no investigation was conducted.
The investigations were carried out between 0 and 10 days (mean 5 days)
after the EWGLINET notification, but for 11 clusters, investigation
took place prior to the EWGLINET notification. For these clusters, as
all patients were French, the French surveillance network was warned
before EWGLINET was.
Among the 36 clusters investigated, water samples were collected in
35. As one campsite was closed when an investigation was requested,
only a risk assessment was carried out.
In 16 (46%) sites, Legionella pneumophila was found at a level more
than 103 cfu/litre and in 6 (17%) Legionella pneumophila was present
at a level between 102 and 103 cfu/litre at the time of investigation.
In 13 (43%) sites, no Legionella pneumophila was found. In 26 (72%)
sites, the assessment identified the low temperature of the hot water
system and closed off water pipes among the risks present.
In all accommodation sites with inadequate risk assessments, control
measures were implemented immediately, and 6 hotels were closed immediately
after the cluster alert.
Form B was sent to the EWGLINET coordinating centre punctually in 35
out of 36 cluster investigations. The name of one campsite was published
on EWGLINET website but then removed when satisfactory measures were
taken by the owner.
Comparison of clinical and environmental isolates by pulsed field gel
electrophoresis (PFGE) at the Centre National de Référence
(CNR) des Legionella (national reference centre for legionellae) was
possible for 3 clusters and identical genomic profile of the isolates
were found in all.
Four accommodation sites had previously been linked with clusters in
2001, 2002 or 2003. At that time, all the control measures have been
taken and controlled by the local health authorities and the form B
has been returned with satisfactory conclusion.
Discussion
Through the network, we detected clusters with small numbers of cases
but we could assume that control measures have prevented a number of
new cases. Good collaboration has meant that numbers of clusters detected
have nearly doubled in the two year period. It is not surprising that
most of the clusters were located in the south of France, a popular
destination for holidays.
The high number of French citizens involved in the clusters can be explained
by the fact that there are more French people than foreigners who travel
in France. In fact, data on tourist origins in France shows that 63%
are French and 37% foreigners (9).
The improvement of our surveillance system in the recent years has also
allowed a rapid detection of clusters.
The previous case definition of a cluster was 2 cases during a six month
period. Using this definition, we would have missed 35% of the clusters.
The risk assessments showed that most of the sites were at risk for
Legionella contamination and infection. In nearly half of the sites,
contamination with Legionella was more than 103 cfu/l which is the level
where action is required to be taken (7). Despite the low proportion
of human cultures obtained, in all the clusters where comparison of
clinical and environmental isolates was possible, we had confirmation
of the source of infection. However, our data shows that 29% of tourists
stay in two or more hotels during the incubation period, highlighting
the problem of interpreting association between cases and possible multiple
sources of infection.
It is worrying that 4 sites were previously linked with clusters and
the subsequently had an extra case. It may be important to implement
regular tests at these sites known to be particularly at risk during
a determined period.
Appropriate surveillance and timely notification is necessary for interruption
of Legionella transmission from ongoing outbreak sources and for implementation
of preventive measures. The European EWGLINET is a unique, sensitive
network (6). It has been very efficient in determining numbers of published
European outbreaks (1-3).
This reinforces the importance of the European surveillance network
and the timely notifications of all the cases to EWGLINET, particularly
national cases travelling inside their own country as these could potentially
be linked to other European cases.
|