| Introduction
In June 2003, the Dutch National Salmonella Centre reported a significant
excess Salmonella isolation rate compared to previous years in most
regional public health laboratories (FIGURE 1). Beginning in May 2003,
the number of laboratory confirmed cases clearly increased to above
the level expected [1], and from June to November, and again since the
beginning of 2004, to above the level of tolerance (a measure for the
significance of an excess). This increase involved only Salmonella
Enteritidis, and not S. Typhimurium (ST), or other Salmonella
serotypes or Campylobacter spp. In this paper, we try to indicate the
possible role in the 2003 excess of the hot summer compared with that
of the increase of imports of (contaminated) eggs due to the concurrent
avian influenza outbreak.
Salmonella surveillance
The data are from the National Salmonella Centre (NSC) and the National
and European Reference Laboratory (CRL) for Salmonella at RIVM that
performs the sero- and phage-typing of isolates taken from humans (mostly
sent by regional public health laboratories,covering 64% of the Dutch
population) and animals, from food, animal feed and from the environment
[2]. The sensitivity to various antibiotics has been quantitatively
determined by the minimal inhibitory concentration (MIC) at the Centraal
Instituut voor DierziekteControle - Lelystad (CIDC- Central Institute
of Animal Disease Control) [3].
The excess isolation rate of S. Enteritidis since May 2003 amounted
to an extra 540 laboratory confirmed cases for the whole of the Netherlands
at the end of 2003 (FIGURE 2, adjusted for the 64% coverage of the laboratory
surveillance). This is 50% higher than excesses found in previous years.
Figure 2 shows that the large increase of cases involved S. Enteritidis
only. Extrapolation using data from a 1999 study [4], then 540 extra
laboratory confirmed cases would mean an estimated 7500 extra cases
of gastroenteritis caused by S. Enteritidis in the total population.
Denmark has a laboratory surveillance system comparable to that of the
Netherlands, and a Danish study has shown that, when compared with controls,
1.5-2.1% of the laboratory confirmed patients with salmonellosis die
within one year, probably due to the infection [5]. This would mean
that the 2003 excess S. Enteritidis infections in the Netherlands caused
8-11 deaths.
Hot summer
The excess of SE cases in June and July was at first attributed
to the exceptionally hot weather that lasted until August, when temperatures
were far higher than normal for that time of year [FIGURE 2]. This was
suggested by the findings in the WHO cCASHh (project (climate Change
and Adaptation Strategies for Human health in Europe: http://www.who.dk/ccashh)
of time series analysis of salmonellosis in 10 European countries. An
additional effect of temperature was demonstrated clearly on the risk
for food poisoning, apart from a general effect of season itself [6].
In the Dutch data (covering the period 1984-2001) this effect was exceptionally
strong for S. Enteritidis (a linear 12.6% increase per oC). The
largest effect of temperature is one week before onset of illness, with
diminishing but positive effects up to 5 weeks [6]. Earlier calculations
of our own, that more strongly adjust for season (covering 1990-1998),
illustrate these findings [FIGURE 3].
Above a threshold of about 6oC the risk linearly increases;
most strongly for S. Enteritidis but, in the Netherlands, hardly
so for ST. The difference between S. Enteritidis and S.
Typhimurium probably derives from the traditional food preparation of
the main food vehicle for S. Enteritidis, eggs, sometimes processed and
consumed raw, whereas S. Typhimurium is mainly associated with
meat from pigs and cattle that normally get a proper heat treatment.
However, an inquiry among the members of the Enter-net surveillance network
revealed that most European countries had not experienced an excess of
Salmonella infections during the same time period, with the exception
of Belgium, and England and Wales. Therefore the hot summer was unlikely
to have had a major role in the excess.. Furthermore, figure 2 shows that
the 'hot summer' occurred during the months of June, July and -August,
when temperatures were on average between 1 and 2.5 oC above normal. This
period was followed by two months when temperatures that were below normal.
Clearly, a 7-13% increase per oC cannot explain the 50% excess of cases
at the end of the year. Note that due to the lag of about one week between
temperature changes and the onset of disease and another three weeks until
the laboratory results appear, the temperature findings in figure 2 should
be compared with the surveillance findings of one month later.
Raw shell eggs
Surveillance programmes in the Netherlands show that the Salmonella
control programme for poultry has been successful in reducing S.
Enteritidis in broilers almost to exclusion [7]. However, in commercial
layers in 2003, more than 6% (9% in 2001 and 14% in 1997) of the flocks
remained S. Enteritidis positive (7). This makes raw shell eggs
the main suspect food vehicle for causing the 2003 excess of S.
Enteritidis infections in humans. However, phage typing of S. Enteritidis,
combined with antimicrobial resistance testing, showed remarkable differences
between human and poultry isolates, pointing to a source from outside
the Netherlands [3]. In 2003, twice as much phage type 1 (PT 1) was
found among S. Enteritidis isolates from Dutch patients (14.5%)
as between 1998-2002, 54% of them being resistant to nalidixic acid
(Na) and with decreased susceptibility to ciprofloxacin. Between 1998-2003,
PT 1 accounted for about 5% of all S. Enteritidis poultry isolates
(SE isolates derive almost exclusively from layer flocks , but none
of these were resistant to nalidixic acid. Human infections with PT
1(Na) in the Netherlands appeared to be travel-related three times more
often than other S. Enteritidis phage types, and more than 50%
of PT1(Na) infections were related to travel to Spain and Portugal .
A series of outbreaks with S. Enteritidis in the United Kingdom
(UK) in 2002 and again in 2003 [8] led to several investigations of
raw shell eggs [9,10]. Among a range of other phage types, PT 1(Na)
was found to be associated with Spanish eggs. Salmonella was found in
0.3% of the eggs produced in the UK and in 5.1% and 6.7% in two surveys
of eggs imported from Spain and was high as well (7.7%) in other imports
where the country of origin was unknown. Salmonella was found in only
0-0.03% of eggs produced in Holland [11], i.e. 10 and 160 times lower
than eggs produced in the UK and Spain respectively. It is nevertheless
estimated that about 35% human salmonellosis cases in the Netherlands
are due to consumption of eggs.
Avian Influenza outbreak in poultry
The Netherlands experienced a major outbreak of avian influenza in
poultry in the spring of 2003 that led to a shortage of eggs on the Dutch
market. Data from EUROSTAT [FIGURE 4] shows that this shortage was compensated
for with egg imports, mainly from Germany, Italy and Spain (>8-fold
increase in the 2nd quarter of 2003 as compared to former quarters) .
In the fourth quarter of 2003, the number of imported eggs was still considerably
higher than in former years. In fact the contribution of eggs imported
from the new EU member states , negligible in previous years, continued
to increase and doubled in the second half of 2003. Figure 1 shows that
in the first months of 2004 there was still an excess of S. Enteritidis
cases, now predominantly PT 8. PT 8 has been reported as a problem in
the poultry industry in several new member states in central Europe (personal
communication with NRL and ENTERNET colleagues) . For several years, central
European countries have been the number one destination of travellers
that returned with a PT 8 infection.
Discussion
In June 2003, the Dutch National Salmonella Centre reported a significant
excess isolation rate of S. Enteritidis when compared with previous
years . The hot summer of 2003 could not explain the findings. Strong
evidence was found to suggest that the increase in importation of contaminated
eggs, as a result of the avian influenza outbreak, was the most probable
reason for this excess.
The lesson is that with the low level of contamination in Dutch eggs,
even small increases in imports of eggs that are relatively highly contaminated
with S. Enteritidis, may have a large impact on the incidence of human
salmonellosis, and may strongly affect both morbidity and mortality. Hence,
major changes in market supply should initially be considered as a potential
serious public health threat. Continuous surveillance, especially of imported
eggs, is therefore strongly recommended. The approaching implementation
of a harmonized system for monitoring and control of Salmonella spp. in
flocks of laying hens in all EU Member States (EC Zoonosis Regulation
2160/2003 is an important, and constructive development in this respect.
Trace back of the source of salmonellosis cases, serotyping and phage
typing of positive findings, together with testing for antimicrobial resistance,
are essential for decision making and providing a basis for intervention.
|