ARTICLE
Auteur(s) : Côme
Daniau1, France Wallet2, Pierre-André
Cabanes2
1InVS Département Santé Environnement 12, rue du Val
d’Osne 94415 Saint-Maurice cedex France
2EDF Direction Dynamique et Politique RH Service
des études médicales 22, rue Joubert 75009 Paris France
Article reçu le 7 Septembre 2009, accepté le 26 Janvier 2010
Several legionellosis outbreaks have been associated with
cooling towers that are relatively low in height, such as those for
air conditioning systems [1-6]. More recently, industrial cooling
towers were incriminated in the largest legionellosis outbreak in
France, which had 86 identified cases [1, 7]. The epidemic
strain was found in water circulating in these towers,
6 meters high. Two other studies, conducted in the United
States, looked specifically for differences in the prevalence of
legionella antibodies in workers in electric power plants with
cooling towers. Although one study did not find a relation between
this prevalence and exposure, the other observed an increase in
seroprevalence at a titer of 1:128 according to exposure to the
plume. Both, however, involved stacks that averaged less than
20 meters in height [8, 9]. To our knowledge, no published
studies establish an association between Legionnaires’ disease and
exposure to plumes of very tall industrial cooling towers, such as
those of most of closed-circuit nuclear power plants, even though
conditions there, especially nutrients and water temperature, are
favorable to the development of Legionella bacteria [10, 11].
The difference in height between these towers and those involved
in the earlier epidemic affects the atmospheric dispersion of the
bacteria. The greater dispersion from taller towers is associated
with delayed arrival of Legionella in the exposure area, a delay
that reduces their viability.
Contact with these bacteria can cause the formation of
antibodies that may last for several months to several years in the
organism without any noticeable symptoms [12]. The aim of this
study was to determine exposure to Legionella pneumophila from the
plume of nuclear power plants by comparing the prevalence of
antibodies against L. pneumophila, an indicator of Legionella
contact, among nuclear power plant workers who were and were not
exposed to plumes from the plants’ cooling towers.
Methods
Occupational physicians recruited workers of four nuclear power
plants during annual routine examinations. Exclusion criteria were:
employment at the plant for less than one year, part-time work
(less than 80%), and sick-leave or maternity leave for two months
or longer during the twelve months preceding inclusion. The survey
took place from January through December 2002 in each plant,
to avoid the biases associated with seasonal factors of Legionella
exposure [2]. Volunteers were included progressively during the
year until we reached the planned sample size of 500 subjects
in each exposure group.
Exposure assessment to the plume
We selected four plants that use the same industrial processes and
were constructed during the same period. Each site has two to four
reactors, for power of 900 to 1300 megawatts. They differ
in their tertiary cooling loops, which are either open or closed.
The Blayais plant, located along the seacoast, cools water in a
tertiary loop open to the sea. It therefore has no cooling tower.
The heights of the cooling towers of the other three plants
differed. Because of landscape constraints, the Chinon plant tower
is only 28 meters high. Despite fans that push the plume
upward, it may be blown back down by the wind to soil level at the
plant site and outside as far away from the plant as 10 km.
Atmospheric modeling indicates that the maximum atmospheric
concentration of Legionella is then at the plant site. The cooling
towers at Belleville-sur-Loire and Cattenom are much higher:
165 meters. Their plumes are widely dispersed over a 20-km
radius, and modeling indicates the maximum Legionella concentration
is at around 2 km from the plant [13].
There are thus three groups with different levels of exposure to
cooling tower plumes:
- – highly exposed: Chinon plant;
- – slightly exposed: Belleville-sur-Loire and Cattenom
plants;
- – unexposed: Blayais plant.
For the three plants with cooling towers – Chinon, Belleville
and Cattenom – regular sampling is performed at representative
points, following the NF T 90-431 culture method. Annual monitoring
of Legionella concentrations shows that all cooling pond water
samples are regularly contaminated. Concentrations are lowest in
the winter, increase in the spring, peak in the summer, and
decrease in autumn. They range from (min and max) 5×103
to 7×105 CFU/l1 for Belleville;
4×104 to 1.1×106 for Cattenom and
2×104 to 1.3×106 for Chinon. All the strains
were L. pneumophila, with serogroups 1-6 accounting for 70 to
90% of the population.
Survey questionnaire
Volunteers completed a standardized questionnaire in the
occupational physician’s presence. It asked about:
- – social and demographic characteristics, including age,
sex, place and type of residence;
- – information about recent medical episodes, fever,
bronchitis or lung disease;
- – information related to potential exposure to
Legionella from the plume of cooling towers: plant of employment,
distance between home and plant;
- – other Legionella exposure factors at work, including
jobs involving direct contact with aerosolized water: 14 tasks
were identified, including working in an operating cooling tower,
cleaning cooling tower basins, and replacing splash slats and
packing in cooling towers, etc.); other work-related exposure
factors included failure to wear protection (such as masks or
ventilated helmets) during tasks at these at-risk jobs [8, 9,
14-16], taking showers at work, and exposure to spray (large water
droplets from the bottom of the tower that cannot be inhaled and
are thus not considered infectious);
- – other non-work related exposure factors were also
recorded: all other showers (location and number per week), recent
plumbing renovations at home, and the method of hot water
production in the primary residence.
Serology
During the routine occupational medicine consultation, an extra
tube of 5 mL of blood was taken for the serology testing. The
tube was centrifuged and the serum stored at -18 °C in two
microtubes. All tubes were sent to the microbiology laboratory at
Henri Mondor Hospital where the antibodies to L. pneumophila were
assayed by two successive techniques. First, enzyme-linked
immunosorbent assays (ELISA) tested for serum immunoglobulin (Ig) G
and M antibodies directed against L. pneumophila serogroups
1 to 7 (Lp 1-7 IgG and IGM Kit SERION ELISA classic: ref. kit
ESR106G and ESR106M) [17, 18]. Samples considered positive or
uncertain for IgG, that is, with a concentration of ≥ 50 arbitrary
units per mL (AU/mL), were then tested by indirect
immunofluorescence (IIF) to measure the antibody titer against
different L. pneumophila serogroups (kit for Legionella IIF
polyvalent antigens, serogroups 1-6, Meridian Diagnoses Europe).
Individual results were obtained for different dilution levels
(cutoff points): 1:16, 1:32, 1:64, and 1:128. According to the
French disease definition, only a titer ≥ 1:128 is meaningful for a
legionellosis diagnosis [1]. The other titers reflect contact with
Legionella but not necessarily any disease.
Statistical analyses
Serological results were studied as binary variables for each
cutoff point and for each exposure category, as defined above (3
classes of nuclear power plants with different types of cooling
tower plumes).
SAS software version 8.2 was used for the statistical
analyses.
Univariate analysis considered the variable to be explained,
that is, the presence of antibodies to Legionella, and the
explanatory variable, defined as plume exposure. We tested
associations with Pearson’s bilateral χ2 test
(parametric test) or with Fisher’s exact test, used for small
sample sizes. Trends were evaluated by the Cochran-Armitage
unilateral χ2 trend test.
We used logistic regression to assess the independent effects of
the exposure variables and other risk factors. The variables used
in the multivariate logistic model were those that were significant
(p < 0.20) in the univariate analysis. Results are expressed as
adjusted odds ratios (OR) and their 95% confidence interval
(CI).
Results
Overall, the study included 1,519 people who completed the
questionnaire and provided blood samples. Although all subjects who
were asked to fill out the questionnaire did so, 59 subjects
completed the questionnaire without providing blood samples and
have been excluded from the analysis. The study population was
almost entirely male (95%), with an age range of 19 to
61 years and a mean age of 40.7 years; 81% lived in rural
areas and 97% in one-family houses. Most of these workers (95%)
lived within a 20-km radius of the plant where they worked, that
is, within the plume dispersion area. Their home hot water was
produced principally by electric water heaters (94%), and very few
had had plumbing work done during the study period (12%). At some
point during the study year, most (88%) had taken showers outside
their homes: in a hotel, at campgrounds, in gyms or spas, or at
work. Nearly half (46%) took showers at work, but only one quarter
of them took more than one such shower a month.
Fewer than half the volunteers were assigned to the tasks
considered to involve important Legionella exposure (42%), beyond
their exposure to the cooling tower plumes: 42% had performed at
least one of the 14 tasks during the year, 26% at least two
and only 13% at least three. Of the workers who were supposed to
wear protective masks during activities that risked Legionella
exposure (59%), three quarters actually did so. Exposure to spray
was frequent at the plants with cooling towers, according to 52% of
the subjects; 15% of them reported exposure to spray at least once
a week.
Overall 2.83% (43 of 1,519) of all samples were positive or
doubtful for IgG in the ELISA test (≥ 50 AU/mL). The tests for
IgM were all negative. The prevalence rate of workers in our sample
with antibodies against Legionella was 1.51% (23/1519) for an
IIF titer ≥ 1:16, 1.05% (16/1519) for ≥ 1:32, and 0.53%
(8/1519) for ≥ 1:64. No volunteer at any plant had positive results
that could be interpreted as a disease diagnosis according to
French criteria [1] (IIF titer ≥ 1:128).
Table 1 presents seroprevalence
results according to the thresholds for positive findings and
exposure group. Seroprevalence according to IIF at cutoff points ≥
1:64 and ≥ 1:32 did not differ between the three exposure levels.
When we add all the samples with a titer ≥1:16, prevalence
rises through the three exposure levels with exposure level, but
not significantly (p = 0.12).
The results showed no statistically significant difference in
antibody prevalence between the unexposed and the (slightly or
highly) exposed population: p = 0.53 for IIF titer 1:16 and p =
0.35 for titer 1:32.
Table 2 presents the associations
between seroprevalence and the risk factors examined here. Analysis
of tasks involving exposure did not identify any particular type of
activity that increased the risk of Legionella exposure. On the
other hand, not wearing masks for respiratory protection when
recommended (for activities involving particular exposure) was
identified as a significant (p = 0.01) risk factor for positive
Legionella serology results at a threshold of 1:16 (OR = 5.94; 95%
CI [1.47; 23.94]). The association was not, however, significant
for a titer ≥ 1:32 (table 3).
Age and failure to wear masks when recommended were taken into
account as adjustment variables in the regression models because
they were distributed differently between the different exposure
groups. After adjustment for them, multivariate analysis showed no
significant increase in antibody prevalence regardless of the
threshold. Results for the slightly exposed versus unexposed (OR =
2.55, 95% CI [0.72; 9.04] for threshold 1:16 and 0.96 95% CI [0.26;
3.62] for threshold 1:32) were similar to those for the highly
exposed versus unexposed (OR = 1.89, 95% CI [0.56; 6.31] for
threshold 1:16 and 1.27, 95% CI [0.37; 4.39] for threshold
1:32).
Table 1 Results of univariate analysis
of seroprevalence of antibodies to Legionella
pneumophila according to IIF titer and plume exposure group,
based on the nuclear power plant of employment with
different cooling tower heights.Tableau 1. Résultats
de l’analyse univariée de la séroprévalence
des anticorps anti-Legionella penumophila selon le titre
en immunofluorescence et le groupe d’exposition basé
sur la centrale d’appartenance avec différentes
hauteurs de tour aéroréfrigérantes.
|
IIF Titer
|
Prevalence (%)
|
Statistics*
|
|
Unexposed group n = 409 (27.0%)
|
Slightly exposed group n = 529 (34.8%)
|
Highly exposed group n = 581 (38.2%)
|
|
≥ 1:64
|
0.73
|
0.38
|
0.52
|
p = 0.77
|
|
≥ 1:32
|
0.98
|
0.95
|
1.20
|
p = 0.90 (trend test: p = 0.35)
|
|
≥ 1:16
|
0.98
|
1.51
|
1.89
|
p = 0.49 (trend test: p = 0.12)
|
Table 2 Associations between seroprevalence
(for the thresholds of 1:16 and 1:32) and all
of the risk factors assessed
in the study.Tableau 2. Associations entre
la séroprévalence (pour les seuils de 1 :16
et 1 :32) et tous les facteurs de risques
évalués dans cette étude.
|
Risk factors
|
Statistical analysis
|
|
IFI Titer 1:16
|
IFI titer 1:32
|
|
Social and demographic characteristics
|
|
Age in 3 categories (19-37 years; 38-45 years; 46-63
years)
|
0.03
|
0.01
|
|
Sex
|
0.35 +
|
0.48 +
|
|
Environment (rural or urban)
|
0.55 +
|
0.61 +
|
|
Housing (multiple dwelling units or single-family homes)
|
0.485 +
|
0.60 +
|
|
Non-work-related exposure factors
|
|
Distance between home and plant (inside or outside
a 3 km perimeter from the plant)
|
0.53 +
|
0.59 +
|
|
Hot water production at home (electric hot water heater or
other)
|
0.25 +
|
0.38 +
|
|
Recent plumbing work at home (within the past year)
|
0.13 +
|
0.23 +
|
|
Has taken showers somewhere other than home or work
in the past year (yes/no)
|
0.57 +
|
0.46 +
|
|
Exposure factors at work
|
|
Has taken showers at work (< or > 1 time per month)
|
0.18
|
0.19
|
|
Performs tasks with particular exposure (never or at least
once)
|
0.61
|
0.42
|
|
Frequency of exposure during those tasks (0, 1, 2 or ≥
3)
|
0.30 ++
|
0.41 ++
|
|
Wears P3 protective mask (yes or no)
|
0.01
|
0.11
|
|
Wears complete helmet (yes or no)
|
0.75
|
0.74
|
|
Exposure to spray (yes or no)
|
0.81
|
0.38
|
|
Frequency of exposure to spray (never, once a year
to once a month; once a month to once
a week; several times a week)
|
0.21 ++
|
0.07 ++
|
|
Recent medical episode
|
|
Fever in the past year
|
0.43
|
0.44
|
|
Bronchitis in the past year
|
0.24 +
|
0.08 +
|
|
Pneumonia in the past year
|
0.68 +
|
0.76 +
|
Table 3 Details of the univariate relation
between the risk factors with a significant relation (p
< 0.20) to seroprevalence for both thresholds: 1:16
and 1:32 (crude odds ratio with 95% confidence intervals).Tableau
3. Détails de la relation univariée entre
les facteurs de risques avec une relation
significative (p < 0,20) à la séroprévalence
pour les deux seuils 1 :16 et 1 :32 (odds ratio
bruts avec intervalle de confiance à 95 %).
|
Risk factor
|
Number
|
IFI titer 1:16
|
IFI titer 1:32
|
|
n
|
%
|
Prevalence (%)
|
OR [95% CI]*
|
Prevalence (%)
|
OR [95% CI]*
|
|
Age in 3 categories
|
|
19-37 years
|
476
|
31
|
1.26
|
1.00
|
0.84
|
1.00
|
|
38-45 years
|
548
|
36
|
2.55
|
2.05 [0.78-5.39]
|
2.01
|
2.42 [0.76-7.64]
|
|
46-63 years
|
494
|
33
|
0.61
|
0.48 [0.12-1.92]
|
0.20
|
0.24 [0.03-2.15]
|
|
Recent plumbing work
|
|
No
|
1,281
|
85
|
1.33
|
1.00
|
-
|
-
|
|
Yes
|
187
|
12
|
1.60
|
1.21 [0.35-4.18]
|
-
|
-
|
|
Showers at plant
|
|
<once a month
|
1,134
|
75
|
1.76
|
1.00
|
1.23
|
1.00
|
|
≥once a month
|
381
|
25
|
0.79
|
0.44 [0.13-1.50]
|
0.52
|
0.42 [0.10-1.87]
|
|
Wears a protective mask
|
|
Yes
|
683
|
46
|
0.44
|
1.00
|
0.44
|
1.00
|
|
No
|
235
|
16
|
2.55
|
5.94 [1.47-23.94]
|
1.28
|
2.93 [0.59-14.62]
|
|
Frequency of exposure to spray at plant
|
|
No
|
681
|
47
|
-
|
-
|
1.32
|
1.00
|
|
Once a year to once a month
|
442
|
30
|
-
|
-
|
1.36
|
1.03 [0.36-2.91]
|
|
Once a month to once a week
|
208
|
14
|
-
|
-
|
0.48
|
0.36 [0.05-2.86]
|
|
> Once a week
|
111
|
7
|
-
|
-
|
0
|
..
|
|
Bronchitis episodes
|
|
No
|
1,229
|
85
|
-
|
-
|
0.90
|
1.00
|
|
Yes
|
221
|
15
|
-
|
-
|
2.26
|
2.56 [0.88-7.45]
|
Discussion
The point of this study was to explore whether exposure
to Legionella from the plume of power plant cooling towers
produces measurable effects on L. pneumophila antibodies among
workers who are and are not exposed, either slightly or highly, to
plumes from the plants’ cooling towers.
This study measured the rate of past contact with Legionella and
not the risk of Legionaires’ disease, for which risk factors
including age and underlying disease would play a role.
Results show that no difference was observed between the three
exposure levels defined by plant of employment.
A nonsignificant trend towards increased seroprevalence with
exposure was observed, but only for an IFI titer of 1:16. No study
volunteer, from any plant, had a serology positive at a
sufficiently high titer to diagnose Legionnaire’s disease (IIF
titer ≥ 1:128).
The low prevalence rates observed reduce the study’s power and
its capacity to show associations with risk factors for positive
Legionella serology results. The study was nonetheless able to show
the importance of wearing a mask for protection against Legionella
exposure in tasks involving particular exposure not related to
exposure to the plume from high cooling towers.
Exposure assessment to Legionella from the plume
of power plant cooling towers
As in other seroprevalence studies, the categories of exposure to
the plumes of power plant cooling towers can be considered only an
indirect and qualitative assessment of exposure to Legionella. This
classification is based, on the one hand, on the existence of a
cooling tower (no exposure at Blayais and exposure at the other
sites). For the sites with cooling towers, similar levels of
Legionella were found in the cooling circuits at all the sites. On
the other hand, the two categories of exposure (slightly and highly
exposed) are based on processes of plume dispersion that differ for
the tall cooling towers and the smaller ones [13], processes that
show that the level of exposure at Chinon would be higher than at
the other sites with tall towers (Belleville and Cattenom).
Even if there is a seasonal variability of Legionella
concentrations in cooling circuits, regular sampling showed that L.
pneumophila was always present in the cooling pond water at
concentrations higher than 1×103 in UFC/l and confirmed
a possible exposure to Legionella from the plume of cooling
tower.
Studies modelling the dispersion of the plumes from high towers
show dispersion of the plume far away from cooling towers [13]. As
the greater part of volunteers of the study (95%) were living
within a 20-km radius of the plant where they worked, exposure to
the plume of tall cooling towers is also possible outside the
site.
However, results of our study do not show any statistical
significant differences between the rates of seroprevalence of L.
pneumophila antibodies at the different sites with and without
cooling towers. These results do not support the hypothesis that
plumes from very tall cooling towers cause measurable exposure to
Legionella bacteria in all the area of dispersion. We can presume
that atmospheric conditions during the plume dispersion from very
tall cooling towers may be incompatible with bacteria viability in
the plume.
Serology as exposure marker
Serum antibody assays have now been replaced by urinary antigen
tests for diagnostic testing. This technique could not be applied
to this study where there was no question of tissue infection.
Moreover it does not appear to be conclusive in cases of Pontiac
fever (a milder form of Legionnaires’ disease) [19, 20]. On the
other hand, serum antibody methods have numerous advantages. In
particular, because these antibodies remain in the body for a long
time, testing for them is much more useful for a cross-sectional
study [9, 12].
This study used two test batteries: an ELISA-IgG and IgM test
for initial screening and an IIF test to quantify the antibody
level. The ELISA test is easy to perform and less burdensome than
the IIF method. The sensitivity of this ELISA test, compared with
the IIF method, varies from 64 to 91% for a 1:128 titer
threshold [18, 21]. The laboratory quality control files show
equivalent results for the samples tested. Moreover, while the IIF
method has advantages for antibody titration, the ELISA method can
determine the presence of IgM, which demonstrates recent
infection.
The serology testing conducted for this study (titers of 1:16,
1:32, and 1:64) detects seroconversion even in the absence of
disease and must be interpreted as a biological exposure marker or
indicator rather than a morbidity indicator. An IIF titer of 1:16,
which corresponds to the resolution limit of the IIF test
technique, is difficult to interpret because of the potential for
false positives from nonspecific or cross-reactions [22, 23]. These
cross-reactions may produce high antibody levels, especially for
infections for some serogroups and species of Chlamydia, for
example. On the other hand, an IIF titer of 1:32 may be interpreted
as slight exposure or old seropositivity or a serologic scar;
cross-reactions remain possible but are less frequent at this
dilution level. A titer of 1:64 shows Legionella exposure with
more certainty and is often considered a high titer in the
literature [24, 25]. Nonetheless, the ELISA and IIF tests used in
this study are normally used for diagnostic purposes, and there are
no expert recommendations about the significance of titers less
than 1:128 with IIF or results below 70 AU/mL for the ELISA
test. Numerous studies show a correlation between the presence of
low antibody titers and exposure level, but not between
seroprevalence and the existence of risk factors associated only
with disease onset [17, 26-31]. In this study, only exposure
factors were taken into account; we did not consider disease risk
factors such as tobacco.
The low number of positive results, however, prevented us from
using the titer of 1:64 as the only indicator of Legionella
exposure. We decided to analyze the results for all thresholds,
while bearing in mind that the meaning of a dilution of 1:16 is far
from clear.
Comparison of the seroprevalence in our
population with prevalence data
in the literature
Seroprevalence rates in the literature vary from less than 1% [32]
to 50% [33]. Variations depend on the type of population studied,
place, season, screening test used, and cutoff point. Table 4 summarizes some IIF results from several
studies in the general population (chosen for comparison because
our results showed no significant difference between exposed and
unexposed personnel). Several multicenter studies show differences
in rates and serogroups-according to geographic location [27]. The
geographic disparities encountered, however, do not suggest the
application of any general laws, such as a north/south gradient. We
therefore decided to compare our sample only with populations in
similar geographic zones and so selected only French studies. The
three studies we found all assessed antibody prevalence in urban
populations.
In 1987, Bornstein tested blood donors in three cities (Lyon,
Nice, and Poitiers; n = 583) and found a seroprevalence of 5.1%
(IIF ≥ 1:16, L. pneumophila serogroups 1-6) [24]. In 1992,
Desenclos (unpublished data) found a seroprevalence of 11.8% (IIF ≥
1:16, L. pneumophila serogroups 1-6) in a population of Paris
blood donors (n = 144). Seroprevalence (IIF ≥ 1:16, L. pneumophila
serogroup 1-6) in our study of nuclear power plant workers at
Blayais, which has no cooling tower, was lower than in these two
studies (0.96%, n = 409). The seroprevalence rates in our study
(0.53% for a titer ≥ 1:64, 1.05% for a titer ≥ 1:32, and 1.51% for
a titer ≥ 1:16) are, however, close to that found earlier by
Dournon [32], 0.90% (IIF ≥ 1:16) in a population of 450 blood
donors in Paris in 1983.
The difference in seroprevalence between later studies of urban
populations may be explained by the increased exposure (more
cooling towers for air conditioning systems). Our study population
is rural and only slightly exposed (lower concentrations due to air
conditioning cooling towers, fewer collective hot water systems),
and composed of healthy workers, compared with the urban
populations used as a reference in the other French studies. The
lack of standardization between laboratories in their methods for
detecting antibodies to L. pneumophila also makes comparisons
difficult. The two-stage strategy we used for analysis, combining
an initial ELISA test, followed by an IFI analysis only if positive
or uncertain, may also account for the low seroprevalence rates we
observed.
Table 4 Several studies of Legionella pneumophila
antibody prevalence in general populations.Tableau 4.
Différentes études de prévalence des anticorps
anti-Legionella pneumophila en population générale.
|
Authors, year
|
Ref.
|
Study population
|
Number
|
Country
|
Prevalence (%)
|
Technique, Titer, Genus, Species, Serogroup
|
|
Storch, 1979
|
[31]
|
Healthy volunteers (≥45 years)
|
1,143
|
USA (4 States)
|
1.7
|
IIF; ≥1/64 Lp
|
|
Snowman, 1982
|
[34]
|
Healthy workers
|
588
|
USA (Ohio)
|
19.9 17.9
|
IIF; ≥1/128 Lp Sg 1 IIF; ≥1/128 Lp Sg 2
|
|
Wilkinson, 1983
|
[35]
|
Blood donors
|
184
|
USA (12 States)
|
12 41
|
IIF; ≥1/256 Lp IIF; ≥1/128 Lp
|
|
Dournon, 1983
|
[32]
|
Blood donors
|
450
|
France (Paris)
|
0.9
|
IIF; ≥1/16 Lp Sg 1
|
|
Nadarajah, 1987
|
[36]
|
Blood donors (patients hospitalized with pneumonia)
|
150 166
|
Singapore
|
20 3
|
IIF; ≥1/16 Lp IIF; ≥1/256 Lp
|
|
Bornstein, 1987
|
[24]
|
Blood donors
|
583
|
France (Lyon, Nice, Poitiers)
|
5.1 0 2.5
|
IIF; ≥1/16 Lp Sg 1 à 6 IIF; ≥1/64 Lp Sg 1 à 6
IIF; ≥1/16 Lp Sg 1
|
|
Romano, 1989
|
[37]
|
Healthy population
|
562
|
Italy
|
23 9.4
|
IIF; ≥1/256 Lp Sg 1 à 6 IIF; ≥1/256 Lp Sg 3
|
|
Desenclos, 1993
|
-
|
Urban blood donors
|
144
|
France (Paris)
|
11.8 4.9 2.1 1.4
|
IIF; ≥1/16 Lp Sg 1 à 6 IIF; ≥1/32 Lp Sg 1 à 6
IIF; ≥1/64 Lp Sg 1 à 6 IIF; ≥1/16 Lp Sg 1
|
|
Lee, 2008
|
[38]
|
Healthy individuals
|
500
|
Korea
|
15.2
|
IIF; ≥1/128 Lsp
|
|
Lobos, 1993
|
[39]
|
Blood donors
|
100
|
Chile
|
5
|
IIF; ≥1/64 Lp Sg 1 à 6
|
|
Rocha, 1995
|
[40]
|
Blood donors
|
503
|
Portugal
|
1.2
|
IIF; ≥1/64 Lp
|
|
Franzin, 1995
|
[29]
|
Blood donors
|
777
|
Italy
|
0.3
|
IIF; ≥1/8 Lp Sg 1
|
|
Heng, 1997
|
[41]
|
Healthy population
|
719
|
Singapore
|
17.5
|
IIF; ≥1/32 Lp Sg 1 à 6
|
|
Rudbeck, 2008
|
[42]
|
Blood donors
|
708
|
Denmark (2 cities)
|
22.9
|
IIF; ≥1/128 Lsp
|
Conclusion
The prevalence of antibodies to L. pneumophila is very low in the
population of nuclear power plant workers exposed to the plumes of
tall cooling towers, even though the cooling circuits are regularly
contaminated. Despite the inclusion of a large proportion of
workers we found no significant difference between the three
exposure levels, characterized by plant of employment and
specifically the existence and height of cooling towers. When we
analyzed all the samples with a titer of at least 1:16, prevalence
rose with exposure level across the three exposure levels, but not
significantly. Nevertheless, the statistical analysis showed no
statistically significant difference in antibody prevalence between
the unexposed and (slightly and highly) exposed populations for
titers of 1:16 and 1:32.
These results do not support the hypothesis that plumes from
very tall cooling towers cause high exposure to Legionella
bacteria, although the limitations of its exposure assessment must
be borne in mind.
The study also shows that for exposure from other Legionella
sources at work not wearing a mask for respiratory protection when
recommended, for activities involving particular exposure, is a
significant risk factor for positive Legionella serology results.
In terms of workplace risk management, this study demonstrates that
wearing a protective mask appears to be effective against
Legionella exposure during tasks identified as entailing a specific
risk of exposure. It is therefore important to make staff aware of
the importance of wearing protective masks in the recommended areas
by signs stating that masks are mandatory and to encourage
compliance with the instructions for their use.
Acknowledgments
We thank Dr Lionel Desforges, microbiologist at the Henri Mondor
Hospital (Créteil, France) for performing the L. pneumophila
antibody assays. We also thank Pascale Bernillon for her support on
the statistic analysis, and the occupational physicians who
participated in the recruitment of volunteers at the four sites.
Disclosures: the authors declare that they have no
competing interests.
References
1 Institut de veille sanitaire (InVS). Numéro spécial consacré à la
légionellose. Bulletin épidémiologique hebdomadaire 2004 ; 36-37 :
173-85.
2 Bentham RH, Broadbent CR. A model for autumn
outbreaks of Legionnaires’ disease associated with cooling towers,
linked to system operation and size. Epidemiol Infect 1993; 111:
287-95.
3 Brown A, Lema MW, Brown-Schlumpf MS. Antigenic
specificity of the antibody response in humans during
legionellosis. Infection 1986; 14: 108-14.
4 Formica N, Tallis G, Zwolak B, et al.
Legionnaires’ disease outbreak: Victoria’s largest identified
outbreak. Commun Dis Intell 2000; 24: 199-202.
5 Garbe PL, Davis BJ, Weisfeld JS, et al.
Nosocomial Legionnaires’ disease. Epidemiologic demonstration of
cooling towers as a source. Jama 1985; 254: 521-4.
6 Keller DW, Hajjeh R, DeMaria A, et al.
Community outbreak of Legionnaires’ disease: an investigation
confirming the potential for cooling towers to transmit Legionella
species. Clin Infect Dis 1996; 22: 257-61.
7 Nguyen TM, Ilef D, Jarraud S, et al. A
community-wide outbreak of legionnaires disease linked to
industrial cooling towers--how far can contaminated aerosols
spread? J Infect Dis 2006; 193: 102-11.
8 Buehler JW, Kuritsky JN, Gorman GW,
Hightower AW, Broome CV, Sikes RK. Prevalence of
antibodies to Legionella pneumophila among workers exposed to a
contaminated cooling tower. Arch Environ Health 1985; 40:
207-10.
9 Deubner DC, MacCormack JN, Kleeman K,
Muhlbaier LH. One-time screening to define the problem:
Legionella exposure in an electric power company. J Occup Med 1986;
28: 670-3.
10 Kurtz JB, Bartlett CL, Newton UA,
White RA, Jones NL. Legionella pneumophila in cooling
water systems. Report of a survey of cooling towers in London and a
pilot trial of selected biocides. J Hyg 1982; 88: 369-81.
11 Witherell LE, Novick LF, Stone KM, et al.
Legionella in cooling towers. J Environ Health 1986; 49: 134-9.
12 Darelid J, Lofgren S, Malmvall BE,
Olinder-Nielsen MA, Briheim G, Hallander H.
Legionella pneumophila serogroup 1 antibody kinetics in
patients with Legionnaires’ disease: implications for serological
diagnosis. Scand J Infect Dis 2003; 35: 15-20.
13 Afsset. Évaluation des risques sanitaires liés aux
proliférations de Legionella dans l’eau des tours aéroréfrigérantes
des centres nucléaires de production d’électricité d’EDF. Première
phase d’analyse : évaluation de la pertinence des seuils d’action
actuellement recommandés. Maisons Alfort : Afsset, 2006.
14 Fraser DW, Deubner DC, Hill DL,
Gilliam DK. Nonpneumonic, short-incubation-period
Legionellosis (Pontiac fever) in men who cleaned a steam turbine
condenser. Science 1979; 205: 690-1.
15 Lauderdale JF, Johnson CC. An outbreak of acute
fever among steam turbine condenser cleaners. Am Ind Hyg Assoc J
1983; 44: 156-60.
16 Werner HP, Pietsch M. Incidence and significance of
Legionella in industrial air conditioning systems. Schriftenr Ver
Wasser Boden Lufthyg 1993; 91: 231-40.
17 Boshuizen HC, Den Boer JW, de Melker H,
et al. Reference values for the SERION classic ELISA for
detecting Legionella pneumophila antibodies. Eur J Clin Microbiol
Infect Dis 2003; 22: 706-8.
18 Etienne J, Vandenesch F, Lina G, Jarraud S. Evaluierung des
SERION ELISA classic Legionella pneumophila 1-7 (IgG, IgM). SERION
Legionella pneumophila 1-7 studies ELISA classic. Leusden : Clindia
Benelux, 1999.
19 Friedman S, Spitalny K, Barbaree J,
Faur Y, McKinney R. Pontiac fever outbreak associated
with a cooling tower. Am J Public Health 1987; 77: 568-72.
20 Gregersen P, Grunnet K, Uldum SA,
Andersen BH, Madsen H. Pontiac fever at a sewage
treatment plant in the food industry. Scand J Work Environ Health
1999; 25: 291-5.
21 Yzerman EP, den Boer JW, Lettinga KD,
Schel AJ, Schellekens J, Peeters M. Sensitivity of
three serum antibody tests in a large outbreak of Legionnaires’
disease in the Netherlands. J Med Microbiol 2006; 55: 561-6.
22 Boswell TC, Marshall LE, Kudesia G.
False-positive legionella titres in routine clinical serology
testing detected by absorption with campylobacter: implications for
the serological diagnosis of legionnaires’ disease. J Infect 1996;
32: 23-6.
23 Musso D, Raoult D. Serological cross-reactions
between Coxiella burnetii and Legionella micdadei. Clin Diagn Lab
Immunol 1997; 4: 208-12.
24 Bornstein N, Janin N, Bourguignon G,
Surgot M, Fleurette J. Prevalence of anti-Legionella
antibodies in a healthy population and in patients with
tuberculosis or pneumonia. Pathol Biol 1987; 35: 353-6.
25 Sampson IA. Prevalence of antibody to Legionella
pneumophila in aborigines and non-aborigines in Western Australia.
Med J Aust 1988; 148: 16-9.
26 Bell JC, Jorm LR, Williamson M, et al.
Legionellosis linked with a hotel car park--how many were infected?
Epidemiol Infect 1996; 116: 185-92.
27 Borella P, Mazzini E, Bargellini A, Montagna MT, Tato D,
Napoli C. Is there a professional risk of legionnaires disease?
Results of a multicentric serological study. European Working group
for Legionella Infections 20th annual meeting. Roma: Istisan
Congressi, 2005.
28 Boshuizen HC, Neppelenbroek SE, van Vliet H,
et al. Subclinical Legionella infection in workers near the
source of a large outbreak of Legionnaires disease. J Infect Dis
2001; 184: 515-8.
29 Franzin L, Scramuzza F. Prevalence of Legionella
pneumophila serogroup 1 antibodies in blood donors. Eur J
Epidemiol 1995; 11: 475-8.
30 Helms CM, Johnson W, Renner ED,
Hierholzer WJJ, Wintermeyer LA, Viner JP. Background
prevalence of microagglutination antibodies to Legionella
pneumophila serogroups 1, 2, 3, and 4. Infect Immun 1980; 30:
612-4.
31 Storch G, Hayes PS, Hill DL, Baine WB.
Prevalence of antibody to Legionella pneumophila in middle-aged and
elderly Americans. J Infect Dis 1979; 140: 784-8.
32 Dournon E, Mayaud C, Bure A, Desplaces N,
Christol D. Epidemiological features of Legionnaires’ disease
in the Paris area. Zentralbl Bakteriol Mikrobiol Hyg A 1983; 255:
76-83.
33 Reinthaler F, Mascher F, Stunzner D.
Legionella pneumophila: seroepidemiologic studies of dentists and
dental personnel in Austria. Zentralbl Bakteriol Mikrobiol Hyg B
1987; 185: 164-70.
34 Wilkinson HW, Reingold AL, Brake BJ,
McGiboney DL, Gorman GW, Broome CV. Reactivity of
serum from patients with suspected legionellosis against
29 antigens of legionellaceae and Legionella-like organisms by
indirect immunofluorescence assay. J Infect Dis 1983; 147:
23-31.
35 Snowman WJ, Holtzhauer FJ, Halpin TJ,
Correa-Villasenor A. The role of indoor and outdoor
occupations in the seroepidemiology of Legionella pneumophila. J
Infect Dis 1982; 145: 275.
36 Nadarajah M, Singam S, Jalil HA. Sero-survey
for Legionella pneumophila antibodies--Singapore experience. Ann
Acad Med Singapore 1987; 16: 583-5.
37 Romano F, Ribera G, D’Errico MM,
Grasso GM, Montanaro D. Levels of anti-Legionella
antibodies in a healthy Neapolitan population. Ann Ig 1989; 1:
629-36.
38 Lee HK, Woo MK, Ju YI, et al. Prevalence
of antibodies in response to Legionella species, analysis of a
healthy population from Jeollanam-do Province, Korea. J Microbiol
2008; 46: 160-4.
39 Lobos T, Moreno R, Carstens M, et al.
Seroprevalence of Legionella pneumophila infection in healthy,
adults from Santiago, Chile. Rev Med Chil 1993; 121: 1123-7.
40 Rocha G, Verissimo A, Bowker R,
Bornstein N, Da Costa MS. Relationship between Legionella
spp. and antibody titres at a therapeutic thermal spa in Portugal.
Epidemiol Infect 1995; 115: 79-88.
41 Heng BH, Goh KT, Ng DL, Ling AE.
Surveillance of legionellosis and Legionella bacteria in the built
environment in Singapore. Ann Acad Med Singapore 1997; 26:
557-65.
42 Rudbeck M, Molbak K, Uldum S. High prevalence
of antibodies to Legionella spp. in Danish blood donors.
A study in areas with high and average incidence of
Legionnaires’ disease. Epidemiol Infect 2008; 136: 257-62.
1 CFU: colony forming unit, unité format
colonie (UFC).
|
Version imprimable
Version PDF
