Alternative Routes for Dissemination of Legionella pneumophila

Oct 15, 2010 - Legionella pneumophila is the etiological agent of Legionnaires' disease (LD) and of the nonpneumonic Pontiac fever. In addition to L...
2 downloads 0 Views 4MB Size
Download PDF
Environ. Sci. Technol. 2010, 44, 8712–8717

Alternative Routes for Dissemination of Legionella pneumophila Causing Three Outbreaks in Norway JARAN STRAND OLSEN,† TONE AARSKAUG,† INGJERD THRANE,† CHRISTINE POURCEL,‡ EIRIK ASK,§ GISLE JOHANSEN,| VIGGO WAAGEN,| A N D J A N E T M A R T H A B L A T N Y * ,† Norwegian Defence Research Establishment (FFI), P.O. Box 25, N-2027 Kjeller, Norway, GPMS, Institut de Ge´ne´tique et Microbiologie, Baˆt 400, Universite´ Paris-Sud, 91405 Orsay cedex, France, Unilabs Telelab AS, P.O. Box 1868, Gulset, N-3703 Skien, Norway, and Borregaard Ind. Ltd., P.O. Box 162, N-1701 Sarpsborg, Norway

Received March 10, 2010. Revised manuscript received September 15, 2010. Accepted September 15, 2010.

Three outbreaks of Legionnaires’ disease were reported in the Fredrikstad/Sarpsborg community, Norway, in 2005 and 2008 caused by the L. pneumophila ST15 and ST462 strains determined by sequence based typing. In this retrospective study, we suggest that the aeration ponds, a part of the biological treatment plant at Borregaard Ind. Ltd., are the main amplifiers and primary disseminators of the outbreak L. pneumophila strains. This result is supported by the finding that the ST15 and ST462 strains were not able to survive in air scrubber liquid media more than two days of incubation at the scrubber’s operating conditions during the 2005 and 2008 outbreaks. In 2008, >1010 CFU/L of L. pneumophila ST462 were detected in the aeration ponds. ST15 and ST462 were also detected in the river Glomma in 2005 and 2008, respectively, downstream of the wastewater outlet from the treatment plant (105 CFU/L). These findings strongly suggest that the presence of L. pneumophila in the river is due to the release of wastewater from the industrial aeration ponds, demonstrating that the river Glomma may be an additional disseminator of L. pneumophila during the outbreaks. This work emphasizes the need for preventive actions against the release of wastewater containing human pathogens to the environment.

Introduction Legionella pneumophila is the etiological agent of Legionnaires’ disease (LD) and of the nonpneumonic Pontiac fever. In addition to L. pneumophila, more than 49 different Legionella species have been described in which 19 species may cause infections in humans (1, 2). The species L. pneumophila contains at least 16 serogroups, in which serogroup 1 (SG1) is most frequently associated with legionellosis. Legionella spp. are ubiquitous in aqueous environments, but the incidence of the different clinical isolates of various Legionella species does not correlate with * Corresponding author phone: +47 63 80 78 27; fax: +47 63 80 75 09; e-mail: [email protected]. † Norwegian Defence Research Establishment (FFI). ‡ GPMS. § Unilabs Telelab AS. | Borregaard Ind. Ltd. 8712

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010

that found in the environment where L. pneumophila SG1 is not widely distributed (1, 3). It has been suggested that the predominance of L. pneumophila SG1 in humans is due to its virulence (4). Outbreaks of the Legionnaires’ disease caused by L. pneumophila SG1 have been traced back to various anthropogenic aerosol generating sources (5-7). In May 2005, an outbreak of LD caused by L. pneumophila SG1 occurred in the Sarpsborg/Fredrikstad region in Norway, and the source was identified as an air scrubber at the woodbased chemical factory Borregaard Ind. Ltd (8). During this outbreak, 56 people were infected over an area of approximately 1200 km2 and ten died (8). It has later been postulated that approximately additional 50 patients were infected by the 2005 outbreak strain (www.fhi.no). In November/December 2005, three new cases of LD occurred in the same region, caused by the same L. pneumophila strain as in the outbreak in May 2005; ST15 (9). This strain was identified to be identical to the L. pneumophila Lens strain responsible for the outbreak in Pas-de-Calais, France, in 2003/ 2004 (5). Five additional cases of LD were reported in Sarpsborg/Fredrikstad in June/July 2008, where the air scrubber, once again, was identified as being involved. All wastewater from Borregaard Ind. Ltd.’s wood refinement processing is biologically treated according to environmental requirements legislated by the Norwegian Environmental Protection Agency before it is released into the river Glomma nearby the wastewater treatment plant. Glomma is the largest river in Norway with a water flow of 200 to more than 2000 m3/s. Prior to the shut down in September 2008, the biological treatment facility at Borregaard Ind. Ltd. consisted of two large aeration ponds containing 30,000 m3 of liquid kept at approximately 37 °C for optimal growth of microorganisms to achieve efficient degradation of organic substances. Various Legionella species are present in the aeration ponds at concentration levels up to 1010 CFU/L (10, 11). Airborne Legionella spp. are disseminated up to 200 m downwind of the aeration ponds at Borregaard Ind. Ltd (11)., but it is speculated that infectious Legionella spp. may travel more than ten (5, 8) kilometers in air. Increased levels of IgG and IgM antibodies to L. pneumophila have been found in employees working proximal to the aeration ponds compared to those working more than 200 m away (12). Thus, there is no doubt that individuals in such working environments are exposed to L. pneumophila. Various molecular techniques have been used for phylogenetic and epidemiological studies of L. pneumophila strains (13). Sequence based typing (SBT) is, according to the European Working Group of Legionella Infections (EWGLI), recommended as the method of choice to generate reliable data for molecular typing of L. pneumophila (14, 15). However, multilocus variable number of tandem repeat analysis (MLVA) has been shown to provide high resolution, good reproducibility, phylogenetic information, low analytical costs, and time-consumption and is thus a favorable supplementary method to the SBT assay (16, 17). In this retrospective study we further investigated the source of dissemination for the LD outbreaks in Sarpsborg/ Fredrikstad, Norway, in 2005 and 2008 by analyzing the presence of L. pneumophila in samples harvested at various locations of Borregaard Ind. Ltd.’s biological treatment plant as well as from the nearby river Glomma. We propose that other disseminators than the air scrubbers at Borregaard Ind. Ltd. need to be evaluated as potential sources causing the outbreaks of LD. 10.1021/es1007774

 2010 American Chemical Society

Published on Web 10/15/2010

FIGURE 1. L. pneumophila SG1 sampled from the river Glomma in 2005 and 2008. Figures in CFU/L. Blue ) 2005, yellow ) 2008. Initial figures in boxes indicate location referred to in Table S2 (SI).

TABLE 1. Number of Samples Analyzed, Indicated by Location and Year of Samplinga year location

2005

scrubber1 scrubber3 cellulose fabric sedimentation pond Glomma (river) pond1 pond2 scrubberX

1/1

2006

2007

4/4 1/1 29/4 3/0 3/0 1/1

12/10 12/0 19/1

1/0 1/0

2008 1/1 2/2 3/3 1/1 12/10 40/38 41/29 1/1

a

Figures show total no. of samples vs no. of samples containing L. pneumophila SG1.

Materials and Methods Bacterial Growth. Environmental samples were collected from six locations at Borregaard Ind. Ltd. industrial area and from 15 sites in the river Glomma, Fredrikstad/Sarpsborg, Norway (Figure 1, Supporting Information (SI) Figure S1). The number of samples harvested at each location and time is presented in Table 1. Enrichment of Legionella spp. and L. pneumophila from the environmental samples was performed according to ISO11731 standard (SI Tables S1 and S2). L. pneumophila isolates were serotyped using Legionella Latex test DR 0800 and Dry Spot Legionella Latex test DR 200M, DR 210M, and DR 220M kit (Oxoid, England). Clinical L. pneumophila isolates (outbreak strains from 2005 and 2008) were obtained from The Norwegian Institute of Public Health. DNA Purification. Molecular analysis of the isolates identified as L. pneumophila SG1 was performed as follows;

total DNA was prepared from one single colony of 58 isolates of SG1 by boiling a loop of cells in 0.75 mL of sterile water for 10 min. Two microliters of the lysate was used as template in PCR. Sequence Based Typing. Each L. pneumophila SG1 isolate was genotyped using sequence based typing (SBT) as described by refs 14 and 15. Amplification was performed by using the LightCycler 480 System (Roche Diagnostics) and the LightCycler 480 SYBR Green I Master kit. PCR products for DNA sequencing were purified using the ExoSAP-IT kit (USB, USA). Nucleotide sequencing was performed in both directions by MWG Biotech Ltd., Germany. The resulting sequences were edited and aligned using the Staden package (18) and ClustalW (MEGA4 software package (19)), respectively. The SBT allele profile was established using the EWGLI bioinformatic recourses (www.ewgli.org). A phylogenetic comparison was made of L. pneumophila genotypes detected in this study to an extended strain collection of 182 L. pneumophila allele profiles from EWGLI (15, 20). The allele profiles were entered into BioNumerics version 5.1 software (Applied-Maths, Belgium) as character values, and a dendrogram was constructed using the categorical coefficient and the WARD algorithm. MLVA. MLVA-11 was performed as described previously (16, 17, see the SI). Survival of L. pneumophila in Air Scrubber Medium. The survival of L. pneumophila ST15 and ST462 at 45 °C, pH 8 (operating conditions of the air scrubbers during the 2005 outbreaks) were analyzed in air scrubber media from the air scrubber1 (AS1) and 3 (AS3). These liquid scrubber media were chosen as L. pneumophila ST15 and ST462 were detected in samples harvested from these scrubbers during the 2005 and 2008 outbreaks (SI Table S2). The pH of the VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8713

fresh scrubber media was adjusted from 3 to 8 using 6 M NaOH and homogenized prior to addition of bacteria. L. pneumophila cells cultivated on BCYE agar were resuspended in 2 mL of PBS and vortexed. Five hundred microliters of the cell suspension was added to 24.5 mL of AS1 and AS3 media (three parallels), obtaining a L. pneumophila concentration of 108-109 CFU/mL. Sterile water was used as control. The cell suspension was incubated in the scrubber media for 0 min, 30 min, 60 min, 180 min, 1 d, 2 d, and 3 d. One hundred microliters of the suspension was harvested and plated out on GVPC agar and incubated at 37 °C. CFU were enumerated after four and ten days. The survival of L. pneumophila in aeration pond liquid was studied under the same conditions as described above. The aeration pond liquid (harvested September 5, 2008) was used after two years of storage at 4 °C (the aeration ponds were nonoperative after September 2, 2008). However, L. pneumophila present in the pond liquid was able to survive during storage as the concentration was determined to 1.0 × 108 CFU/L compared to the original concentration of 6.0 × 108 CFU/L. Nineteen milliliters of scrubber media was preheated to 45 °C prior to addition of 1 mL of aeration pond liquid. Sterile water was used as control. The concentration of L. pneumophila in the mixed pond liquid/scrubber medium was determined to 1.5 × 103 CFU/mL. Further incubation, harvesting and culturing was performed as described above. mip-PCR analysis (21) was performed on the resulting colonies obtained on GVPC and BCYE (no growth at blood agar) to confirm the presence of L. pneumophila (data not shown). Also, survival of L. pneumophila in air scrubber medium (AS1 and 3) was analyzed at 45 °C, pH 3, resembling the ongoing operating conditions of the scrubbers and scrubber3 prior to and during the outbreak in 2008. Increased lignosulfonate concentrations and temperature (55 °C) were also tested as possible countermeasures against the survival of L. pneumophila in the scrubber (further details are presented in the SI).

Results Detection of L. pneumophila at Borregaard Ind. Ltd. L. pneumophila isolates harvested from the aeration ponds were typed to either SG1 and/or 2-14. In time periods between 2005 and 2008, when data were available, high concentration levels (106-1010 CFU/L) of L. pneumophila SG1 were detected in the aeration ponds (SI Table S1 and Figure S4). During June 2008-September 2008, L. pneumophila SG1 was present at 108-1010 CFU/L and constituted the majority of the Legionella pool in the aeration ponds (SI Table S1). L. pneumophila SG2-14 and other Legionella spp. were only sporadically detected in these periods. In other time periods, L. pneumophila SG2-14 and other Legionella spp. were periodically detected at concentrations levels of 105-1010 CFU/L (SI Table S1 and Figure S4). Non-L. pneumophila seemed to dominate the Legionella pool in the ponds during October 2005-March 2007 (no data available between February and August 2006 as no samples were harvested). Samples were harvested from the air scrubbers and other locations (except the aeration ponds) at Borregaard Ind. Ltd. weekly or every second week from 2005- (data not shown). These samples did not reveal any presence of L. pneumophila SG1, except in May 2005 and June 2008 (SI Table S2). Investigations of the 2008 outbreak (sampling performed June 25, 2008), identified the L. pneumophila ST462 strain in five samples harvested at Borregaard Ind. Ltd.: aeration ponds (two samples), scrubber1 (one sample), and scrubber3 (two samples) (SI Figure S1 and Table S2). L. pneumophila Detected in the River. After the 2005 outbreak (May-June), samples were harvested from the river Glomma at various locations up- and downstream of the 8714

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010

outlet from Borregaard’s wastewater treatment facility, and the Legionella concentration levels were determined (Figure 1). Results showed that the highest concentration level of L. pneumophila SG2-14 was detected at the outlet (1.9 × 106 CFU/L), while the lowest level (4.0 × 104 CFU/L) was identified about 1.6 km downstream of the outlet (data not shown). SG1 isolates were identified in four out of 16 river samples, all located >300 m downstream of the plant’s outlet (Figure 1). L. pneumophila was not detected upstream of the outlet. An additional one-day study (August 8, 2005) detected only L. pneumophila SG2-14 in Glomma at the outlet (3.0 × 105 CFU/L) and at two other locations at low concentrations (data not shown). Analysis of L. pneumophila from river samples was also performed in August 2008 as five additional cases of LD were reported in June/July 2008. The highest concentration level of L. pneumophila SG1 was detected by the outlet of the biological treatment plant (2.1 × 105 CFU/L). L. pneumophila SG1 was not detected at about 18.5 km downstream of the outlet where the river flow is mixed with seawater in the Fredrikstad estuary (Figure 1). These results demonstrated that the river Glomma was contaminated with L. pneumophila SG1 originating from Borregaard’s biological treatment plant in 2005 and 2008. In addition, one sample positive for L. pneumophila ST462 (September 19, 2008) was identified in an air scrubber located at an industrial plant about 12 km downstream of Borregaard Ind. Ltd. (SI Table S2). Sequence Based Typing. The SBT profile was determined from a total of 58 environmental and clinical isolates of L. pneumophila SG1 sampled from various locations at Borregaard Ind. Ltd. and Glomma (SI Table S2). In total, three different genotypes were identified: ST15 (12,9,26,5,26,17,15) (outbreak strain 2005), ST462 (12,9,2,5,50,17,15) (outbreak strain 2008), and ST458 (24,6,3,3,13,11,11). The six L. pneumophila SG1 isolates sampled mainly from Glomma (one sample from the scrubber) in 2005 and 2006 were identified as ST15 (SI Table S2). This is in agreement with Nygård et al. (8), who proved that one sample from both Glomma and the air scrubber were positive for the ST15 strain. Interestingly, no L. pneumophila SG1 was identified in the aeration ponds in May 2005, but river samples harvested in May/June 2005 and August 2008 were positive for both the ST15 and ST462 strains, respectively (SI Table S2). The ST458 strain was only detected in a drying machine of the cellulose plant, during April 2007- August 2008 (SI Figure S1 and Table S2). Phylogenetic Analysis. The genetic relationship of the three different L. pneumophila genotypes identified at Borregaard Ind. Ltd. was compared to a broader collection of L. pneumophila isolates; a dendrogram of the SBT profiles was established based on data from 185 different L. pneumophila isolates (SI Figure S2). The 2005 and 2008 outbreak strains clustered along with 28 other strains close to the L. pneumophila Lens strain (4). This cluster is herein described as the Lens-cluster. All strains within the Lens-cluster belong to SG1, except for the SG3 EULV1536 and EULV1539 Canadian clinical isolates. MLVA. As the 2008 outbreak strain ST462 showed an identical SBT profile to the L. pneumophila SG3 strain EULV1536, concern was raised to the methods resolution in order to identify a link between the corresponding clinical and environmental 2005 and 2008 outbreak strains. A high resolution MLVA assay was used (16, 17) to support the SBT data. All SBT profiles were confirmed by the MLVA-11 assay, except for one isolate (SI Table S2) which was sampled from the aeration pond January 5, 2006 showing a different Lpms37 allele compared to all other ST15 isolates studied in this work. TheMLVA-11profileoftheEUL1536strain(8,7,9,4,2,4,13.5,2,1,23,11) was different from the 2008 outbreak strain in locus Lpms4, Lpms31, and Lpms37 (underlined in the allele profile above).

FIGURE 2. Survival of L. pneumophila in two various air scrubber media at 45 °C, pH 8 in a) AS1 and b) AS3. Bars indicate standard deviation of three parallels. Cont. ) control. Survival of L. pneumophila in Air Scrubber Medium. The survival of L. pneumophila ST15 and ST462 strains was studied in two air scrubber media (AS1 and 3) as the air scrubbers may be effective disseminators of L. pneumophila. The conditions for analysis (45 °C, pH 8) were chosen according to the operating conditions of the scrubbers during the outbreak in 2005 and 2008 (SI Table S3). The hypothesis is that if L. pneumophila is not able to multiply and/or survive in the harsh conditions provided by the scrubbers, it is less likely that such devices running at these operating conditions were the main disseminators of L. pneumophila-containing aerosols during the outbreaks. Survival of L. pneumophila ST15 and ST462 was not observed after one day and three hour incubation, respectively, at 45 °C, pH 8, in AS1 media (Figure 2). In AS3 media, survival of L. pneumophila ST15 and ST462 was not observed after two and one day incubation, respectively. These data show that the air scrubber media have an impact on the survival of the outbreak strains. Also, L. pneumophila from the aeration ponds and strain ST462 were not able to survive after a six hour incubation in AS1 and AS3 medium at 45 °C, pH 3 (prevailing conditions in air scrubber3 during the 2008 outbreak) (SI Figure S3) supporting our finding that the conditions provided by the air scrubbers do not promote survival of L. pneumophila. The survival of L. pneumophila, originally present in aeration pond liquid, was studied under similar conditions as described above. This analysis was performed to assess whether the survival of L. pneumophila in scrubber media could be enhanced, compared to L. pneumophila cells grown in laboratory, due to a more environmental and maybe stressed growth state of the L. pneumophila cells provided by the aeration pond liquid. L. pneumophila was not able to survive after one and three hour incubation in AS1 and AS3 media, respectively (Figure 2).

Discussion Previous studies identified the air scrubber at Borregaard Ind. Ltd. as the source of dissemination causing LD outbreaks in the Sarpsborg/Fredrikstad region, Norway, in 2005 based on i) the clinical ST15 outbreak strain was identified in the air scrubber1, ii) the scrubber was regarded as an effective aerosol generator, and iii) favorable growth conditions for Legionella in the air scrubbers were assumed (8). However, only two samples were harvested from Borregaard Ind. Ltd. during the 2005 investigation; aeration pond and air scrubber liquid medium, in which only the latter was positive for L.

pneumophila ST15 (8) at concentration level of 2000 CFU/L. The analysis of these low number of samples collected in 2005, together with our data showing high concentration levels of L. pneumophila ST15 (>105 CFU/L) in the river Glomma (May 2005) downstream of the outlet from the biological treatment plant (Figure 1), strongly indicate that the aeration ponds were considerably contaminated with L. pneumophila ST15 during the 2005 outbreak. This is further supported by the presence of high concentration levels (>1010 CFU/L) of ST462 in the aeration ponds in 2008 and the resulting ST462 concentration gradient in Glomma from Borregaard’s wastewater outlet to the Fredrikstad estuary (Figure 1). We have previously identified that Legionellacontaining aerosols are able to be transported at least up to 200 m downwind of the aeration ponds (11), which is equivalent to the distance between the air scrubber1 and the aeration ponds. The aerosol generation potential from the aeration ponds should be considered as reasonable calculations suggest that the rate of aerosol generation per areal unit from the aeration ponds at Borregaard Ind. Ltd. is equivalent to aerosol generation rates for whirlpool spa (ref 22, SI). Further, our institute is currently analyzing whether the aerosol dispersion pathways from the aeration ponds and the air scrubber at Borregaard Ind. Ltd. are comparable at far distances, i.e. > 1 km. As the aerosol plume concentrations from the aeration ponds (and the L. pneumophila concentration in the ponds (this study)) are larger than that from the air scrubbers, but indistinguishable at far distances (29), we find it reasonable to suggest that the aeration ponds had a significant potential to disperse L. pneumophila containing aerosols. Therefore, we suggest that the aeration ponds have functioned as amplifiers and disseminators of Legionella causing the LD outbreaks in both 2005 and 2008. The L. pneumophila ST15 strain has previously been identified in an aeration pond during a LD outbreak in Pasde-Calais, France, 2001, in which ST15 was identified in air up to 300 m downstream of the ponds (5). In this outbreak a nearby cooling tower was identified as the dissemination source of L. pneumophila. Thus, the aeration pond at Pasde-Calais may also have functioned as a local amplifier of the outbreak strain. However, it cannot be ruled out that L. pneumophila containing aerosols generated from the aeration ponds at Borregaard Ind. Ltd. and Pas-de-Calais might have been able to contaminate the air scrubber and the cooling tower, respectively, through the air inlet, as these devices are located in short distances from the aeration ponds. However, in contrast to air scrubbers (see below), cooling towers are regarded to possess both an amplifying and a dissemination potential of L. pneumophila containing aerosols (5). Due to environmental regulations (25, 26), the air scrubbers at Borregaard Ind. Ltd. were subject to routine investigations for L. pneumophila from 2005 (still ongoing). No growth of L. pneumophila was identified in any of these control samples (except during the outbreaks in 2005 and 2008) and neither for those harvested from the inside of the scrubbers. Also, our study showed that the outbreak strains were not able to survive after two days incubation in AS1 and AS3 media (Figure 2). Thus, we strongly believe that the harsh operating conditions in the scrubber do not promote amplification of L. pneumophila. Surprisingly, the survival of L. pneumophila originally present in the aeration pond liquid was lower compared to plate grown ST15 and ST462 (nor the control sample did show any survival after one day) (Figure 2). This may be due to the long-term storage of the pond liquid, in which L. pneumophila cells have been adapted to the storage temperature (4 °C), and that a rise in temperature will inhibit survival of L. pneumophila. Also, the initial concentration level of L. pneumophila used was different in these experiments; ∼103 CFU/ml L. pneumophila in the pond liquid vs ∼108 CFU/mL for plate grown bacteria. VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8715

In addition, experiments performed at operational conditions from scrubber3 during the outbreak in 2008 (45 °C, pH 3) did not show any survival of L. pneumophila after six hours incubation in AS1 and AS3 (SI Figure S3). These data suggest that the air scrubbers at Borregaard Ind. Ltd. provide unfavorable growth conditions for L. pneumophila and propose that the scrubbers should not to be regarded as the primary disseminator source of the outbreak strains in 2005 and 2008. If the contradictory appears to be realistic, positive detection of L. pneumophila should have been observed, at least occasionally, during routine investigations (inside) of the scrubbers. Potential transfer of L. pneumophila containing aerosols from the aeration ponds via scrubber1 to ambient air may seem reasonable as these two devices are located only 200 m in distance from each other (11). However, as outdoor air entering the scrubber is mixed with process air from the lignin plant, holding a temperature of 80-90 °C, the concentration level of airborne L. pneumophila originating from the scrubber will be significantly reduced. The D-value (decimal reduction time) for L. pneumophila is 15 s at 80 °C (23). Despite even a few second retention time in the scrubber (high air flow rate), it is likely to assume that a considerable direct transport of viable L. pneumophila cells through the scrubber is not to occur. It may be speculated that Legionella is able to contaminate the scrubbers through the scrubbers’ water inlet and that bacterial cells may stick to the surfaces of the scrubbers and/ or be protected by amoeba and biofilm (24). This may explain why at least four samples harvested from the air scrubbers (2005-2010) were positive for L. pneumophila SG1 (SI Table S2). The water used in Borregaard’s air scrubbers originates from Glomma, upstream of Borregaard’s wastewater outlet where L. pneumophila SG1 was not detected. Still, our study shows that L. pneumophila is not able to multiply in the air scrubbers at the operating (imitating) conditions during the outbreak in 2005 and 2008. All together, these arguments support our hypothesis that a L. pneumophila contamination established in the air scrubbers, of any reasons, may retain at low concentrations for a limited time period and then diminish after a short time period (1-2 days) as demonstrated by the survival experiments in this study. Unfortunately, data from environmental investigations are limited due to the low number of samples harvested (Table 1). However, the biased number of samples when comparing data from 2005 and 2008 may call upon alternative explanations for the difference in the number of LD cases reported in 2005 (56 cases) compared to that reported in 2008 (five cases). The high number of LD cases in 2005 may be due to the less stringent decontaminations procedures of the air scrubbers prior to the 2005 outbreak compared to the operating and decontamination procedures in 2005-2008. Enhanced national legislations regarding use of aerosol generating equipment were established by law after the 2005 outbreak and, thus, also implemented at Borregaard Ind. Ltd. (25, 26). Borregaard Ind. Ltd. improved their disinfection routines of the scrubbers; monthly use of steam (hot-water) (2006-2008) and later weekly washing with hypochlorite solution (Dec 2008-). However, the incidence of LD may correlate with an array of various factors, e.g. weather conditions, source of dispersion (aerosol generation rate, concentration of pathogen), host exposure time, host susceptibility, and strain variations (e.g., virulence, environmental stability). For instance, the survival experiment indicated that ST15 had an enhanced survival rate compared to ST462 in both AS1 and 3 (Figure 2). It is therefore challenging to describe when and to what extent a LD outbreak may take place, as illustrated by the 2008 outbreak, which did not continue for a more extensive time period (ended in July) even though high concentration levels of 8716

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010

ST462 were detected in the aeration ponds (108-109 CFU/L) during July-September 2008. Several industries are located along the river Glomma, and some of these use untreated river water for industrial processes, suggesting that other sources, including air scrubbers/cooling towers/the Glomma river itself, located outside Borregaard Ind. Ltd. may take part in dispersing L. pneumophila containing aerosols. This is the first report on addressing the river Glomma as a potential source for L. pneumophila dissemination. The majority of the LD cases in 2005 were reported by the population residing along the river establishing two main clusters of cases (8). On September 17, 2008, ST462 was detected in an air scrubber at an industrial plant and in the river (sampled September 3) downstream of Borregaard Ind. Ltd. after the wastewater effluent from Borregaard’ aeration ponds was shut down. It cannot be ruled out that devices using untreated water contaminated by L. pneumophila are able to contribute to the aerosol dispersion of L. pneumophila. L. pneumophila cells originating from the contaminated river, and which has not been in contact with air scrubber medium, may be able to survive and be expelled from the scrubber as aerosols immediately after passing the shower nozzles which are used to clean the air before release. Such low level supplement of L. pneumophila containing aerosols to a community may increase the rates of LD (27). The MLVA-11 assay confirmed the SBT analysis of L. pneumophila (SI Table S2). Interestingly, ST462 and EULV1536 demonstrated identical genotype despite that these isolates possess different serotypes and are found at two locations in far distance from each other. These results confirm that there is not a stringent relation between the genotype and serotype of L. pneumophila consistent with previous studies (20). Cross reactions have occurred during serotyping of SG1, 3, and 6 (28), but this seems not to be the case in this study as typing of the serogroups were confirmed (personal communication, C. Guyard). This suggests that several molecular typing methods, in addition to SBT, should be complementary used for epidemiological typing. Our data showed that single infectious strains of L. pneumophila SG1 was most probably established in the aeration ponds over extended time periods (2005-2008) and that there seems to be a switch of the dominant ST15 strain to ST462 in these ponds during the time period studied. No indications of coexistence were detected but cannot be ruled out. Both genotypes are related to the Lens group (SI Figure S2). It seems as the L. pneumophila Lens group strains are able to survive and efficiently grow in such environments as no other SBT profiles were detected throughout May 2005December 2008. Our work emphasizes the need for preventive actions against the release of wastewater containing human pathogens to the environment and encourages thorough investigations in order to identify environmental sources in epidemiological studies.

Acknowledgments We would like to thank Dr. Ingeborg S. Aaberge at the Norwegian Institute of Public Health for providing L. pneumophila strains from the air scrubbers at Borregaard Ind. Ltd. and the clinical isolates from the outbreaks in 2005 and 2008. We are also grateful to Dr. Cyril Guyard for forthcoming information and provision of L. pneumophila EUL1536 DNA. Thanks also to health authorities of the city of Fredrikstad by Reidun Ottosen and Steinar Haugsten for making a valuable isolate of L. pneumophila SG1 from Fredrikstad accessible for this study and Prof. Per Einar Granum and Dr. Øyvind Andreassen for valuable discussions.

Supporting Information Available Theoretical calculations of the aerosol generation rate from the aeration ponds at Borregaard Ind. Ltd., description of

the MLVA method, supplementary survival experiments, sampling sites, growth data in aeration ponds, SBT and MLVA data, temperature and pH in air scrubbers, and phylogenetic comparison of the L. pneumophila outbreak strains. This material is available free of charge via the Internet at http:// pubs.acs.org.

(14)

(15)

Literature Cited (1) Edelstein, P. H. Clinical features of Legionnaires’ disease: A selective review. In Legionella: State of the art 30 years after its recognition; Cianciotto, N. P., Kwaik, Y. A., Edelstein, P. H., Fields, B. S., Geary, D. F., Harrison, T. G., Joseph, C. A., Ratcliff, R. M., Stout, J. E., Swanson, M. S., Eds.; ASM Press: Washington, DC, 2006. (2) Dideren, B. M. W. Legionella spp. and Legionnaires’ disease. J. Infect. 2008, 56, 1–12. (3) Muder, R. R.; Yu, V. L. Infection due to Legionella species other than L. pneumophila. Emerg. Inf. 2002, 35, 990–998. (4) Cazalet, C.; Rusniok, C.; Bru ¨ ggemann, H.; Zidane, N.; Magnier, A.; Ma, L.; Tichit, M.; Jarraud, S.; Bouchier, C.; Vandenesch, F.; Kunst, F.; Etienne, J.; Glaser, P.; Buchrieser, C. Evidence in the Legionella pneumophila genome for exploitation of host cell functions and high genome plasticity. Nat. Genet. 2004, 36, 1165– 1173. (5) Nguyen, T. M. N.; Ilef, D.; Jarraud, S.; Rouil, L.; Campese, C.; Che, D.; Haeghebaert, S.; Ganiayre, F.; Marcel, F.; Etienne, J.; Desenclos, J. A community-wide outbreak of Legionnaires disease linked to industrial cooling towers - How far can contaminated aerosols spread. J. Infect. Dis. 2006, 193, 102– 111. (6) Palmore, T. N.; Stock, F.; White, M.; Bordner, M.; Michelin, A.; Bennett, J. E.; Murray, P. R.; Henderson, D. K. A cluster of cases of nosocomial legionnaires disease linked to a contaminated hospital decorative water fountain. Infect. Control. Hosp. Epidemiol. 2009, 30, 764–768. (7) Jernigan, D. B.; Hofmann, J.; Cetron, M. S.; Genese, C. A.; Nuorti, J. P.; Fields, B. S.; Benson, R. F.; Carter, R. J.; Edelstein, P. H.; Guerrero, I. C.; Paul, S. M.; Lipman, H. B.; Breiman, R. Outbreak of Legionnaires’ disease among cruise ship passengers exposed to a contaminated whirlpool spa. Lancet 1996, 347, 494–499. (8) Nygård, K.; Werner-Johansen, Ø.; Rønsen, S.; Caugant, D. A.; Simonsen, Ø.; Kanestrøm, A.; Ask, E.; Ringstad, J.; Ødegård, R.; Jensen, T.; Krogh, T.; Høiby, E. A.; Ragnildstveit, E.; Aaberge, I. S.; Aavidtsland, P. An outbreak of Legionnaires’ disease caused by long-distance spread from an air scrubber in Sarpsborg, Norway. Clin. Infect. Dis. 2008, 46, 61–69. (9) Folkehelseinstituttet. Utbrudd av legionellose i Østfold junijuli 2008. (Outbreak of legionellosis in Østfold. June-July 2008.) Available at http://www.fhi.no/dav/ce069fc583.pdf. (10) Allestam, G.; Långmark, J. Legionella i bioreningsanlaggningar, Kartlaggning och riskbedømning 2005-2007 (Legionella in biological treatment plants; Survey and risk assessment 20052007); Swedish Institute for Infectious Disease Control: 2007. (11) Blatny, J. M.; Reif, B. A. P.; Skogan, G.; Andreassen, Ø.; Høiby, E. A.; Ask, E.; Waagen, V.; Aanonsen, D.; Aaberge, I. S.; Caugant, D. A. Tracking Airborne Legionella spp. and Legionella pneumophila at a biological treatment plant. Environ. Sci. Technol. 2008, 42, 7360–73687. (12) Wedege, E.; Bergdal, T.; Bolstad, K.; Caugant, D. A.; Efskind, J.; Heier, H. E.; Kanestrøm, A.; Strand, B. H.; Aaberge, I. S. Seroepidemiological study after a long-distance industrial outbreak of Legionnnaires’ disease. Clin. Vac. Immunol. 2009, 16, 528–534. (13) Scaturro, M.; Losardo, M.; De Ponte, G.; Ricci, M. L. Comparison of three molecular methods used for subtyping of Legionella pneumophila strains isolated during an epidemic of Legion-

(16)

(17)

(18) (19) (20)

(21) (22)

(23)

(24) (25) (26)

(27) (28)

(29)

ellosis in Rome. J. Clin. Microbiol. 2005, 43, 5348–5350 (inn I stedet for original refs 14-16 ). Gaia, V.; Fry, N. K.; Afshar, B.; Luck, P. C.; Meugnier, H.; Etienne, J.; Peduzzi, R.; Harrison, T. G. Consensus sequence-based scheme for epidemiological typing of clinical and environmental isolates of Legionella pneumophila. J. Clin. Microbiol. 2005, 43, 2047–2052. Ratzow, S.; Gaia, V.; Helbig, J. H.; Fry, N. K.; Lu ¨ ck, P. C. Addition of neuA, the gene encoding N-acylneuraminate cytidylyl transferase, increases the discriminatory ability of the consensus sequence-based scheme for typing Legionella pneumophila serogroup 1 strains. J. Clin. Microbiol. 2007, 45, 1965–1968. Pourcel, C.; Visca, P.; Afshar, B.; D’Arezzo, S.; Vergnaud, G.; Fry, N. K. Identification of variable-number tandem-repeat (VNTR) sequences in Legionella pneumophila and development of an optimized multiple-locus VNTR analysis typing scheme. J. Clin. Microbiol. 2007, 45, 1190–1199. Nederbragt, A. J.; Balasingham, A.; Sirevåg, R.; Utkilen, H.; Jakobsen, K. S.; Anderson-Glenna, M. J. Multiple-locus variablenumber tandem repeat analysis of Legionella pneumophila using multi-colored capillary electrophoresis. J. Microbiol. Methods 2008, 73, 111–117. Staden, R. The Staden sequence analysis package. Mol. Biotechnol. 1996, 5, 233–241. Tamura, K.; Dudley, J.; Nei, M.; Kumar, S. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol. Biol. Evol. 2007, 24, 1596–1599. Borchardt, J.; Helbig, J. H.; Lu ¨ ck, P. C. Occurrence and distribution of sequence types among Legionella pneumophila strains isolated from patients in Germany: common features and differences to other regions of the world. Eur. J. Clin. Microbiol. Infect. Dis. 2008, 27, 29–36. Wellinghausen, N.; Frost, C.; Marre, R. Detection of legionellae in hospital water samples by quantitative real-time LightCycler PCR. Appl. Environ. Microbiol. 2001, 67, 3985–3993. Armstrong, T. W.; Haas, C. N. Quantitative microbial risk assessment model for Legionnaires’ disease: assessment of human exposures for selected spa outbreaks. J. Occup. Environ. Hyg. 2007, 4, 634–646. Stout, J. E.; Best, M. G.; Yu, V. L. Susceptibility of members of the family Legionellaceae to thermal stress: implications for heat eradication methods in water distribution systems. Appl. Environ. Microbiol. 1986, 52, 396–399. Declerck, P. Biofilms: the environmental playground of Legionella pneumophila. Environ. Microbiol. 2009, 12. [Epub ahead of print]. Lee, J. V. New Norwegian legionella legislation is consistent with the European guidelines. Euro. Surveill. 2005, 10, E050714.3. Social and Health Directorate. Midlertidig forskrift om tiltak for å hindre overføring av Legionella via aerosol. Sosial- og helsedirektoratet, 2005. (Preliminary regulations for preventive measurements of transmission of Legionella-containing aerosols. July 12, 2005.) Available at http://www.shdir.no/vp/ multimedia/archive/00003/Midlertidig_forskrift_3362a.pdf [in Norwegian]. Armstrong, T. W.; Haas, C. N. Legionnaires’ disease: evaluation of a quantitative microbial risk assessment model. J. Water Health 2008, 6, 149–166. Ditommaso, S.; Giacomuzzi, M.; Gentile, M.; Zotti, C. M. Antibody detection and cross-reactivity among species and serogroups of Legionella by indirect immunofluorescence test. J. Microbiol. Methods 2008, 75, 350–353. Tutkun, M.; Andreassen, Ø.; Reif, B. A. Dispersion of Aerosols from Two Different Sources at a Biological Treatment Plant in Norway. Boundary-Layer Meteorology, submitted.

ES1007774

VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8717