- Email: [email protected]
Control of Legionella contamination in a hospital water distribution system by monochloramine
American Journal of Infection Control 40 (2012) 279-81
Contents lists available at ScienceDirect
American Journal of Infection Control
American Journal of Infection Control
journal homepage: www.ajicjournal.org
Brief report
Control of Legionella contamination in a hospital water distribution system by monochloramine Isabella Marchesi PhD a, Stefano Cencetti MD b, Patrizia Marchegiano MD b, Giuseppina Frezza PhD a, Paola Borella MD a, Annalisa Bargellini PhD a, * a b
Department of Public Health Sciences, University of Modena and Reggio Emilia, Modena, Italy University Hospital Policlinico of Modena, Modena, Italy
Key Words: Legionella Pseudomonas Monochloramine Chlorine dioxide Hospital Hot water distribution system
Background: We report the results of 1-year application of monochloramine to control Legionella pneumophila contamination in a hospital hot water distribution system. Methods: In the main building of the hospital, a device continuously distributing monochloramine was installed. Legionella pneumophila and Pseudomonas spp contamination was followed in comparison with 2 other water networks in the same building using chlorine dioxide. Results: Monochloramine significantly reduced the number of contaminated sites compared with baseline (from 97.0% to 13.3%, respectively), chlorine dioxide device I (from 100% to 56.7%, respectively), and device II (from 100% to 60.8%, respectively). No positive sample exceeded 104 colony-forming units/L versus 59.4% at baseline. Conclusion: Monochloramine could represent a good alternative to chlorine dioxide in controlling legionellae contamination in public and private buildings. Copyright Ó 2012 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved.
The control of Legionella spp contamination in plumbing systems of large buildings may be a difficult challenge. This is of particular concern in hospital settings where contamination of hot water distribution system actually is the most important risk factor for nosocomial Legionella pneumonia.1,2 National and international guidelines aimed at preventing Legionella infection suggest the use of biocides for the routine treatment of contaminated hot water.3,4 Methods adopted include superheating, ultraviolet light, coppersilver ionization, ozone, hyperchlorination, chlorine dioxide, and point-of-use water filters. To date, the best procedure is not established, but the effectiveness of different systems in terms of reduction of contamination and related costs has been recently evaluated.5 Monochloramine is considered more effective than free chlorine in controlling Legionella contamination of longer distribution systems for its ability to better penetrate biofilms.6 Indeed, when adopted for disinfection of municipal water supplies in the United States, it was associated with a lower prevalence of
* Address correspondence to Annalisa Bargellini, PhD, Department of Public Health Sciences, University of Modena and Reggio Emilia, Via Campi 287, 41125 Modena, Italy. E-mail address: [email protected] (A. Bargellini). Supported by hospital funds. Conflicts of interest: None to report.
Legionella colonization in potable water systems of public and private buildings6,7 and decreased risk of nosocomial outbreaks of Legionnaires’ disease.8,9 In this study, we present the preliminary results of 1-year continuous injection of monochloramine in a hot water distribution system of a large hospital. A comparison with chlorine dioxide applied in 2 other water distribution systems of the same building is also presented. METHODS The experiment was conducted in a 9-story hospital building where 3 water networks distribute in parallel hot water. A device aimed to continuously produce and dispense monochloramine was installed from March 2009 on 1 of the 3 networks of this building. The monochloramine device (Sanipur, S.r.l., Brescia, Italy) was specifically developed to keep the level of monochloramine between 1.5 and 3.0 mg/L in the recirculation system, minimizing production of ammonia and other by-products. In the other 2 water networks of the same building, continuous chlorine dioxide devices furnished by different producers (device I: Sanipur S.r.l., Brescia, Italy, and device II: Culligan Italiana S.p.A., Bologna, Italy) were installed, assuring approximately 0.3 mg/L at distal outlets.5 Hot water samples were collected at regular intervals from the same sites to evaluate Legionella spp and Pseudomonas spp
0196-6553/$36.00 - Copyright Ó 2012 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.ajic.2011.03.008
10/41 (24.4)x 29 (1-46,000) 7/11 (63.6) 170 (10-850,000) 13/28 (46.4) 88x (1-13,000) 6/8 (75.0) 3332 (150-118,400) 13/60 (21.7) 169 (10-2,000) cfu, colony-forming units. *P < .001. y P < .002. z Only water positive. x P < .05.
6/19 (31.5) 70 (3-930)
15/15 (100) 10,752 (120-280,000) 9/15 (60.0) 34/60 (56.7)y 435* (25-120,000) 4/60 (6.7)* 14/14 (100) 11,714 (1,020-248,200) 7/14 (50.0)
After Before Before
After
Chlorine dioxide: device II
After
8/60 (13.3)* 335* (25-4,950) 0/60 (0)* 31/32 (97.0) 17,242 (100-946,000) 19/32 (59.4)
One-year application of monochloramine in a hot water distribution system heavily contaminated by L pneumophila gave satisfactory results both per se and compared with chlorine dioxide. Only 8 out of 60 samples were positive, and no sample exceeded 104 cfu/L versus 59.4% at baseline. In addition, after the first month, the positive samples were in total 4 of 48 (8.3%), all below 103 cfu/L. In previous studies, the percentage of water sites testing positive (30%) has been found to be a better predictor of hospital-acquired legionellosis than is the quantitative concentration,11 thus suggesting that monochloramine may be useful to reduce the risk of case appearance since the beginning of its application. The monochloramine level associated with bacteria below the detection limit approximates 3 mg/L, but 2 mg/L was sufficient for reducing legionellae below 100 cfu/L. Positive effects were not documented against Pseudomonas spp contamination, but this can depend on a limited pretreatment in the water network where monochloramine was applied. The significant reduction noticed by using chlorine dioxide suggests a higher capability to counteract
Before
DISCUSSION
Chlorine dioxide: device I
Table 1 shows Legionella pneumophila and Pseudomonas spp contamination in the water distribution systems treated by either monochloramine or chlorine dioxide. In the examined hospital building, only Legionella pneumophila was isolated. Before any treatment, L pneumophila concentration did not differ among the 3 distribution networks. Compared with chlorine dioxide, monochloramine exhibited a higher reduction in the number of both contaminated sites and sites exceeding 104 colony-forming units (cfu)/L. Chlorine dioxide significantly decreased number and/or concentration of Pseudomonas spp, whereas monochloramine had no significant effect. Figure 1 shows the trend of L pneumophila contamination with time by monochloramine. Taking the pretreatment samples all together as reference, the contamination risk was strongly reduced by monochloramine (odds ratio [OR], 0.003; confidence interval [CI]: 0.001-0.021, P < .001), followed by chlorine dioxide device I (OR, 0.022; CI: 0.003-0.168, P < .05) and device II (OR, 0.026; CI: 0.003-0.202, P < .05). A negative correlation was observed between residual monochloramine and L pneumophila concentration (r ¼ 0.341, P < .01, on total samples; r ¼ 725, P < .05, on positive samples). The regression line parameters revealed that 2 mg/L maintained L pneumophila below 100 cfu/L and 3 mg/L below the detection limit (25 cfu/L). For chlorine dioxide, taking the 2 systems together, the coefficients of correlation were r ¼ 0.606 on all samples and r ¼ 0.610 on positive samples (P < .001). Residual chlorine dioxide concentrations of 0.3 and 0.6 mg/L were requested at distal sites to obtain <100 and <25 cfu/L legionellae, respectively. Before any treatment, 5 of 60 positive samples (8.4%) were colonized by L pneumophila serogroup 1. The trend did not change after application of monochloramine (1/8 positive samples, 12.5%), whereas the percentage significantly increased after application of chlorine dioxide: 35 of 65 (53.8%), c2 ¼ 30.21, P < .001. No variation in both nitrite and nitrate water concentration was observed during the entire study, and no sample exceeded the limit of 0.50 mg/L and 50 mg/L, respectively, requested for potable water.
Table 1 Legionella pneumophila and Pseudomonas spp contamination in 3 hot water distribution systems of the same building before and after application of continuous disinfecting procedures
RESULTS
Legionella pneumophila positive, n (%) Geometric mean cfu/L (range)z 104 cfu/L, n (%) Pseudomonas spp positive, n (%) Geometric mean cfu/100mL (range)z
colonization using standardized cultural procedures and to measure residual disinfectant concentration.10 Water samples were taken from heater, return loop, and distal outlets (shower and/or tap selected on the basis of distance and exposure risk), without flaming and after flushing for 1 minute. Only viable planktonic bacteria were enumerated.
31/51 (60.8)y 1297* (50-22,500) 5/51 (9.8)*
I. Marchesi et al. / American Journal of Infection Control 40 (2012) 279-81
Monochloramine
280
I. Marchesi et al. / American Journal of Infection Control 40 (2012) 279-81
281
L. pneumophila Log cfu/L
5 Monochloramine 4
3
return loop
heat exchanger
tap A shower C
tap B
2
1
mt h 15
mt h 14
mt h 12
mt h 10
8m th
6m th
4m th
3m th
1m th
1w ee k
be for
e
0
Time Fig 1. Trend of L pneumophila contamination before and after monochloramine continuous injection.
Pseudomonas spp but could also depend on their higher number and/or concentration before treatment. The presence of L pneumophila serogroup 1 was not influenced by monochloramine, whereas chlorine dioxide increased the number of points colonized by this serogroup, which is more frequently associated with the disease. This is in line with the decreased number of Legionnaires’ disease cases observed in the United States by using monochloramine for residual disinfection.8,9 One possible limitation of our study is the absence of a control building, but the Italian guidelines require intervention when contamination exceeds 104 cfu/L.3 The hospital has other 4 buildings: 2 contaminated and 2 not. In those contaminated, shock treatments are periodically adopted, insufficient to control contamination as documented in our previous study5; but, at present, 1 building is closed, and the other actually avoids patient exposure to hot water. Our results suggest that continuous injection of monochloramine in hot water distribution systems can control L pneumophila and Pseudomonas spp in contaminated buildings without modifying water content in nitrite and nitrate. Investigations are in progress to evaluate the production of disinfection by-products and the long-term effect of monochloramine also on biofilm development and amoeba presence associated with Legionella. Acknowledgment The author thanks Sanipur S.r.l. for technical support and equipment provided.
References 1. Yu VL, Stout JE. Hospital characteristics associated with colonization of water systems by Legionella and risk of nosocomial Legionnaires’ disease: a cohort study of 15 hospitals. Infect Control Hosp Epidemiol 2000;21: 434-5. 2. Sabria M, Yu VL. Hospital-acquired legionellosis: solutions for a preventable infection. Lancet Infect Dis 2002;2:368-73. 3. Italian guidelines for prevention and control of legionellosis G.U. Gazzetta Ufficiale della Repubblica Italiana, serie generale, n.103. 2000. 4. World Health Organization. Legionella and the prevention of legionellosis. Geneva: World Health Organization; 2007. 5. Marchesi I, Marchegiano P, Bargellini A, Cencetti S, Frezza G, Miselli M, et al. Effectiveness of different methods to control legionella in the water supply: ten-year experience in an Italian university hospital. J Hosp Infect 2011;77: 47-51. 6. Moore MR, Pryor M, Fields B, Lucas C, Phelan M, Besser RE. Introduction of monochloramine into a municipal water system: impact on colonization of buildings by Legionella spp. Appl Environ Microbiol 2006;72:378-83. 7. Flannery B, Gelling LB, Vugia DJ, Weintraub JM, Salerno JJ, Conroy MJ, et al. Reducing Legionella colonization in water systems with monochloramine. Emerg Infect Dis 2006;12:588-96. 8. Kool JL, Carpenter JC, Fields BS. Effect of monochloramine disinfection of municipal drinking water on risk of nosocomial Legionnaires’ disease. Lancet 1999;353:272-7. 9. Heffelfinger JD, Kool JL, Fridkin S, Fraser VJ, Hageman J, Carpenter J, et al. Risk of hospital-acquired Legionnaires’ disease in cities using monochloramine versus other water disinfectants. Infect Control Hosp Epidemiol 2003;24: 569-74. 10. Borella P, Montagna MT, Stampi S, Stancanelli G, Romano-Spica V, Triassi M, et al. Legionella contamination in hot water of Italian hotels. Appl Environ Microbiol 2005;71:5805-13. 11. Stout JE, Muder RR, Mietzner S, Wagener MM, Perri MB, DeRoos K, et al. Role of environmental surveillance in determining the risk of hospital-acquired legionellosis: a national surveillance study with clinical correlations. Infect Control Hosp Epidemiol 2007;28:818-24.
Contents lists available at ScienceDirect
American Journal of Infection Control
American Journal of Infection Control
journal homepage: www.ajicjournal.org
Brief report
Control of Legionella contamination in a hospital water distribution system by monochloramine Isabella Marchesi PhD a, Stefano Cencetti MD b, Patrizia Marchegiano MD b, Giuseppina Frezza PhD a, Paola Borella MD a, Annalisa Bargellini PhD a, * a b
Department of Public Health Sciences, University of Modena and Reggio Emilia, Modena, Italy University Hospital Policlinico of Modena, Modena, Italy
Key Words: Legionella Pseudomonas Monochloramine Chlorine dioxide Hospital Hot water distribution system
Background: We report the results of 1-year application of monochloramine to control Legionella pneumophila contamination in a hospital hot water distribution system. Methods: In the main building of the hospital, a device continuously distributing monochloramine was installed. Legionella pneumophila and Pseudomonas spp contamination was followed in comparison with 2 other water networks in the same building using chlorine dioxide. Results: Monochloramine significantly reduced the number of contaminated sites compared with baseline (from 97.0% to 13.3%, respectively), chlorine dioxide device I (from 100% to 56.7%, respectively), and device II (from 100% to 60.8%, respectively). No positive sample exceeded 104 colony-forming units/L versus 59.4% at baseline. Conclusion: Monochloramine could represent a good alternative to chlorine dioxide in controlling legionellae contamination in public and private buildings. Copyright Ó 2012 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved.
The control of Legionella spp contamination in plumbing systems of large buildings may be a difficult challenge. This is of particular concern in hospital settings where contamination of hot water distribution system actually is the most important risk factor for nosocomial Legionella pneumonia.1,2 National and international guidelines aimed at preventing Legionella infection suggest the use of biocides for the routine treatment of contaminated hot water.3,4 Methods adopted include superheating, ultraviolet light, coppersilver ionization, ozone, hyperchlorination, chlorine dioxide, and point-of-use water filters. To date, the best procedure is not established, but the effectiveness of different systems in terms of reduction of contamination and related costs has been recently evaluated.5 Monochloramine is considered more effective than free chlorine in controlling Legionella contamination of longer distribution systems for its ability to better penetrate biofilms.6 Indeed, when adopted for disinfection of municipal water supplies in the United States, it was associated with a lower prevalence of
* Address correspondence to Annalisa Bargellini, PhD, Department of Public Health Sciences, University of Modena and Reggio Emilia, Via Campi 287, 41125 Modena, Italy. E-mail address: [email protected] (A. Bargellini). Supported by hospital funds. Conflicts of interest: None to report.
Legionella colonization in potable water systems of public and private buildings6,7 and decreased risk of nosocomial outbreaks of Legionnaires’ disease.8,9 In this study, we present the preliminary results of 1-year continuous injection of monochloramine in a hot water distribution system of a large hospital. A comparison with chlorine dioxide applied in 2 other water distribution systems of the same building is also presented. METHODS The experiment was conducted in a 9-story hospital building where 3 water networks distribute in parallel hot water. A device aimed to continuously produce and dispense monochloramine was installed from March 2009 on 1 of the 3 networks of this building. The monochloramine device (Sanipur, S.r.l., Brescia, Italy) was specifically developed to keep the level of monochloramine between 1.5 and 3.0 mg/L in the recirculation system, minimizing production of ammonia and other by-products. In the other 2 water networks of the same building, continuous chlorine dioxide devices furnished by different producers (device I: Sanipur S.r.l., Brescia, Italy, and device II: Culligan Italiana S.p.A., Bologna, Italy) were installed, assuring approximately 0.3 mg/L at distal outlets.5 Hot water samples were collected at regular intervals from the same sites to evaluate Legionella spp and Pseudomonas spp
0196-6553/$36.00 - Copyright Ó 2012 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.ajic.2011.03.008
10/41 (24.4)x 29 (1-46,000) 7/11 (63.6) 170 (10-850,000) 13/28 (46.4) 88x (1-13,000) 6/8 (75.0) 3332 (150-118,400) 13/60 (21.7) 169 (10-2,000) cfu, colony-forming units. *P < .001. y P < .002. z Only water positive. x P < .05.
6/19 (31.5) 70 (3-930)
15/15 (100) 10,752 (120-280,000) 9/15 (60.0) 34/60 (56.7)y 435* (25-120,000) 4/60 (6.7)* 14/14 (100) 11,714 (1,020-248,200) 7/14 (50.0)
After Before Before
After
Chlorine dioxide: device II
After
8/60 (13.3)* 335* (25-4,950) 0/60 (0)* 31/32 (97.0) 17,242 (100-946,000) 19/32 (59.4)
One-year application of monochloramine in a hot water distribution system heavily contaminated by L pneumophila gave satisfactory results both per se and compared with chlorine dioxide. Only 8 out of 60 samples were positive, and no sample exceeded 104 cfu/L versus 59.4% at baseline. In addition, after the first month, the positive samples were in total 4 of 48 (8.3%), all below 103 cfu/L. In previous studies, the percentage of water sites testing positive (30%) has been found to be a better predictor of hospital-acquired legionellosis than is the quantitative concentration,11 thus suggesting that monochloramine may be useful to reduce the risk of case appearance since the beginning of its application. The monochloramine level associated with bacteria below the detection limit approximates 3 mg/L, but 2 mg/L was sufficient for reducing legionellae below 100 cfu/L. Positive effects were not documented against Pseudomonas spp contamination, but this can depend on a limited pretreatment in the water network where monochloramine was applied. The significant reduction noticed by using chlorine dioxide suggests a higher capability to counteract
Before
DISCUSSION
Chlorine dioxide: device I
Table 1 shows Legionella pneumophila and Pseudomonas spp contamination in the water distribution systems treated by either monochloramine or chlorine dioxide. In the examined hospital building, only Legionella pneumophila was isolated. Before any treatment, L pneumophila concentration did not differ among the 3 distribution networks. Compared with chlorine dioxide, monochloramine exhibited a higher reduction in the number of both contaminated sites and sites exceeding 104 colony-forming units (cfu)/L. Chlorine dioxide significantly decreased number and/or concentration of Pseudomonas spp, whereas monochloramine had no significant effect. Figure 1 shows the trend of L pneumophila contamination with time by monochloramine. Taking the pretreatment samples all together as reference, the contamination risk was strongly reduced by monochloramine (odds ratio [OR], 0.003; confidence interval [CI]: 0.001-0.021, P < .001), followed by chlorine dioxide device I (OR, 0.022; CI: 0.003-0.168, P < .05) and device II (OR, 0.026; CI: 0.003-0.202, P < .05). A negative correlation was observed between residual monochloramine and L pneumophila concentration (r ¼ 0.341, P < .01, on total samples; r ¼ 725, P < .05, on positive samples). The regression line parameters revealed that 2 mg/L maintained L pneumophila below 100 cfu/L and 3 mg/L below the detection limit (25 cfu/L). For chlorine dioxide, taking the 2 systems together, the coefficients of correlation were r ¼ 0.606 on all samples and r ¼ 0.610 on positive samples (P < .001). Residual chlorine dioxide concentrations of 0.3 and 0.6 mg/L were requested at distal sites to obtain <100 and <25 cfu/L legionellae, respectively. Before any treatment, 5 of 60 positive samples (8.4%) were colonized by L pneumophila serogroup 1. The trend did not change after application of monochloramine (1/8 positive samples, 12.5%), whereas the percentage significantly increased after application of chlorine dioxide: 35 of 65 (53.8%), c2 ¼ 30.21, P < .001. No variation in both nitrite and nitrate water concentration was observed during the entire study, and no sample exceeded the limit of 0.50 mg/L and 50 mg/L, respectively, requested for potable water.
Table 1 Legionella pneumophila and Pseudomonas spp contamination in 3 hot water distribution systems of the same building before and after application of continuous disinfecting procedures
RESULTS
Legionella pneumophila positive, n (%) Geometric mean cfu/L (range)z 104 cfu/L, n (%) Pseudomonas spp positive, n (%) Geometric mean cfu/100mL (range)z
colonization using standardized cultural procedures and to measure residual disinfectant concentration.10 Water samples were taken from heater, return loop, and distal outlets (shower and/or tap selected on the basis of distance and exposure risk), without flaming and after flushing for 1 minute. Only viable planktonic bacteria were enumerated.
31/51 (60.8)y 1297* (50-22,500) 5/51 (9.8)*
I. Marchesi et al. / American Journal of Infection Control 40 (2012) 279-81
Monochloramine
280
I. Marchesi et al. / American Journal of Infection Control 40 (2012) 279-81
281
L. pneumophila Log cfu/L
5 Monochloramine 4
3
return loop
heat exchanger
tap A shower C
tap B
2
1
mt h 15
mt h 14
mt h 12
mt h 10
8m th
6m th
4m th
3m th
1m th
1w ee k
be for
e
0
Time Fig 1. Trend of L pneumophila contamination before and after monochloramine continuous injection.
Pseudomonas spp but could also depend on their higher number and/or concentration before treatment. The presence of L pneumophila serogroup 1 was not influenced by monochloramine, whereas chlorine dioxide increased the number of points colonized by this serogroup, which is more frequently associated with the disease. This is in line with the decreased number of Legionnaires’ disease cases observed in the United States by using monochloramine for residual disinfection.8,9 One possible limitation of our study is the absence of a control building, but the Italian guidelines require intervention when contamination exceeds 104 cfu/L.3 The hospital has other 4 buildings: 2 contaminated and 2 not. In those contaminated, shock treatments are periodically adopted, insufficient to control contamination as documented in our previous study5; but, at present, 1 building is closed, and the other actually avoids patient exposure to hot water. Our results suggest that continuous injection of monochloramine in hot water distribution systems can control L pneumophila and Pseudomonas spp in contaminated buildings without modifying water content in nitrite and nitrate. Investigations are in progress to evaluate the production of disinfection by-products and the long-term effect of monochloramine also on biofilm development and amoeba presence associated with Legionella. Acknowledgment The author thanks Sanipur S.r.l. for technical support and equipment provided.
References 1. Yu VL, Stout JE. Hospital characteristics associated with colonization of water systems by Legionella and risk of nosocomial Legionnaires’ disease: a cohort study of 15 hospitals. Infect Control Hosp Epidemiol 2000;21: 434-5. 2. Sabria M, Yu VL. Hospital-acquired legionellosis: solutions for a preventable infection. Lancet Infect Dis 2002;2:368-73. 3. Italian guidelines for prevention and control of legionellosis G.U. Gazzetta Ufficiale della Repubblica Italiana, serie generale, n.103. 2000. 4. World Health Organization. Legionella and the prevention of legionellosis. Geneva: World Health Organization; 2007. 5. Marchesi I, Marchegiano P, Bargellini A, Cencetti S, Frezza G, Miselli M, et al. Effectiveness of different methods to control legionella in the water supply: ten-year experience in an Italian university hospital. J Hosp Infect 2011;77: 47-51. 6. Moore MR, Pryor M, Fields B, Lucas C, Phelan M, Besser RE. Introduction of monochloramine into a municipal water system: impact on colonization of buildings by Legionella spp. Appl Environ Microbiol 2006;72:378-83. 7. Flannery B, Gelling LB, Vugia DJ, Weintraub JM, Salerno JJ, Conroy MJ, et al. Reducing Legionella colonization in water systems with monochloramine. Emerg Infect Dis 2006;12:588-96. 8. Kool JL, Carpenter JC, Fields BS. Effect of monochloramine disinfection of municipal drinking water on risk of nosocomial Legionnaires’ disease. Lancet 1999;353:272-7. 9. Heffelfinger JD, Kool JL, Fridkin S, Fraser VJ, Hageman J, Carpenter J, et al. Risk of hospital-acquired Legionnaires’ disease in cities using monochloramine versus other water disinfectants. Infect Control Hosp Epidemiol 2003;24: 569-74. 10. Borella P, Montagna MT, Stampi S, Stancanelli G, Romano-Spica V, Triassi M, et al. Legionella contamination in hot water of Italian hotels. Appl Environ Microbiol 2005;71:5805-13. 11. Stout JE, Muder RR, Mietzner S, Wagener MM, Perri MB, DeRoos K, et al. Role of environmental surveillance in determining the risk of hospital-acquired legionellosis: a national surveillance study with clinical correlations. Infect Control Hosp Epidemiol 2007;28:818-24.