Abstract
The Legionella colonization frequency at 385 Greek hotel hot and cold water distribution systems was 20.8%. Legionella contamination was associated with the presence of an oil heater (odds ratio [OR] = 2.04, 95% confidence interval [CI] = 1.12 to 3.70), with the sample temperature (OR = 0.26, 95% CI = 0.1 to 0.5), with seasonal operation (OR = 3.23, 95% CI = 1.52 to 6.87), and with the presence of an independent disinfection system (OR = 0.30, 95% CI = 0.15 to 0.62). The same water temperatures, free-chlorine levels, and pHs differently affect the survival of various Legionella spp.
Legionnaires' disease is often associated with travel and with staying in hotels. Few reports, covering a small number of hotels, have studied Legionella colonization of water distribution systems of hotels.
In Greece, Legionella pneumophila was isolated and identified in hotel water distribution systems associated with cases of legionellosis for the first time in 1989 (2). Moreover, 344 cases of Legionella infection were associated with traveling in Greece by the European Surveillance Scheme for Travel Associated Legionnaires' Disease network from 1987 to 2005 (http://www.ewgli.org/).
To investigate legionella contamination of the water distribution systems of hotels accommodating athletes and tourists during the Athens 2004 Olympic Games, a descriptive multicentric study was conducted at seven locations in Greece. The aims of this study were to evaluate the frequency of Legionella colonization of water distribution systems of hotels, to identify risk factors for Legionella contamination associated with water distribution systems and water characteristics, and to identify remedial action needed to improve hotel water distribution systems.
MATERIALS AND METHODS
From January 2003 through September 2004, 1,086 samples were collected from 385 hotels in Athens, Volos, Chalkida, Korinthos (central Greece), Thessaloniki (northern Greece), Patras (western Greece), and Iraklio (southern Greece). The sample collection and microbiological analysis were part of the Environmental Health Surveillance Program developed by the Olympic Planning Unit and implemented for the Athens 2004 Olympic Games, which is described elsewhere (8, 9). Methods used for sample collection, storage conditions, and microbiological analysis have been described elsewhere (7).
Microbiologic analyses for Legionella spp. were performed by the National Legionella Reference Laboratory of Southern Greece in Athens and the National Legionella Reference Laboratory of Northern Greece in Thessaloniki. Microbiological testing for total plate count, coliform bacteria, Escherichia coli, intestinal enterococci, and Clostridium perfringens (including spores) was conducted in the Central Public Health Laboratory in accordance with the methods specified in the standing European legislation (5a).
A detailed standardized form was developed to register the hotel water supply systems at the seven locations (8, 9). Data included in the registry were used to evaluate risk factors possibly associated with Legionella colonization.
Data were analyzed with Epi-Info 2000 (Centers for Disease Control and Prevention, Atlanta, GA) and SPSS for Windows release 11.0.1 software (SPSS Inc., Chicago, IL) by t test or Mann-Whitney test for quantitative data and by chi-square test or Fisher exact test for qualitative data. Relative risk (RR) and 95% confidence interval (CI) were calculated to assess categorical risk variables associated with legionella-positive test results. Variables that were significant in the univariate analysis were entered into a multiple logistic regression model. By using conditional logistic regression models, independent predictors of colonization were established. Variables were retained in the model if the likelihood ratio test result was significant (P < 0.05).
RESULTS
In 80 hotel hot water distribution systems (20.8%, 95% CI = 16.9 to 25.2) the Legionella count was >500 CFU liter−1 in at least one sample, while in 25 hotels (6.5%) the Legionella count was ≥104 CFU liter−1 in at least one sample. About 35 hotel hot water distribution systems (9%) needed remedial action because the Legionella count was ≥104 CFU liter−1 in at least one sample or >103 but <104 CFU liter−1 in more than two samples (6).
A total of 268 water samples (24.7%) were contaminated by Legionella spp. with concentrations of ≥500 CFU liter−1 (Table 1). Of the total of 277 isolates, L. pneumophila was the most frequently isolated species (87%), while 14.9% of the positive samples contained ≥104 CFU liter−1 Legionella spp. and 92.5% of these samples were contaminated by L. pneumophila.
TABLE 1.
Legionella contamination in hot and cold water samples examined
| Organism | No. of samples/total (%)
|
|||||||
|---|---|---|---|---|---|---|---|---|
| Hot water samples
|
Cold water samples
|
|||||||
| Positive | With 500-999 CFU liter−1 | With 103-9,999 CFU liter−1 | With ≥104 CFU liter−1 | Positive | With 500-999 CFU liter−1 | With 103-9,999 CFU liter−1 | With ≥104 CFU liter−1 | |
| Legionella spp. | 260/962 (27.0) | 89/260 (34.2) | 131/260 (50.4) | 40/260 (15.4) | 8/124 (6.4) | 3/8 (37.5) | 5/8 (62.5) | 0/8 (0.0) |
| L. pneumophila serogroup 1 | 120/962 (12.5) | 42/120 (35.0) | 59/120 (49.7) | 19/120 (15.8) | 1/124 (0.8) | 0/1 (0.0) | 1/1 (100.0) | 0/1 (0.0) |
| L. pneumophila serogroups 2-14 | 117/962 (12.2) | 38/117 (32.5) | 61/117 (52.1) | 18/117 (15.4) | 3/124 (2.1) | 1/3 (33.3) | 2/3 (66.7) | 0/3 (0.0) |
| Legionella spp. other than L. pneumophila | 32/962 (3.3) | 19/32 (59.4) | 10/32 (31.2) | 3/32 (9.4) | 4/124 (3.2) | 2/4 (50.0) | 2/4 (50.0) | 0/4 (0.0) |
Factors associated with contamination by Legionella spp. upon application of univariate analysis and logistic regression are shown in Tables 2 and 3, respectively.
TABLE 2.
Association of hot water distribution systems and hotel characteristics with Legionella contamination by univariate analysis
| Characteristic | No. (%) of samples:
|
RR (95% CI) | |
|---|---|---|---|
| With Legionella spp. | Without Legionella spp. | ||
| No. of rooms, ≥80 | 31 (27.0) | 49 (18.2) | 1.5 (1.0-2.2)a |
| Hotel age, ≥30 yr | 34 (17.2) | 42 (28.0) | 0.6 (0.4-0.9)a |
| Boiler age, ≥20 yr | 5 (20.8) | 58 (22) | 0.9 (0.4-2.1) |
| Apartments | 73 (21.2) | 6 (23.1) | 0.9 (0.4-1.9) |
| Bungalows | 11 (44) | 68 (19.8) | 2.2 (1.3-3.6)b |
| Seasonal operation | 16 (42.1) | 58 (18.4) | 2.3 (1.5-3.5)b |
| Water source, municipal supply | 70 (20.5) | 10 (23.3) | 0.8 (0.5-1.6) |
| Water source, well | 3 (25.0) | 77 (20.6) | 1.2 (0.4-3.3) |
| Water tank | 30 (27.5) | 44 (17.1) | 1.6 (1.0-2.4)a |
| Operation temp, ≥55°C | 15 (27.3) | 58 (19.0) | 1.4 (0.9-2.3) |
| Thermal insulation of system | 50 (20.4) | 19 (19.4) | 1.0 (0.6-1.7) |
| Steel plumbing | 5 (26.3) | 75 (20.5) | 1.2 (0.6-2.8) |
| Cu-PVC plumbing | 3 (23.1) | 77 (20.7) | 1.1 (0.4-3.0) |
| Cu plumbing | 32 (21.1) | 48 (20.6) | 1.0 (0.7-1.5) |
| PVC plumbing | 9 (24.3) | 71 (20.4) | 1.2 (0.6-2.2) |
| Independent disinfection system | 9 (10.8) | 245 (29.3) | 0.4 (0.2-0.7)c |
| Independent water filtration | 10 (23.8) | 49 (21.6) | 1.1 (0.6-2.0) |
| Water storage tank protected | 22 (30.1) | 3 (18.8) | 1.6 (0.4-4.7) |
| Oil heat | 55 (25.1) | 25 (15.1) | 1.7 (1.0-2.5)a |
| Electric heat | 8 (14.8) | 72 (21.8) | 0.6 (0.3-1.3) |
| Solar heat | 8 (17.4) | 72 (21.2) | 0.8 (0.4-1.6) |
P = 0.01 to 0.03.
P = 0.001 to 0.007.
P = 0.00001.
TABLE 3.
Multiple logistic regression of hot water distribution system and water characteristics associated with Legionella contamination
| Characteristic | ORc | 95% CI |
|---|---|---|
| Hotel age, ≥30 yr | 0.65 | 0.37-1.15 |
| No. of rooms, ≥80 | 1.14 | 0.60-2.15 |
| Bungalows | 1.74 | 0.61-4.90 |
| Seasonal operation | 3.23 | 1.52-6.87a |
| Water tank present | 1.55 | 0.82-2.93 |
| Oil heat | 2.04 | 1.12-3.70a |
| Sample temp, ≥55°C | 0.26 | 0.1-0.5b |
| Independent disinfection system | 0.30 | 0.15-0.62a |
P = 0.01 to 0.02.
P = 0.00001.
OR, odds ratio.
Legionellae were not isolated when the sample temperature was above 60.3°C or below 23.7°C. Legionella contamination was not associated with positive microbiological results for coliform bacteria, E. coli, intestinal enterococci, and Clostridium perfringens, including spores (RR = 1.4, 95% CI = 0.3 to 7.0) but was associated with the total plate count (Table 4).
TABLE 4.
Association of physical, chemical, and microbiologic water characteristics with Legionella contamination
| Characteristic and organism(s) | No. of samples/total (%)
|
RR | P value | |
|---|---|---|---|---|
| Positive for Legionella spp. | Negative for Legionella spp. | |||
| pH of ≥7.8 | ||||
| Legionella spp. | 41/176 (24.1) | 210/828 (25.4) | 0.9 | 0.4 |
| L. pneumophila serogroup 1 | 28/170 (16.5) | 84/128 (10.1) | 1.6 | 0.01 |
| L. pneumophila serogroups 2-14 | 12/170 (7.1) | 102/828 (12.3) | 0.57 | 0.02 |
| Legionella spp. other than L. pneumophila | 1/170 (0.6) | 33/828 (4.0) | 0.14 | 0.012 |
| Sample temp of ≥55 | ||||
| Legionella spp. | 15/145 (10.3) | 239/773 (30.9) | 0.33 | <0.00001 |
| L. pneumophila serogroup 1 | 2/145 (1.4) | 114/773 (14.7) | 0.09 | <0.00001 |
| L. pneumophila serogroups 2-14 | 10/145 (6.9) | 105/773 (13.6) | 0.50 | 0.01 |
| Legionella spp. other than L. pneumophila | 3/145 (2.1) | 29/773 (3.8) | 0.55 | 0.2 |
| Free-chlorine level of ≥0.2 mg/liter | ||||
| Legionella spp. | 14/91 (15.4) | 143/460 (31.1) | 0.49 | 0.001 |
| L. pneumophila serogroup 1 | 3/91 (3.3) | 68/460 (14.8) | 0.22 | 0.0008 |
| L. pneumophila serogroups 2-14 | 8/91 (8.8) | 73/460 (15.9) | 0.55 | 0.05 |
| Legionella spp. other than L. pneumophila | 3/91 (3.3) | 10/460 (2.2) | 1.51 | 0.3 |
| Total plate count of ≥400 CFU | ||||
| Legionella spp. | 83/213 (39.0) | 141/569 (24.8) | 1.57 | 0.00008 |
| L. pneumophila serogroup 1 | 34/213 (16.0) | 78/569 (13.7) | 1.16 | 0.2 |
| L. pneumophila serogroups 2-14 | 39/213 (18.3) | 56/569 (9.8) | 1.86 | 0.0013 |
| Legionella spp. other than L. pneumophila | 13/213 (6.1) | 13/569 (2.3) | 2.67 | 0.01 |
Concentrations of Legionella spp. in positive immediate samples and postflush samples are shown in Table 5.
TABLE 5.
Legionella contamination of immediate samples and postflush samples of the hot water collected
| No. of Legionella CFU liter−1 | No. (%) of positive samples of hot water
|
|
|---|---|---|
| Immediate sample | Postflush sample | |
| 500-999 | 43 (48) | 45 (52) |
| 103-9,999 | 73 (60.5) | 54 (39.5) |
| ≥104 | 27 (73.4) | 11 (26.6) |
| Total | 143 (57.5) | 110 (42.5) |
Legionella spp. and the number of positive samples differed with relation to the geographic location. In Athens, Thessaloniki, and Patras, the isolation rates were 19.6% (CI = 16.3 to 23.3), 17.4% (CI = 12.7 to 23.0), and 40.0% (CI = 24.9 to 56.7), respectively. In Volos, Chalkida, Korinthos, and Iraklio, the isolation rates were 28.1% (CI = 17.0 to 41.5), 70.0% (CI = 45.7 to 88.1), 65.0% (CI = 40.8 to 84.6), and 6.1% (CI = 0.7 to 20.2), respectively. Hotels in Athens and Thessaloniki, the two largest cities in Greece, had lower contamination rates than the ones in small towns (RR = 0.42, CI = 0.30 to 0.57, P > 0.00001).
The isolation rates per positive hotel distribution system were higher during summer (24.1%) than in winter, autumn, or spring (the isolation rates were 17.5%, 16.2%, and 17.4%, respectively).
No cases of Legionnaires' disease were linked to the hotels inspected during the study period.
DISCUSSION
Our data indicate that at the time of this study, 20.8% of the water distribution systems in Greek hotels were positive for Legionella spp. (≥500 CFU liter−1). Other investigators have observed percentages of Legionella contamination ranging from 63.6% (11) to 75% (4) in 11 and 40 Italian hotels, respectively, but the detection limit of the microbiological examination procedure was 25 CFU liter−1 while in our study it was 500 CFU liter−1. Moreover, in our study the numbers of hotels and samples examined were higher (385 hotels, 1,086 samples), resulting in a relatively narrow range of CI limits of the percentage of positive distribution systems (20.8%) of 16.9 to 25.2.
In our study 4.2% of the hot water samples had concentrations of ≥104 CFU liter−1, while in the Italian studies 11.8% (4) and 17.4% (11) of the samples examined had concentrations of ≥104 CFU liter−1. Furthermore, L. pneumophila was isolated from 24.6% of the samples examined, while in the studies of the Italian hotels L. pneumophila was found in 62.2% (4) and 60.9% (11) of the samples examined. To assess if a difference in the colonization rates of hotels in Italy and Greece does exist, we compared the cases of legionellosis among tourists who visited Greece and Italy reported to the European Surveillance Scheme for Travel Associated Legionnaires' Disease. In 2003 and 2004, traveling to Greece was associated with 28 and 24 cases of Legionnaires' disease, respectively, while traveling to Italy was associated with 123 and 111 cases, respectively, (http://www.ewgli.org/data/data_tables/year_onset_country_travel.asp). During the 2-year period, the numbers of cases per million tourists were 1.85 (52 cases per 28 million tourists) in Greece and 3.00 (234 cases per 78 million tourists) in Italy. This higher rate of cases in Italy is consistent with the higher rates of Legionella colonization of the Italian hotels reported in these studies.
Internationally, about 90% of Legionnaires' disease cases are due to L. pneumophila, and predominant serogroup 1 of L. pneumophila accounts for 84% of the cases (13). Our findings support the idea that a small proportion of Greek hotels posed a risk for Legionnaires' disease at the time of the study, due to the high concentrations of L. pneumophila (24.6% of the total hot water samples examined) and the number of positive distribution systems which exceeded the safety levels of the European Surveillance Scheme for Travel Associated Legionnaires' Disease guidelines (9% of the hotel water distribution systems examined). Regulations regarding the control of Legionella spp. in water distribution systems should be established in Greece.
Seasonal hotel operation was the main predictor of the presence of legionellae in hotel water systems, while in small towns with hotels operating to a major degree seasonally, the isolation rates were higher than in big towns, where hotels generally operate all year round. Control measures should be applied intensively during summer, especially in hotels operating seasonally.
Storage water heaters powered directly by electricity and solar water heaters showed a lower prevalence of colonization by legionellae than those heated by oil. Similar were the findings in previous studies concerning domestic hot water in Wellington, New Zealand (3), and in six towns in Italy (5), while the opposite was observed in a previous study in Quebec City (1). In Greece, oil heaters are part of the hotel central heating system while electric heaters are used as point-of-use heating systems, have minimal storage, and are placed in each hotel apartment or hotel room.
The temperature of the hot water samples was negatively associated with contamination with L. pneumophila serogroup 1 and to a lesser degree with contamination with serogroups 2 to 14, suggesting that serogroup 1 is more sensitive to high temperatures. This is the opposite of the findings of the Italian study, which indicates that L. pneumophila serogroup 1 is more resistant to higher temperatures (4).
In our study, increases in pH correlated negatively with the counts of Legionella spp. other than L. pneumophila and of L. pneumophila serogroups 2 to 14 and positively with counts of L. pneumophila serogroup 1. These findings confirm other reports, which indicated a positive association of L. pneumophila with pH (10, 11, 12). The differences in distribution of species according to water characteristics confirm the hypothesis of other reports, which indicated that Legionella strains differ in sensitivity to environmental risk factors and have different ecological niches (4, 5, 11).
Contamination by Legionella spp. was associated with total plate counts above the acceptable level in samples collected from the same water distribution system. This association may be explained by the presence in contaminated distribution systems of biofilms consisting of bacteria and other microorganisms embedded in a protective layer with entrained debris attached to a surface (6). The total plate count could be used as an indicator of the presence of Legionella spp.
Future studies in Greece should directly evaluate levels of water distribution system contamination and water and system characteristics in relation to Legionnaires' disease cases and outbreaks.
Acknowledgments
The Hellenic Ministry of Health funded this study.
We thank all environmental health inspectors of Olympic cities and the staff of the National School of Public Health. We also thank S. D. Alexiou, A. Mavridou, Vasilios D. Daniilides, L. Georgiadou, and George Panagako.
Footnotes
Published ahead of print on 19 January 2007.
REFERENCES
- 1.Alary, M., and J. R. Joly. 1991. Risk factors for contamination of domestic hot water systems by legionellae. Appl. Environ. Microbiol. 57:2360-2367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Alexiou, S. D., A. Antoniadis, J. Papapaganagiotou, and T. Stefanou. 1989. Isolation of Legionella pneumophila from hotels of Greece. Eur. J. Epidemiol. 5:47-50. [DOI] [PubMed] [Google Scholar]
- 3.Bates, M. N., E. Maas, T. Martin, D. Harte, M. Grubner, and T. Margolin. 2000. Investigation of the prevalence of Legionella species in domestic hot water systems. N. Z. Med. J. 113:218-220. [PubMed] [Google Scholar]
- 4.Borella, P., M. T. Montagna, S. Stampi, G. Stancanelli, V. Romano-Spica, M. Triassi, I. Marchesi, A. Bargellini, D. Tato, C. Napoli, F. Zanetti, E. Leoni, M. Moro, S. Scaltriti, G. Ribera D'Alcala, R. Santarpia, and S. Boccia. 2005. Legionella contamination in hot water of Italian hotels. Appl. Environ. Microbiol. 71:5805-5813. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Borella, P., M. T. Montagna, V. Romano-Spica, S. Stampi, G. Stancanelli, M. Triassi, R. Neglia, I. Marchesi, G. Fantuzzi, D. Tato, C. Napoli, G. Quaranta, P. Laurenti, E. Leoni, G. De Luca, C. Ossi, M. Moro, and G. Ribera D'Alcala. 2004. Legionella infection risk from domestic hot water. Emerg. Infect. Dis. 10:457-464. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5a.European Economic Community. 1998. Council directive 98/83/EC of 3 November 1998 on the quality of water intended for human consumption. Off. J. Eur. Communities L330/32:L330/36, L330/49. [Google Scholar]
- 6.European Working Group for Legionella Infections. 2005. European guidelines for control and prevention of travel associated Legionnaires' disease, p. 14, 42, 60. European Surveillance Scheme for Travel Associated Legionnaires' Disease and European Working Group for Legionella Infections, London, United Kingdom. http://www.ewgli.org/data/european_guidelines/european_guidelines_jan05.pdf.
- 7.Hadjichristodoulou, C., G. Goutziana, V. Mouchtouri, C. Kapoula, A. Konstantinidis, E. Velonakis, A. Vatopoulos, and J. Kremastinou. 2006. Evaluation of standardized scored inspections for Legionnaires' disease prevention, during the Athens 2004 Olympics. Epidemiol. Infect. 134:1074-1081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Hadjichristodoulou, C., V. Mouchtouri, E. S. Soteriades, V. Vaitsi, V. Kolonia, A. P. Vasilogiannacopoulos, and J. Kremastinou. 2005. Mass gathering preparedness: the experience of the Athens 2004 Olympic and Para-Olympic Games. J. Environ. Health 67:52-57. [PubMed] [Google Scholar]
- 9.Hadjichristodoulou, C., E. S. Soteriades, V. Kolonia, M. E. Falagas, E. Pantelopoulos, G. Panagakos, V. Mouchtouri, and J. Kremastinou. 2005. Methodological aspects of a GIS-based environmental health inspection program used in the Athens 2004 Olympic and Para Olympic Games. BMC Public Health 5:93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kusnetsov, J., E. Torvinen, O. Perola, T. Nousiainen, and M. L. Katila. 2003. Colonization of hospital water systems by legionellae, mycobacteria and other heterotrophic bacteria potentially hazardous to risk group patients. APMIS 111:546-556. [DOI] [PubMed] [Google Scholar]
- 11.Leoni, E., G. De Luca, P. P. Legnani, R. Sacchetti, S. Stampi, and F. Zanetti. 2005. Legionella waterline colonization: detection of Legionella species in domestic, hotel and hospital hot water systems. J. Appl. Microbiol. 98:373-379. [DOI] [PubMed] [Google Scholar]
- 12.Marrie, T., P. Green, S. Burbridge, G. Bezanson, S. Neale, P. S. Hoffman, and D. Haldane. 1994. Legionellaceae in the potable water of Nova Scotia hospitals and Halifax residences. Epidemiol. Infect. 112:143-150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Yu, V. L., J. F. Plouffe, M. C. Pastoris, J. E. Stout, M. Schousboe, A. Widmer, J. Summersgill, T. File, C. M. Heath, D. L. Paterson, and A. Chereshsky. 2002. Distribution of Legionella species and serogroups isolated by culture in patients with sporadic community-acquired legionellosis: an international collaborative survey. J. Infect. Dis. 186:127-128. [DOI] [PubMed] [Google Scholar]