- Email: [email protected]
Legionella in hospitals: a review
Journal of Hospital Infection
(1991)
Legionella
18 (Supplement
in hospitals:
C. A. Hart Department
of Medical
A), 481-489
a review
and T. Makin
Microbiology, University of Liverpool, Liverpool L69 3BX, UK
P.O. Box 147,
Summary: Although epidemics of nosocomial Legionnaires’ disease attract great attention, up to 30% of sporadic cases of hospital-acquired pneumonia are caused by legionellae. Legionellae are ubiquitous contaminants of potable water and can achieve high numbers in the hot-water systems of large buildings such as hospitals. They are present in the mains water supply in small numbers but are amplified considerably in the hospital’s hot-water system. This is encouraged by water temperatures below SO”C, areas of stagnation and sludge formation, the presence of amoebae and other bacteria and the materials used in the piping. Formation of aerosols from contaminated water is a major mode of spread of legionellae, but there is evidence to suggest that aspiration is also a mode of entry. Safe levels of legionellae in cooling towers have been defined, but not for hot-water systems. A combination of culture and antigen detection by immunofluorescence offer the best method for enumerating legionellae in environmental samples. Control involves a mixture of physical (heat, UV irradiation, sanitation) and chemical (hypochlorite, ozone) methods combined with good plumbing practice (e.g. arrangement of pumps and calorifiers, elimination of dead-‘legs). Adequate control can be costly and requires considerable attention to detail. Keywords:
Legionellae;
nosocomial
infection;
hospital
reservoirs.
Introduction Since the initial description of an outbreak of pneumonia associated with an American Legion convention in Philadelphia in 1976l there has been great interest in legionellae in both the scientific and popular press. Although the initial outbreak was linked to a hotel, it is now clear that hospital-associated Legionnaires’ disease, both epidemic2B3 and sporadic4,’ is not uncommon. The infection generally presents as either a severe pneumonia with or without extrapulmonary manifestation or a milder ‘flu-like’ febrile illness (Pontiac fever),‘j but infection at other sites can occur. Such infections include pericarditis,’ endocarditis’ and cutaneous abscesses.’ In order for legionellae to produce disease, several factors must be considered. Firstly, because human-to-human transmission does not occur, there must be an environmental reservoir. Secondly there must be a mechanism for getting the bacterium from the reservoir to the site of Correspondence 0195%6701/91/06A481+09
to: Professor C. A. Hart. $03.00/O
0 1991 The Hospital
481
Infection
Society
482
C. A. Hart and T. Makin
infection and thirdly the organism must lodge in a susceptible host. In this review we will discuss the properties of legionellae, their reservoirs and modes of spread in hospitals and methods of detection and control of legionellae in the hospital environment. The bacterium
There are currently 14 recognized serogroups of Legionella pneumophila and at least 29 other Legionella species (Table I) although there is still some disagreement over speciation. Not all of these bacteria have been associated with disease in man, and even within L. pneumophila serogroup 1 there appear to be subtypes of differing virulence.” Legionellae are not thermophilic but are thermotolerant, i.e. they are able to survive at relatively high temperatures. For example, legionellae will multiply over the temperature range 2CM3”C and can survive for varying periods at temperatures 40-60°C. At SO”C, 90% of the organisms are killed within 2 h. Legionellae can survive in water for over a year.” Legionellae are able to survive aerosolized in droplets 1-5 pm in diameter and there is evidence to suggest that virulent strains survive longer.‘* Legionellae appear to be able to adhere to substrata such as pipes, rubber, plastics and sediment and can thus persist even when there is constant flushing of a piped-water system. It is possible that they exist in biofilms with release of planktonic organisms. This may explain the discrepancy between the ability of disinfectants to kill legionellae in the test tube and that in piped water systems. Finally, there is some evidence that legionellae are engulfed by amoebae and are able to survive and multiply inside them.13 This might also be a mechanism prolonging survival of legionellae in water. It is also noteworthy that amoebae constantly generate superoxide,14 thus the ability of
Table Legionella pneumophila L. a&a* L. birminghamensis L. boxernaG* L. cherrii L. cincinattiensis L. dumofii* L. erythra L. feeleii* L. geestiae L. gormanii* L. rhackeliae* L. israelensis L. jamestowniensis L. jordanis*
I. Legionella (serogroups
* Associated with human disease.
pneumophila l-14)*
and Legionella L. L. L. L. L. L. L. L. L. L. L. L. L. L. L.
spp.
londiniensis longbeachi? maceacherni? micdadei* nautarum oakridgensis parisiensis quateriensis rubrilucens sainthelensi santicrucis spiritensis steigwaltii wadsworthii* worsliensis
A review
of Legionella
legionellae to survive in amoebae might in a patient’s alveolar macrophages.
in hospitals
be related to their ability
483
to survive
Reservoirs
contaminants of potable water. Legionella Legionellae are common pneumophila has been isolated from drinking water from various parts of the world with a frequency ranging from 0 to 8.8% and isolation occurs more frequently in warm summer months. is It is probable that the small numbers of legionellae in the incoming water are amplified by the hot water systems in large buildings such as hospitals and hotels. For example, the water systems of over 50% of large establishments have been shown to contain disease have resulted L. pneumophila. i6 Large outbreaks of Legionnaires’ from contamination of air conditioning and ventilation systems,3,4T’7 and legionellae can be found in large numbers in water in cooling towers. The other major site for colonization and multiplication of legionellae is the hot-water system, or other areas where water temperature is raised into the range 2540°C. In many halspitals, hot water is produced by taking water through a number of heaters or calorifiers. The water then circulates throughout the hospital and returns to the calorifiers. The factors promoting the expansion and persistence of legionella populations have been described by others. 2118They include the arrangement of calorifiers, which should be in series rather than parallel, the presence of a dead space at the bottom of the calorifier which allows sludge and legionellae to persist, and maintenance of the hot-water output below 57°C. The presence of dead-legs in the hot-water system, e.g. associated with showers,19-21 hosepipes or fire sprinkler systems, the use of spas or jacuzzis, and the addition of hot water to humidifiers or nebulizers** can all serve to amplify Legionella populations. We have even been able to find legionellae in the melt-water from ward ice-making machines. The pH, ionic composition and conductivity of the water and the nature and age of materials used to make pipes and valves also affect the growth and survival of legionellae. For example, some silicone-based rubbers enhance the growth of L. pneumophila by a factor of 105, whereas rubbers containing the accelerator thiuram inhibit ;growth.23 There has been considerable debate over the relative importance of the two major reservoirs, namely cooling towers and evaporative condensers and the hot-water system. It is clear that episodes of infection have arisen from contamination of water in cooling towers,3,24 and the evidence linking such reservoirs and development of infection is strong. Indeed, in one outbreak of Legionnaires’ disease in a retirement hotel in which L. pneumophila serogroup 1 was found in both the hot-water system and an evaporative condenser, it was demonstrated using air sampling and monoclonal antibody subtyping that the condenser was the source.25 However, it has also been shown that Legionnaires’ disease could occur in
484
C. A. Hart and T. Makin
areas where there was no possibility of aerosols from evaporative condensers.20,26 Similarly outbreaks have been controlled by treatment of the water system alone even though a cooling tower was present.27 It would appear that the two reservoirs are equally important. Guidelines on acceptable levels of legionella contamination of cooling towers have been produced28 but no such guidance is available for hot-water systems. In cooling towers Legionella counts of less than lo3 I-’ have been suggested as being acceptable, those above lo5 1-l indicate that corrective action is required. It has been demonstrated that there is a link between the extent of colonization of a hospital’s hot-water system and the development of nosocomial Legionnaires’ disease. 4,29 Thus, in an Ear Nose and Throat hospital where 67% of sites sampled contained legionellae, 30% of the cases of hospital-acquired pneumonia were due to legionellae. After chlorination of the system only 3.5% of sites were contaminated and none of the cases of hospital-acquired pneumonia were caused by legionellae. The patient group studied were resident on head and neck oncology wards and such patients are known to be highly susceptible to legionella infection. Nevertheless, it appears that once contamination of a hospital’s hot-water system involves more than 30% of sites sampled, nosocomial Legionnaires’ disease occurs.19 Unfortunately, there is little information on the relationship between numbers of legionellae in the hot-water systems and development of infection. However, we were able to demonstrate that a few days prior to acquisition of pneumonia due to L. pneumophila serogroup 12 by a renal transplant recipient, the water from her hand basin hot tap contained 6 x lo3 cfu 1-l of the same bacterium.5 Modes
of spread
The most likely mode of spread is by aerosolized particles < 5 pm diameter containing legionellae. It is possible to infect guinea-pigs with aerosols containing as few as 3.7 X lo4 cfu L. pneumophila serogroup 13’ and the bacteria can survive for up to 2 h in such aerosols.31 Aerosols can be produced easily in evaporative condensers, cooling towers, jacuzzis, showers, nebulizers and in large buildings even by normal tap pressure. There has even been an outbreak of Pontiac fever due to L. anisa where an aerosol created by an artificial fountain was implicated.32 It has also been suggested that aspiration or direct inoculation of legionellae into the trachea-bronchial tree via respiratory manipulation are important modes of transmission especially in head and neck surgery patients.2,4 Detection
Perhaps the environmental
of legionellae
in the environment
most difficult question to answer is should there be testing and if so, how frequently? In general the indication
A review
of Legionella
in hospitals
485
for environmental testing in hospitals is the presence of a cluster of cases of Legionnaires’ disease or an increasing number of sporadic cases. Because legionellae are ubiquitous and because of the high cost, routine surveillance some is not recommended.33 Perhaps in the lig ht of recent developments re-appraisal of this position is necessary. Guidelines have been produced giving acceptable levels of legionellae in cooling towers, in order to comply with these it is thus necessary to measure levels actually present. There does appear to be a relationship between the extent of environmental colonization and development of nosocomial Legionnaires’ disease.4 Methods for blanket control of Legionella in the hot-water system are expensive and not uniformly effective (see below) and we may now be reaching the point where it is more cost-effective to monitor environmental contamination than implement continuous control. Finally, with patients increasingly taking recourse to the law, the high cost of hospital-acquired Legionnaires’ disease might make regular testing more cost-effective. One author has suggested that hot water tanks and selected distal outlets be tested three to four times a year.* We would also suggest that the hot-water systems in areas where patients such as the immunosuppressed, those undergoing head and neck surgery and those with chronic pulmonary disease might warrant regular testing. Detection of legionellae in the environment is now becoming simpler, cheaper and more accurate. The methods available for detecting legionellae in environmental samples fall into three categories: culture, antigen detection and genome detect:ion.2~6~‘5~20~3~36 Quantitative culture usually involves concentration of water samples by centrifugation followed by incubation of the deposit on suitable selective media (e.g. buffered charcoal yeast extract agar with and without cysteine) at 37°C in a humid atmo’sphere for 7-10 days.37 Colonies with the appropriate morphology are then serotyped with specific antisera. Although this is taken as the ‘bench-mark’ methodology it is rather slow and can be rendered imprecise by the presence of biocides and other bacteria that produce inhibitory compounds. 38 Another problem is the presence of non-culturable L. pneumophila. l5 These are bacteria that are visible by indirect or direct immunofluorescence but not recoverable on culture. It is possible that this is a form of ‘heat shock’ and it has been shown that if appropriate resuscitation procedures are employed the ‘non-culturable’ legionellae can be grown.” Antigen detection is usually by immunofluorescence, which provides quantification. Although enzyme-linked immunosorbent assay (ELISA) systems have been devised, 36,39they are not so readily rendered quantitative, since they will detect both bacterium-associated and soluble antigen. Immunofluorescence can be highly sensitive (94%) but is less specific (52%) compared with culture,35 presumably because of the non-culturable legionellae. i5 Perhaps the use of vital dyes (e.g. that change colour when the bacterial respiratory chain is active) or antisera directed against Legionella heat-shock proteins might improve the specificity of such tests.
C. A. Hart and T. Makin
486
Methods involving genome detection, e.g. DNA hybridization or polymerase chain reaction (PCR) are likely to suffer the same problems of quantification as ELISA, although PCR could determine viability if it were used to detect mRNA. At present it would appear that a combination of culture and antigen detection offer the best method for quantification of legionellae in environmental waters.35 Control
of legionellae
in the hospital
environment
Methods for control of legionellae in cooling towers and evaporative condensers are well-established and involve the use of biocides combined with good engineering practice.40 There are many methods, of varying efficacy, for controlling legionellae in hospital hot water systems. They fall into three categories: physical, chemical and good plumbing practice (Table II).
Physical methods Raising the intermittently elimination2>27 Disadvantages bacterial killing and under-used
Table 1. Physical (a) Heat
output temperature in the hot-water system, either or continually is a well-tried and successful method of and is often used in combination with hyperchlorination. of this method include: cost, risk of scalding and incomplete because of the temperature gradient between the calorifier peripheral outlets. Methods to circumvent this problem
II.
Some methods for control
of legionellae
in hot-water
-Raising calorifier temperatures -Instantaneous steam heat systems -Self-regulating trace heat systems
(b) UV irradiation (c) Sonication (d) Draining and flushing
pipes with
compressed
air
2. Chemical (a) To prevent scale formation (b) To maintain sediment in suspension -Sodium hypochlorite (c) Biocides --Ozone (d) Charcoal filters 3. Good (a) (b) (c) (d) (e) (f)
plumbing practice Maintenance Pumps and calorifiers in series not parallel Elimination of dead-space in calorifiers Removal of dead-legs Regular flushing of outlets. Use of Water Research Centre recommended
components
systems
A review
of Legionella
in hospitals
487
include generation of hot-water at outlets via instantaneous steam heat’ and the use of self-regulating trace heat elements that can maintain water in peripheral outlets at 50”C.41 Legionellae are readily killed by UV irradiation.42 We have been able to demonstrate that fitting a IJV irradiation device to a shower was able to maintain it free from legionellae following initial hyperchlorination (unpublished observation). Interestingly although this device did not eliminate other bacteria such as pseudomonads, no amoebae were found. On cessation of UV treatment amoebae and legionellae reappeared concurrently. Chemical methods These have been discussed extensively elsewhere.6 Problems of their use include: inability to kill bacteria in biofilms, corrosion of pipework and generation of organic chlor,ine compounds. Ozonization of water presents another method of disinfection, and a comparison of various methods for elimination of legionellae has been produced.* Finally, charcoal filters have been shown to be of limited benefit in maintaining dental chair water systems free of legionellae.43 Good plumbing practice Design and maintenance of hot-water systems are of prime importance in controlling legionellae in hospitals.40 For example, elimination of dead spaces by avoiding the use of pumps and calorifiers in parallel, the use of devices to ensure complete water circulation in calorifiers and the elimination of dead legs are all of great importance. We have demonstrated that the fitting of automatic drain valves was ineffective in maintaining a reduction in the number of legionellae in shower water2’ but that regular flushing of the system was highly effective. In conclusion, our understanding of the ecology, microbiology and pathogenesis of legionella infections has increased tremendously over the last decade. However, legiolnellae are still an important cause of hospitaland community-acquired pneumonia. For hospital-acquired Legionnaires’ disease to develop, a series of events is required. Firstly, there must be a way for legionellae to gain access to the hospital’s water system and it is likely that low numbers of legionellae gain access through the mains water supply. Next there must be an amplification step. Legionellae are thermotolerant, being able to survive at rel,atively high temperatures and they are able to multiply in lukewarm water both in sessile and planktonic forms. There must also be a way for the bacteria to gain access to their host’s alveoli. This is usually by the aerosols that are readily produced in hospitals, but may occasionally be due to aspiration. Finally the aerosolized bacteria must lodge in a susceptible host. To prevent nosocomial Legionnaires’ disease one or more of these four stages must be prevented. Efforts to decrease or prevent the amplification step are the most likely to be fruitful.
C. A. Hart and T. Makin
488
References 1. Fraser DW, Tsai TR, Orenstein W et al. Legionnaires’ disease: description of an epidemic of pneumonia. N Engl J Med 1977; 297: 1189-l 197. 2. Yu VL. Nosocomial legionellosis: current epidemiological issues. In: Remington JS, Swartz MN, Eds. Current Clinical Topics in Infectious Diseases. New York: McGraw-Hill 1986: 7: 239-253. of Inquiry. First Report of the Committee of Inquiry into the Outbreak of 3. Committee Legionnaires’ Disease in Stafford in April 1985, London: HMSO 1986. legionellosis in surgical patients with 4. Johnson JT, Yu VL, Best MG et al. Nosocomial head and neck cancer: implications for epidemiological reservoir and mode of Lancet 1985; 2: 298-300. transmission. RE, Makin T, Scott MH, Hart CA. Legionella pneumophila serogroup 12 5. Meigh pneumonia in a renal transplant recipient: case report and environmental observations. J
Hosp Infect 1989; 13: 315-319. CLR, Macrae AD, Macfarlane JT. Legionella Infections. London: Edward 6. Bartlett Arnold 1986; 37-55. 7. Durand S. Darnecour C. Iacotot B. A case of Legionnaires’ disease with pericardial involvement. Nouv Press; Med 1984; 13: 1516. L, Roessler BJ, Redd SC, Markowitz LE, Cohen ML. Legionella prosthetic 8. Tompkins valve endocarditis. N EnglJ Med 1988; 318: 530-535. 9. Ampel NM, Ruben FL, Norden CW. Cutaneous abscess caused by Legionella micdadei in an immunosunnressed natient. Ann Intern Med 1985: 102: 630-632. 10. Plouffe JF, Pa& -MF, Maher WE, Hackman B, Webster C. Subtypes of Legionellae pneumophila serogroup 1 associated with different attack rates. Lancet 1983, 2: 649-650. HV. Survival of the Legionnaires’ disease bacterium m water. Ann 11. Skaliy P, McEachern
Intern Med 1979; 90: 662-663. PJ, Lee JV. Differences in aerosol survival between pathogenic and 12. Dennis 1. J Appl Bacterial 1988; non-pathogenic strains of Legionella pneumophilia serogroup 64: 135-141. 13. Editorial. Legionella and amoebae. Lancet 1981; 1: 703-704. M, Calef E, Gitler C. The role of the plasma membrane in the response of 14. Mogyoros parasites to stress. In: McAdam KPWJ, Ed. N ew Strategies in Parasitology. Edinburgh: Churchill Livingstone 1989; 77-94. JS, Dennis PJ. The ecology and survival of Legionella pneumophila. J Inst 15. Colbourne
Water Environ Manag 1989; 3: 345-350. PHLS Report to the DHSS on a collaborative study of Legionella species in water systems 1981-l 985. Colindale, London: Public Health Laboratory Service 1985. Legionnaires’ disease: airborne M, Davis JP et al. Epidemic 17. Band JD, LaVenture transmission down a chimney. JAMA 1981; 245: 2404-2407. the risk of Legionnaires’ disease. Br MedJ 1988; 296: 1343-1344. 18. Finch R. Minimizing 16. Anon.
in the hospital 19. Best M. Yu VL. Stout I, Goetz A. Mulder RR. Taylor F. Legionellaceae water supply. Lancet 1983; 2: 307-310. MS et al. Legionnaires’ disease in a transplant unit: 20. Tobin JO’H, Beare J, Dunnill isolation of a causative agent from shower baths. Lancet 1980; 2: 118-l 21. T, Hart CA. The efficacy of control measures for eradicating legionellae in 21. Makin showers. 7 Hasp Infect 1990: 16: l-7. 22. Verissimo A, Vesey 6, Rocha GM et al. A hot water supply as the source of Legionella pneumophila in incubators of a neonatal unit. J Hosp Infect 1990; 15: 255-263. CJ, Pot FM, Meenhorst PL. Effect of rubbers and their constituents on 23. Niedeveld proliferation of Legionella pneumophila in naturally contaminated hot water. Lancet 1986: 2: 1X&184. TJ, Rendtorff RC, Mallison GF et al. An outbreak of Legionnaires’ disease 24. Dondero associated with a contaminated air conditioning cooling tower. N Ennl.7 - - Med 1980; 302: 365-370. RF, Cozen W, Fields BS et al. Role of air sampling in investigation of an 25. Brieman outbreak of Legionnaires’ disease associated with exposure to aerosols from an evaporative condenser. J Infect Dis 1990; 161: 1257-1261. 26. Tobin JO’H, Bartlett CLR, Waitkins SA et al. Legionnaires’ disease: further evidence to implicate water storage and distribution systems as sources. Br MedJ 1981; 282: 573.
A review
of Legionella
in hospitals
489
27. Fisher-Hoch SP, Bartlett CLFI, Tobin JO’H et al. Investigation and control of an outbreak of Legionnaires’ disease in a district general hospital. Lancer 1981; 1: 932-936. 28. Anon. Second Report of the Committee of Inquiry into the Outbreak of Legionnaires’ Disease in Stafford in April 1985, London: HMSO 1987. 29. Brennen C, Vickers RM, Yu RL, Puntereri A, Yee YC. Discovery of occult Legionella pneumonia in a long stay hospital: results of prospective serological survey. Br Med J 1987; 295: 306307. 30. Baskerville A, Fitzgeorge RB, Broster M, Hambleton P, Dennis PJ. Experimental transmission of Legionnaires’ disease by exposure to aerosols of Legionella pneumophila. Lancet 1981; 2: 138991390. 31. Hambleton P, Broster MG, Dennis PJ et al. Survival of virulent LegionelZapneumophila in aerosols. J Hyg (Camb) 1983; 90: 451460. 32. Fenstersheib MD, Miller M, Diggins C et al. Outbreak of Pontiac fever due to Legionella anisa. Lancet 1990; 336: 35-37. 33. Editorial. Should hospital water be checked for Legionella? Hosp Infect Control 1983; 10: 125-129. 34. Colbourne JS, Pratt DJ, Smith MG, Fisher-Hoch SP, Harper D. Water fittings as sources of Legionella pneumophila in a hospital plumbing system. Lancet 1984; 1: 210-213. 35. Makin T, Hart CA. Detection of Legionellapneumophila in environmental water samples using a fluorescein conjugated monoclonal antibody. Epidemiol Infect 1989; 103: 105-l 12. 36. Tang PW, Toma S. Broad-spectrum enzyme-linked immunosorbent assay of Legionella soluble antigens. J Clin Microbial 1986; 24: 556-558. 37. Edelstein PH. Improved semi-selective medium for isolation of Legionella pneumophila from contaminated clinical and environmental specimens. J Clin Microbial 1981; 14: 298-303. 38. Makin T. Inhibition of Legionella by other organisms. Abstracts of papers presented at the 18th Triennial Conference. Southampton: Institute of Medical Laboratory Sciences 1986; 9. 39. Walker RA. Testing for Legionella bacteria. Occup Health Rev 1988; 16: 17-18. 40. Anon. The Control of Legionellae in Health Care Premises: A Code of Practice. London: HMSO 1988. 41. Makin T, Hart CA. The effect of a self-regulating trace heating element on legionella within a shower. J Appl Bacterial 1991; 70: 258-264. 42. Knudson G. Photoinactivation. of UV irradiated Legionella pneumophila and other Legionella species. Appl Environ Microbial 1985; 49: 975979. 43. Pankhurst CL, Philpott-Howa.rd JN, Hewitt JH, Casewell MW. The efficacy of chlorination and filtration in the control and eradication of Legionella from dental chair water systems. J Hosp Infect 1990; 16: 9-18.
(1991)
Legionella
18 (Supplement
in hospitals:
C. A. Hart Department
of Medical
A), 481-489
a review
and T. Makin
Microbiology, University of Liverpool, Liverpool L69 3BX, UK
P.O. Box 147,
Summary: Although epidemics of nosocomial Legionnaires’ disease attract great attention, up to 30% of sporadic cases of hospital-acquired pneumonia are caused by legionellae. Legionellae are ubiquitous contaminants of potable water and can achieve high numbers in the hot-water systems of large buildings such as hospitals. They are present in the mains water supply in small numbers but are amplified considerably in the hospital’s hot-water system. This is encouraged by water temperatures below SO”C, areas of stagnation and sludge formation, the presence of amoebae and other bacteria and the materials used in the piping. Formation of aerosols from contaminated water is a major mode of spread of legionellae, but there is evidence to suggest that aspiration is also a mode of entry. Safe levels of legionellae in cooling towers have been defined, but not for hot-water systems. A combination of culture and antigen detection by immunofluorescence offer the best method for enumerating legionellae in environmental samples. Control involves a mixture of physical (heat, UV irradiation, sanitation) and chemical (hypochlorite, ozone) methods combined with good plumbing practice (e.g. arrangement of pumps and calorifiers, elimination of dead-‘legs). Adequate control can be costly and requires considerable attention to detail. Keywords:
Legionellae;
nosocomial
infection;
hospital
reservoirs.
Introduction Since the initial description of an outbreak of pneumonia associated with an American Legion convention in Philadelphia in 1976l there has been great interest in legionellae in both the scientific and popular press. Although the initial outbreak was linked to a hotel, it is now clear that hospital-associated Legionnaires’ disease, both epidemic2B3 and sporadic4,’ is not uncommon. The infection generally presents as either a severe pneumonia with or without extrapulmonary manifestation or a milder ‘flu-like’ febrile illness (Pontiac fever),‘j but infection at other sites can occur. Such infections include pericarditis,’ endocarditis’ and cutaneous abscesses.’ In order for legionellae to produce disease, several factors must be considered. Firstly, because human-to-human transmission does not occur, there must be an environmental reservoir. Secondly there must be a mechanism for getting the bacterium from the reservoir to the site of Correspondence 0195%6701/91/06A481+09
to: Professor C. A. Hart. $03.00/O
0 1991 The Hospital
481
Infection
Society
482
C. A. Hart and T. Makin
infection and thirdly the organism must lodge in a susceptible host. In this review we will discuss the properties of legionellae, their reservoirs and modes of spread in hospitals and methods of detection and control of legionellae in the hospital environment. The bacterium
There are currently 14 recognized serogroups of Legionella pneumophila and at least 29 other Legionella species (Table I) although there is still some disagreement over speciation. Not all of these bacteria have been associated with disease in man, and even within L. pneumophila serogroup 1 there appear to be subtypes of differing virulence.” Legionellae are not thermophilic but are thermotolerant, i.e. they are able to survive at relatively high temperatures. For example, legionellae will multiply over the temperature range 2CM3”C and can survive for varying periods at temperatures 40-60°C. At SO”C, 90% of the organisms are killed within 2 h. Legionellae can survive in water for over a year.” Legionellae are able to survive aerosolized in droplets 1-5 pm in diameter and there is evidence to suggest that virulent strains survive longer.‘* Legionellae appear to be able to adhere to substrata such as pipes, rubber, plastics and sediment and can thus persist even when there is constant flushing of a piped-water system. It is possible that they exist in biofilms with release of planktonic organisms. This may explain the discrepancy between the ability of disinfectants to kill legionellae in the test tube and that in piped water systems. Finally, there is some evidence that legionellae are engulfed by amoebae and are able to survive and multiply inside them.13 This might also be a mechanism prolonging survival of legionellae in water. It is also noteworthy that amoebae constantly generate superoxide,14 thus the ability of
Table Legionella pneumophila L. a&a* L. birminghamensis L. boxernaG* L. cherrii L. cincinattiensis L. dumofii* L. erythra L. feeleii* L. geestiae L. gormanii* L. rhackeliae* L. israelensis L. jamestowniensis L. jordanis*
I. Legionella (serogroups
* Associated with human disease.
pneumophila l-14)*
and Legionella L. L. L. L. L. L. L. L. L. L. L. L. L. L. L.
spp.
londiniensis longbeachi? maceacherni? micdadei* nautarum oakridgensis parisiensis quateriensis rubrilucens sainthelensi santicrucis spiritensis steigwaltii wadsworthii* worsliensis
A review
of Legionella
legionellae to survive in amoebae might in a patient’s alveolar macrophages.
in hospitals
be related to their ability
483
to survive
Reservoirs
contaminants of potable water. Legionella Legionellae are common pneumophila has been isolated from drinking water from various parts of the world with a frequency ranging from 0 to 8.8% and isolation occurs more frequently in warm summer months. is It is probable that the small numbers of legionellae in the incoming water are amplified by the hot water systems in large buildings such as hospitals and hotels. For example, the water systems of over 50% of large establishments have been shown to contain disease have resulted L. pneumophila. i6 Large outbreaks of Legionnaires’ from contamination of air conditioning and ventilation systems,3,4T’7 and legionellae can be found in large numbers in water in cooling towers. The other major site for colonization and multiplication of legionellae is the hot-water system, or other areas where water temperature is raised into the range 2540°C. In many halspitals, hot water is produced by taking water through a number of heaters or calorifiers. The water then circulates throughout the hospital and returns to the calorifiers. The factors promoting the expansion and persistence of legionella populations have been described by others. 2118They include the arrangement of calorifiers, which should be in series rather than parallel, the presence of a dead space at the bottom of the calorifier which allows sludge and legionellae to persist, and maintenance of the hot-water output below 57°C. The presence of dead-legs in the hot-water system, e.g. associated with showers,19-21 hosepipes or fire sprinkler systems, the use of spas or jacuzzis, and the addition of hot water to humidifiers or nebulizers** can all serve to amplify Legionella populations. We have even been able to find legionellae in the melt-water from ward ice-making machines. The pH, ionic composition and conductivity of the water and the nature and age of materials used to make pipes and valves also affect the growth and survival of legionellae. For example, some silicone-based rubbers enhance the growth of L. pneumophila by a factor of 105, whereas rubbers containing the accelerator thiuram inhibit ;growth.23 There has been considerable debate over the relative importance of the two major reservoirs, namely cooling towers and evaporative condensers and the hot-water system. It is clear that episodes of infection have arisen from contamination of water in cooling towers,3,24 and the evidence linking such reservoirs and development of infection is strong. Indeed, in one outbreak of Legionnaires’ disease in a retirement hotel in which L. pneumophila serogroup 1 was found in both the hot-water system and an evaporative condenser, it was demonstrated using air sampling and monoclonal antibody subtyping that the condenser was the source.25 However, it has also been shown that Legionnaires’ disease could occur in
484
C. A. Hart and T. Makin
areas where there was no possibility of aerosols from evaporative condensers.20,26 Similarly outbreaks have been controlled by treatment of the water system alone even though a cooling tower was present.27 It would appear that the two reservoirs are equally important. Guidelines on acceptable levels of legionella contamination of cooling towers have been produced28 but no such guidance is available for hot-water systems. In cooling towers Legionella counts of less than lo3 I-’ have been suggested as being acceptable, those above lo5 1-l indicate that corrective action is required. It has been demonstrated that there is a link between the extent of colonization of a hospital’s hot-water system and the development of nosocomial Legionnaires’ disease. 4,29 Thus, in an Ear Nose and Throat hospital where 67% of sites sampled contained legionellae, 30% of the cases of hospital-acquired pneumonia were due to legionellae. After chlorination of the system only 3.5% of sites were contaminated and none of the cases of hospital-acquired pneumonia were caused by legionellae. The patient group studied were resident on head and neck oncology wards and such patients are known to be highly susceptible to legionella infection. Nevertheless, it appears that once contamination of a hospital’s hot-water system involves more than 30% of sites sampled, nosocomial Legionnaires’ disease occurs.19 Unfortunately, there is little information on the relationship between numbers of legionellae in the hot-water systems and development of infection. However, we were able to demonstrate that a few days prior to acquisition of pneumonia due to L. pneumophila serogroup 12 by a renal transplant recipient, the water from her hand basin hot tap contained 6 x lo3 cfu 1-l of the same bacterium.5 Modes
of spread
The most likely mode of spread is by aerosolized particles < 5 pm diameter containing legionellae. It is possible to infect guinea-pigs with aerosols containing as few as 3.7 X lo4 cfu L. pneumophila serogroup 13’ and the bacteria can survive for up to 2 h in such aerosols.31 Aerosols can be produced easily in evaporative condensers, cooling towers, jacuzzis, showers, nebulizers and in large buildings even by normal tap pressure. There has even been an outbreak of Pontiac fever due to L. anisa where an aerosol created by an artificial fountain was implicated.32 It has also been suggested that aspiration or direct inoculation of legionellae into the trachea-bronchial tree via respiratory manipulation are important modes of transmission especially in head and neck surgery patients.2,4 Detection
Perhaps the environmental
of legionellae
in the environment
most difficult question to answer is should there be testing and if so, how frequently? In general the indication
A review
of Legionella
in hospitals
485
for environmental testing in hospitals is the presence of a cluster of cases of Legionnaires’ disease or an increasing number of sporadic cases. Because legionellae are ubiquitous and because of the high cost, routine surveillance some is not recommended.33 Perhaps in the lig ht of recent developments re-appraisal of this position is necessary. Guidelines have been produced giving acceptable levels of legionellae in cooling towers, in order to comply with these it is thus necessary to measure levels actually present. There does appear to be a relationship between the extent of environmental colonization and development of nosocomial Legionnaires’ disease.4 Methods for blanket control of Legionella in the hot-water system are expensive and not uniformly effective (see below) and we may now be reaching the point where it is more cost-effective to monitor environmental contamination than implement continuous control. Finally, with patients increasingly taking recourse to the law, the high cost of hospital-acquired Legionnaires’ disease might make regular testing more cost-effective. One author has suggested that hot water tanks and selected distal outlets be tested three to four times a year.* We would also suggest that the hot-water systems in areas where patients such as the immunosuppressed, those undergoing head and neck surgery and those with chronic pulmonary disease might warrant regular testing. Detection of legionellae in the environment is now becoming simpler, cheaper and more accurate. The methods available for detecting legionellae in environmental samples fall into three categories: culture, antigen detection and genome detect:ion.2~6~‘5~20~3~36 Quantitative culture usually involves concentration of water samples by centrifugation followed by incubation of the deposit on suitable selective media (e.g. buffered charcoal yeast extract agar with and without cysteine) at 37°C in a humid atmo’sphere for 7-10 days.37 Colonies with the appropriate morphology are then serotyped with specific antisera. Although this is taken as the ‘bench-mark’ methodology it is rather slow and can be rendered imprecise by the presence of biocides and other bacteria that produce inhibitory compounds. 38 Another problem is the presence of non-culturable L. pneumophila. l5 These are bacteria that are visible by indirect or direct immunofluorescence but not recoverable on culture. It is possible that this is a form of ‘heat shock’ and it has been shown that if appropriate resuscitation procedures are employed the ‘non-culturable’ legionellae can be grown.” Antigen detection is usually by immunofluorescence, which provides quantification. Although enzyme-linked immunosorbent assay (ELISA) systems have been devised, 36,39they are not so readily rendered quantitative, since they will detect both bacterium-associated and soluble antigen. Immunofluorescence can be highly sensitive (94%) but is less specific (52%) compared with culture,35 presumably because of the non-culturable legionellae. i5 Perhaps the use of vital dyes (e.g. that change colour when the bacterial respiratory chain is active) or antisera directed against Legionella heat-shock proteins might improve the specificity of such tests.
C. A. Hart and T. Makin
486
Methods involving genome detection, e.g. DNA hybridization or polymerase chain reaction (PCR) are likely to suffer the same problems of quantification as ELISA, although PCR could determine viability if it were used to detect mRNA. At present it would appear that a combination of culture and antigen detection offer the best method for quantification of legionellae in environmental waters.35 Control
of legionellae
in the hospital
environment
Methods for control of legionellae in cooling towers and evaporative condensers are well-established and involve the use of biocides combined with good engineering practice.40 There are many methods, of varying efficacy, for controlling legionellae in hospital hot water systems. They fall into three categories: physical, chemical and good plumbing practice (Table II).
Physical methods Raising the intermittently elimination2>27 Disadvantages bacterial killing and under-used
Table 1. Physical (a) Heat
output temperature in the hot-water system, either or continually is a well-tried and successful method of and is often used in combination with hyperchlorination. of this method include: cost, risk of scalding and incomplete because of the temperature gradient between the calorifier peripheral outlets. Methods to circumvent this problem
II.
Some methods for control
of legionellae
in hot-water
-Raising calorifier temperatures -Instantaneous steam heat systems -Self-regulating trace heat systems
(b) UV irradiation (c) Sonication (d) Draining and flushing
pipes with
compressed
air
2. Chemical (a) To prevent scale formation (b) To maintain sediment in suspension -Sodium hypochlorite (c) Biocides --Ozone (d) Charcoal filters 3. Good (a) (b) (c) (d) (e) (f)
plumbing practice Maintenance Pumps and calorifiers in series not parallel Elimination of dead-space in calorifiers Removal of dead-legs Regular flushing of outlets. Use of Water Research Centre recommended
components
systems
A review
of Legionella
in hospitals
487
include generation of hot-water at outlets via instantaneous steam heat’ and the use of self-regulating trace heat elements that can maintain water in peripheral outlets at 50”C.41 Legionellae are readily killed by UV irradiation.42 We have been able to demonstrate that fitting a IJV irradiation device to a shower was able to maintain it free from legionellae following initial hyperchlorination (unpublished observation). Interestingly although this device did not eliminate other bacteria such as pseudomonads, no amoebae were found. On cessation of UV treatment amoebae and legionellae reappeared concurrently. Chemical methods These have been discussed extensively elsewhere.6 Problems of their use include: inability to kill bacteria in biofilms, corrosion of pipework and generation of organic chlor,ine compounds. Ozonization of water presents another method of disinfection, and a comparison of various methods for elimination of legionellae has been produced.* Finally, charcoal filters have been shown to be of limited benefit in maintaining dental chair water systems free of legionellae.43 Good plumbing practice Design and maintenance of hot-water systems are of prime importance in controlling legionellae in hospitals.40 For example, elimination of dead spaces by avoiding the use of pumps and calorifiers in parallel, the use of devices to ensure complete water circulation in calorifiers and the elimination of dead legs are all of great importance. We have demonstrated that the fitting of automatic drain valves was ineffective in maintaining a reduction in the number of legionellae in shower water2’ but that regular flushing of the system was highly effective. In conclusion, our understanding of the ecology, microbiology and pathogenesis of legionella infections has increased tremendously over the last decade. However, legiolnellae are still an important cause of hospitaland community-acquired pneumonia. For hospital-acquired Legionnaires’ disease to develop, a series of events is required. Firstly, there must be a way for legionellae to gain access to the hospital’s water system and it is likely that low numbers of legionellae gain access through the mains water supply. Next there must be an amplification step. Legionellae are thermotolerant, being able to survive at rel,atively high temperatures and they are able to multiply in lukewarm water both in sessile and planktonic forms. There must also be a way for the bacteria to gain access to their host’s alveoli. This is usually by the aerosols that are readily produced in hospitals, but may occasionally be due to aspiration. Finally the aerosolized bacteria must lodge in a susceptible host. To prevent nosocomial Legionnaires’ disease one or more of these four stages must be prevented. Efforts to decrease or prevent the amplification step are the most likely to be fruitful.
C. A. Hart and T. Makin
488
References 1. Fraser DW, Tsai TR, Orenstein W et al. Legionnaires’ disease: description of an epidemic of pneumonia. N Engl J Med 1977; 297: 1189-l 197. 2. Yu VL. Nosocomial legionellosis: current epidemiological issues. In: Remington JS, Swartz MN, Eds. Current Clinical Topics in Infectious Diseases. New York: McGraw-Hill 1986: 7: 239-253. of Inquiry. First Report of the Committee of Inquiry into the Outbreak of 3. Committee Legionnaires’ Disease in Stafford in April 1985, London: HMSO 1986. legionellosis in surgical patients with 4. Johnson JT, Yu VL, Best MG et al. Nosocomial head and neck cancer: implications for epidemiological reservoir and mode of Lancet 1985; 2: 298-300. transmission. RE, Makin T, Scott MH, Hart CA. Legionella pneumophila serogroup 12 5. Meigh pneumonia in a renal transplant recipient: case report and environmental observations. J
Hosp Infect 1989; 13: 315-319. CLR, Macrae AD, Macfarlane JT. Legionella Infections. London: Edward 6. Bartlett Arnold 1986; 37-55. 7. Durand S. Darnecour C. Iacotot B. A case of Legionnaires’ disease with pericardial involvement. Nouv Press; Med 1984; 13: 1516. L, Roessler BJ, Redd SC, Markowitz LE, Cohen ML. Legionella prosthetic 8. Tompkins valve endocarditis. N EnglJ Med 1988; 318: 530-535. 9. Ampel NM, Ruben FL, Norden CW. Cutaneous abscess caused by Legionella micdadei in an immunosunnressed natient. Ann Intern Med 1985: 102: 630-632. 10. Plouffe JF, Pa& -MF, Maher WE, Hackman B, Webster C. Subtypes of Legionellae pneumophila serogroup 1 associated with different attack rates. Lancet 1983, 2: 649-650. HV. Survival of the Legionnaires’ disease bacterium m water. Ann 11. Skaliy P, McEachern
Intern Med 1979; 90: 662-663. PJ, Lee JV. Differences in aerosol survival between pathogenic and 12. Dennis 1. J Appl Bacterial 1988; non-pathogenic strains of Legionella pneumophilia serogroup 64: 135-141. 13. Editorial. Legionella and amoebae. Lancet 1981; 1: 703-704. M, Calef E, Gitler C. The role of the plasma membrane in the response of 14. Mogyoros parasites to stress. In: McAdam KPWJ, Ed. N ew Strategies in Parasitology. Edinburgh: Churchill Livingstone 1989; 77-94. JS, Dennis PJ. The ecology and survival of Legionella pneumophila. J Inst 15. Colbourne
Water Environ Manag 1989; 3: 345-350. PHLS Report to the DHSS on a collaborative study of Legionella species in water systems 1981-l 985. Colindale, London: Public Health Laboratory Service 1985. Legionnaires’ disease: airborne M, Davis JP et al. Epidemic 17. Band JD, LaVenture transmission down a chimney. JAMA 1981; 245: 2404-2407. the risk of Legionnaires’ disease. Br MedJ 1988; 296: 1343-1344. 18. Finch R. Minimizing 16. Anon.
in the hospital 19. Best M. Yu VL. Stout I, Goetz A. Mulder RR. Taylor F. Legionellaceae water supply. Lancet 1983; 2: 307-310. MS et al. Legionnaires’ disease in a transplant unit: 20. Tobin JO’H, Beare J, Dunnill isolation of a causative agent from shower baths. Lancet 1980; 2: 118-l 21. T, Hart CA. The efficacy of control measures for eradicating legionellae in 21. Makin showers. 7 Hasp Infect 1990: 16: l-7. 22. Verissimo A, Vesey 6, Rocha GM et al. A hot water supply as the source of Legionella pneumophila in incubators of a neonatal unit. J Hosp Infect 1990; 15: 255-263. CJ, Pot FM, Meenhorst PL. Effect of rubbers and their constituents on 23. Niedeveld proliferation of Legionella pneumophila in naturally contaminated hot water. Lancet 1986: 2: 1X&184. TJ, Rendtorff RC, Mallison GF et al. An outbreak of Legionnaires’ disease 24. Dondero associated with a contaminated air conditioning cooling tower. N Ennl.7 - - Med 1980; 302: 365-370. RF, Cozen W, Fields BS et al. Role of air sampling in investigation of an 25. Brieman outbreak of Legionnaires’ disease associated with exposure to aerosols from an evaporative condenser. J Infect Dis 1990; 161: 1257-1261. 26. Tobin JO’H, Bartlett CLR, Waitkins SA et al. Legionnaires’ disease: further evidence to implicate water storage and distribution systems as sources. Br MedJ 1981; 282: 573.
A review
of Legionella
in hospitals
489
27. Fisher-Hoch SP, Bartlett CLFI, Tobin JO’H et al. Investigation and control of an outbreak of Legionnaires’ disease in a district general hospital. Lancer 1981; 1: 932-936. 28. Anon. Second Report of the Committee of Inquiry into the Outbreak of Legionnaires’ Disease in Stafford in April 1985, London: HMSO 1987. 29. Brennen C, Vickers RM, Yu RL, Puntereri A, Yee YC. Discovery of occult Legionella pneumonia in a long stay hospital: results of prospective serological survey. Br Med J 1987; 295: 306307. 30. Baskerville A, Fitzgeorge RB, Broster M, Hambleton P, Dennis PJ. Experimental transmission of Legionnaires’ disease by exposure to aerosols of Legionella pneumophila. Lancet 1981; 2: 138991390. 31. Hambleton P, Broster MG, Dennis PJ et al. Survival of virulent LegionelZapneumophila in aerosols. J Hyg (Camb) 1983; 90: 451460. 32. Fenstersheib MD, Miller M, Diggins C et al. Outbreak of Pontiac fever due to Legionella anisa. Lancet 1990; 336: 35-37. 33. Editorial. Should hospital water be checked for Legionella? Hosp Infect Control 1983; 10: 125-129. 34. Colbourne JS, Pratt DJ, Smith MG, Fisher-Hoch SP, Harper D. Water fittings as sources of Legionella pneumophila in a hospital plumbing system. Lancet 1984; 1: 210-213. 35. Makin T, Hart CA. Detection of Legionellapneumophila in environmental water samples using a fluorescein conjugated monoclonal antibody. Epidemiol Infect 1989; 103: 105-l 12. 36. Tang PW, Toma S. Broad-spectrum enzyme-linked immunosorbent assay of Legionella soluble antigens. J Clin Microbial 1986; 24: 556-558. 37. Edelstein PH. Improved semi-selective medium for isolation of Legionella pneumophila from contaminated clinical and environmental specimens. J Clin Microbial 1981; 14: 298-303. 38. Makin T. Inhibition of Legionella by other organisms. Abstracts of papers presented at the 18th Triennial Conference. Southampton: Institute of Medical Laboratory Sciences 1986; 9. 39. Walker RA. Testing for Legionella bacteria. Occup Health Rev 1988; 16: 17-18. 40. Anon. The Control of Legionellae in Health Care Premises: A Code of Practice. London: HMSO 1988. 41. Makin T, Hart CA. The effect of a self-regulating trace heating element on legionella within a shower. J Appl Bacterial 1991; 70: 258-264. 42. Knudson G. Photoinactivation. of UV irradiated Legionella pneumophila and other Legionella species. Appl Environ Microbial 1985; 49: 975979. 43. Pankhurst CL, Philpott-Howa.rd JN, Hewitt JH, Casewell MW. The efficacy of chlorination and filtration in the control and eradication of Legionella from dental chair water systems. J Hosp Infect 1990; 16: 9-18.