Detection of Legionella by quantitative-polymerase chain reaction (qPCR) for monitoring and risk assessment
- Louise H Krøjgaard†1, 2Email author,
- Karen A Krogfelt†1,
- Hans-Jørgen Albrechtsen†2 and
- Søren A Uldum†1Email author
© Krøjgaard et al; licensee BioMed Central Ltd. 2011
Received: 22 June 2011
Accepted: 21 November 2011
Published: 21 November 2011
Culture and quantitative polymerase chain reaction (qPCR) assays for the detection of Legionella were compared on samples from a residential area before and after two interventions. A total of 84 samples were collected from shower hoses and taps as first flush samples and at constant temperature. Samples were grouped according to the origin of the sample, a) circulation water b) water from empty apartments c) water from shower hoses. The aims were to investigate the usefulness of qPCR compared to culture for monitoring remedial actions for elimination of Legionella bacteria and as a tool for risk assessment.
In water collected from the apartments Legionella spp were detected by qPCR in the concentration range from LOQ to 9.6*105GU/L while L. pneumophila were detected in a range from LOQ to 6.8*105 GU/L. By culturing, the legionellae were detected in the range from below detection limit (> 10 CFU/L) to 1.6*106 CFU/L. In circulating water and in first flush water from shower hoses, culture and qPCR showed the same tendencies. The overall correlation between the bacteria number detected by culture and the two developed qPCR assays (L. spp and L. pneumophila) was relatively poor (r2 = 0.31 for culture and Legionella spp. assay, r2 = 0.20 for culture and L. pneumophila assay).
Detection by qPCR was suitable for monitoring changes in the concentration of Legionella but the precise determination of bacteria is difficult. Risk assessment by qPCR only on samples without any background information regarding treatment, timing, etc is dubious. However, the rapid detection by qPCR of high concentrations of Legionella - especially Legionella pneumophila - is valuable as an indicator of risk, although it may be false positive compared to culture results. On the other hand, the detection of a low number of bacteria by qPCR is a strong indication for the absence of risk.
Legionella bacteria are ubiquitous in nature and are often found in natural water sources as well as in man-made water systems. Humans may be infected through inhalation of contaminated aerosolised water droplets. Symptoms range from influenza-like disease (Pontiac fever) to severe pneumonia (Legionnaires' disease, LD) with a high mortality rate [1, 2]. More than 50 Legionella species are described, but Legionella (L.) pneumophila is the cause of more than 95% of LD cases [3, 4]. High concentrations (104-1010 Legionella CFU/L) of Legionella in the water sources are considered a risk of infection [5–8]. Being able to determine the concentration of Legionella in water is, therefore, highly relevant in risk assessments and transmission tracing. The reference method for enumeration of Legionella in water is culture . Culture is, however, hampered by a long incubation time (7 to 10 days) whereas qPCR can be performed within three hours. By culture, only bacteria cultivable under the given conditions can be quantified in environmental samples with mixed cultures of different bacteria including different Legionella species. Quantitative detection of L. pneumophila (which is the most significant Legionella species for risk assessment) is difficult by culture. In qPCR, with specific L. pneumophila primers, only L. pneumophila will be amplified irrespectively of background flora etc. The important bias of qPCR compared to culture is that also dead and otherwise not culturable Legionella will be quantified.
The aim of this work was to clarify under which circumstances and in which samples qPCR is useful for monitoring and risk assessment. The investigation was performed in a newly built residential area where two males contracted LD.
The sampling area and the interventions
A newly built residential area associated with two cases was investigated . The area consisted of 225 apartments distributed on 6 blocks; around 210 apartments were inhabited at the time of the sampling period. The two cases and the interventions done to overcome the Legionella colonisation and the risk factors found to be associated with the residential area were published previously .
Two interventions were conducted to control the Legionella contamination of the hot water system. The first was a 12 h heat treatment of the boilers (approximately 70°C) together with a request to all residents to flush their taps for 5 minutes. Subsequently, the water in the boilers was completely replaced with fresh water and the temperature was lowered to 60°C. Circulation pumps were set at maximum flow.
As the first intervention did not reduce the Legionella count satisfactorily, a second intervention was performed three weeks later, which consisted of an increase of the water temperature in the boilers to approximately 70°C for 24 hours. During this time, all taps were flushed for 5 min. The boilers were hyperchlorinated and the temperature was set at 65°C. All shower hoses were replaced with new ones in all apartments, and over the next month the boiler temperature was regulated to ensure that the water in the most distant taps was kept at > 50°C. Taps of empty apartments were flushed weekly for 5 min with water from the hot water taps.
Sampling and analysis
Water samples (a total of 84) were collected at ten occasions from January 9, 2009 to September 7, 2009: One sampling before any of the interventions, three samplings between the first and the second interventions and six samplings after the second intervention.
Comparison of culture and qPCR for quantification of Legionella.
Number of positive samples
Type of sample
No of samples
Legionella spp 104 CFU/L
Legionella spp104 GU/L
Legionella pneumophila 104 GU/L
Before the first intervention
0.005 - 1.2
0.77 - 2.9
0.6 - 2.6
After the first intervention
1.9 - 33
2.9 - 24
4.9 - 19
0.8 - 160
3.5 - 96
1.1 - 43
BD - 2.0
After the second intervention
BD - 0.001
3.2 - 55
3.7 - 68
0.17 - 2.3
0.033 - 3.2
Culture and extraction of DNA for qPCR
Culture procedure followed the ISO standard 11731-2: 2006 on both MWY (Modified Wadowsky Yee) (Oxoid, Greve, Denmark) and GVPC (Glycine, Vancomycin, polymyxin, Cycloheximide) (Oxoid, Greve, Denmark) agar plates and based on three different concentration steps. DNA extraction was performed from a 100 fold concentration of the water samples, with Chelex®100 (Bio-Rad California, USA) (900 μL sample, 150 μL Chelex®100) before qPCR. Culturing of samples was previously described in detail in Krøjgaard et al (2011) 
qPCR Legionella species and Legionella pneumophilaassay
qPCR was performed with primers and a probe detecting Legionella species (targeting the 5S rRNA gene) and primers and a probe detecting L. pneumophila (the mip gene); both primers and probes were optimized to a TaqMan assay. Internal process controls (IPCs) for Legionella spp. and L. pneumophila were included in order to assess inhibition or suboptimal reaction conditions. The IPC was co-amplified in every qPCR reaction together with target DNA .
Standard dilution was correlated to the French Legionella pneumophila DNA standard (SMR_ LEGDNA_01 standard from Legionelles centre National de Référence, Lyon, France). The limit of detection (LOD) was 2.5 GU/reaction (5 μL) for the Legionella species assay corresponding to 833 GU/L. The limit of quantification (LOQ) was 10 GU/5 μL and 3333 GU/L. The LOD for the L. pneumophila assay was 15 GU/5 μL / 5000 GU/L and the LOQ was 25 GU/5 μL/8333 GU/L.
Results & discussion
Overall correlation between qPCR and culture
Legionella species were detected by qPCR in all 84 water samples. Four samples were below LOD. L. pneumophila were detected in 75 of the 84 samples, in 34 samples below LOQ. Forty-tree of the 84 samples were found positive by culture. The amount found by culture and qPCR for all (84 samples) did not correlate well (r2 = 0.31 L. species assay, r2 = 0.20 L. pneumophila assay). Poor correlations were also found in others investigations comparing culture and qPCR results for samples from hot water systems [12–15]. qPCR amplifies DNA from both living and dead Legionella still harbouring the DNA. An effective heat treatment would, therefore, not necessarily significantly change the amount detected by qPCR in the short term, as long as the cells not had lost their DNA. When Legionella DNA is still in the water system, it can be amplified by qPCR. By culture depending on living and culturable bacteria, no or only limited growth would be expected after effective heat treatment. This effect of temperature is supported by Lee et al (2011)  who found a significantly higher mean log difference comparing culture and qPCR at temperatures above 50°C than at lower temperatures. Comparison of the methods for samples collected from water systems where no interventions have been conducted, would probably give a better agreement between the two methods.
After the first intervention, Legionella was found in lower concentrations (< 10*103 Legionella CFU/L) in six out of ten samples. The four other samples were in the range 1.2*103 - 1.2*104 Legionella CFU/L. The range measured by qPCR after the first intervention (both assays) was from 6.0*103 to 2.9*104 GU/L (Table 1).
After the second intervention, no legionellae were detected by culture, but the range found by qPCR for the L. species assay was 4.0*103 to 1.9*104 GU/L with an median of 6.2*103 GU/L. For the L. pneumophila assay, three samples were negative, but the 13 other samples were positive ranging from 6.7*102 to 2.0*104 GU/L (Table 1).
The second intervention seemed to kill or make Legionella uncultivable but the results from qPCR showed that they were still present in the system as dead or uncultivable bacteria. There was no obvious difference between the amount detected just after the second intervention and seven months after measuring Legionella species. By qPCR, the amount of L. pneumophila was found to decrease slightly with time. The ranges in which Legionella were detected before and after the second intervention measured by qPCR on circulation water samples were overlapping. Therefore, it is difficult to draw conclusions on the effect of the remedial actions and to form a picture of the risk using the distinct values from circulation water provided by qPCR; however, trends or tendencies can be detected.
First flush from empty apartments
Regular flushing seemed to control the level of live Legionella in the distant parts of the pipes in the empty apartments over a long period, but this could not be demonstrated with qPCR. Sudden opening of the tap could probably flush out biofilm with dead Legionella which could be detected by qPCR but not by culture. It can be concluded that qPCR could not be used for risk assessment or for monitoring the effect of the remedial actions on stagnant water in an empty apartment. It should be noted that only water from one apartment was sampled after the second intervention.
First flush from shower hoses
The conclusion for samples from shower hoses is the same as for circulation water. qPCR is suitable for monitoring a change in the bacterial concentration, whereas the specific number of bacteria is difficult to use for risk assessment.
Overall, when using qPCR, background information on the system from where samples have been collected is helpful in the interpretation of the results. Knowledge of any treatments of the water and the temperature profile of the system is essential for the interpretation. If water is collected at a temperature known to support the growth of Legionella, and no disinfection has been conducted, the high amount of Legionella found by qPCR is also found by culture. If disinfection of some kind was used it is more difficult to correlate results from the two methods since some or all Legionella could have been killed. However, on some occasions it could be interesting to monitor the level of dead or unculturable Legionella, since a high level measured by qPCR could indicate a current or recent colonisation of the system, which could indicate a potential risk even though the bacteria do not grow.
As also discussed in Joly et al 2006  a negative or low level of Legionella detected by qPCR is a quite good predictor of a negative culture result. Unfortunately, this selection is difficult to establish based on detection of Legionella species since all tested samples were found to contain Legionella DNA. Using the Legionella pneumophila assay, eight of ten samples found negative by qPCR were also negative by culture.
It has been suggested to improve the usefulness of qPCR by pre-treatment with the DNA-dye Propidium monoazide to discriminate between dead and live bacteria . Previous work with dying DNA of membrane compromised cells focused on the use of the dye ethidium monoazide  but Propidium monoazide has been found to show less cytotoxicity . Nevertheless, optimization of the use of the dyes is still needed.
We found that detection of Legionella in water samples by qPCR was suitable for monitoring changes in the concentration of Legionella over time, whereas the specific number measured by qPCR was difficult to use for risk assessment. Results for both culture and qPCR followed the same decreasing tendencies for circulating water and first flush water samples from shower hoses. In first flush samples from empty apartments, before the second intervention, culture and qPCR results were generally at the same level, but the two samples collected after the second intervention showed different tendencies with the two methods.
Background information about the water system is necessary to interpret the qPCR results, but low amounts of Legionella pneumophila detected by qPCR is a good indicator of low risk, and detection of high levels in untreated water systems is a good indicator of colonisation and risk.
Acknowledgements and Funding
We would like to thank laboratory technicians of the Dept. of Microbiological Diagnostics, Statens Serum Institut, Berit Larsen, Bente Tangvig and Gitte H. Riisgaard for their practical help with the culturing of water samples.
Louise Hjelmar Krøjgaard was partly financially supported by the Graduate School UrbanWaterTech. All authors declare no conflicts of interest.
Parts of the results have been presented as a poster at the 25th European working group for Legionella Infections meeting in Copenhagen, Denmark, 15-17 September 2010.
- Rosa F: Legionnaires' Disease prevention and Control. 1993, Troy Michigan: Business News Publishing CompanyGoogle Scholar
- Fliermans CB, Cherry WB, Orrison LH, Smith SJ, Tison DL, Pope DH: Ecological distribution of Legionella pneumophila. Appl Environ Microbiol. 1981, 41 (1): 9-16.PubMedPubMed CentralGoogle Scholar
- Bartram J, Chartier Y, Lee JV, Pond K, Surman-Lee S, (editors): Legionella and prevention of legionellosis. World Health Organization. 2007
- Joseph CA, Ricketts KD: Legionnaires disease in Europe 2007-2008. Euro Surveill. 2010, 15 (8):
- Ferre MR, Arias C, Oliva JM, Pedrol A, Garcia M, Pellicer T, Roura P, Dominguez A: A community outbreak of Legionnaires' disease associated with a cooling tower in Vic and Gurb, Catalonia (Spain) in 2005. Eur J Clin Microbiol Infect Dis. 2009, 28 (2): 153-159. 10.1007/s10096-008-0603-6.PubMedView ArticleGoogle Scholar
- Borgen K, Aaberge L, Werner-Johansen O, Gjosund K, Storsrud B, Haugsten S, Nygard K, Krogh T, Høiby EA, Caugant DA, Kanestrøm A, Simonsen Ø, Blystad H: Cluster of Legionnaires disease linked to an industrial plant in southeast Norway, June - July 2008. Euro Surveill. 2008, 13 (38):
- Castilla J, Barricarte A, Aldaz J, Garcia CM, Ferrer T, Pelaz C, Pineda S, Baladron B, Martin I, Goni B, Aratajo P, Chamorro J, Lameiro F, Torroba L, Dorronsoro L, Martinez-Artola V, Esparza MJ, Gastaminza MA, Fraile P, Aldaz P: A large Legionnaires' disease outbreak in Pamplona, Spain: early detection, rapid control and no case fatality. Epidemiol Infect. 2008, 136 (6): 823-832.PubMedPubMed CentralGoogle Scholar
- Rota MC, Caporali MG, Massari M: European Guidelines for Control and Prevention of Travel Associated Legionnaires' Disease: the Italian experience. Euro Surveill. 2004, 9 (2):
- ISO 11731-2:2006 Dansk Standard Water quality-Detection and enumeration of Legionella-Part 2: Direct membrane filtration method for waters with low bacterial counts.
- Krojgaard LH, Krogfelt KA, Albrechtsen HJ, Uldum SA: Cluster of Legionnaires disease in a newly built block of flats, Denmark, December 2. Euro Surveill. 2011, 16 (1):
- Jensen JS, Borre MB, Dohn B: Detection of Mycoplasma genitalium by PCR amplification of the 16S rRNA gene. J Clin Microbiol. 2003, 41 (1): 261-266. 10.1128/JCM.41.1.261-266.2003.PubMedPubMed CentralView ArticleGoogle Scholar
- Bonetta S, Bonetta S, Ferretti E, Balocco F, Carraro E: Evaluation of Legionella pneumophila contamination in Italian hotel water systems by quantitative real-time PCR and culture methods. J Appl Microbiol. 2010, 108 (5): 1576-1583. 10.1111/j.1365-2672.2009.04553.x.PubMedView ArticleGoogle Scholar
- Wellinghausen N, Frost C, Marre R: Detection of legionellae in hospital water samples by quantitative real-time LightCycler PCR. Appl Environ Microbiol. 2001, 67 (9): 3985-3993. 10.1128/AEM.67.9.3985-3993.2001.PubMedPubMed CentralView ArticleGoogle Scholar
- Joly P, falconnet P-A, André J, Weill N, Reyrolle M, Vandenesch F, Maurin M, Etienne J, Jarraud S: Quantitative Real-Time Legionella PCR for environmental water samples:Data interpretation. Appl Environ Microbiol. 2006, 72 (4): 2801-2808. 10.1128/AEM.72.4.2801-2808.2006.PubMedPubMed CentralView ArticleGoogle Scholar
- Yanez MA, Carrasco.Serrano C, Barberá VM, Catalán V: Quantitative detection of Legionella pneumophila in water samples by immunomagnetioc purification and real-time PCR amplification of the dotA gene. Appl Environ Microbiol. 2005, 71 (7): 3433-3441. 10.1128/AEM.71.7.3433-3441.2005.PubMedPubMed CentralView ArticleGoogle Scholar
- Lee JV, Lai S, Exner M, Lenz J, Gaia V, Casati S, Hartemann P, Luck C, Pangon B, Ricci ML, Scaturro M, Fontana S, Sabria M, Sanchez I, Assaf S, Surman-Lee S: An international trial of quantitative PCR for monitoring Legionella in artificial water systems. J Appl Microbiol. 2011, 110 (4): 1032-1044.PubMedPubMed CentralView ArticleGoogle Scholar
- Walker JT, Mackerness CW, Mallon D, Makin T, Williets T, Keevil CW: Control of Legionella pneumophila in a hospital water system by chlorine dioxide. J Ind Microbiol. 1995, 15 (4): 384-390. 10.1007/BF01569995.PubMedView ArticleGoogle Scholar
- Yanez MA, Nocker A, Soria-Soria E, Murtula R, Martinez L, Catalan V: Quantification of viable Legionella pneumophila cells using propidium monoazide combined with quantitative PCR. J Microbiol Methods. 2011, 85 (2): 124-130. 10.1016/j.mimet.2011.02.004.PubMedView ArticleGoogle Scholar
- Nogva HK, Drømtorp SM, Nissen H, Rudi K: Ethidium monoazide for DNA-based differentiation of viable and dead bacteria by 5'nuclease PCR. Biotechniques. 2003, 34 (4): 804-808.PubMedGoogle Scholar