Water Stagnation and Flow Obstruction Reduces the Quality of Potable Water and Increases the Risk of Legionelloses

Legionella is an opportunistic waterborne pathogen associated with Legionnaires' disease and Pontiac fever. Despite improved public awareness, the incidence of Legionella associated infections has been increasing. Aerosols generated from engineered potable water systems are a demonstrated cause of both nosocomial and community-acquired legionellosis. The ecology of Legionella in these systems is complex with multiple factors impacting their colonization and persistence. Flow dynamics has been identified as an important factor and stagnation in cooling towers is an accepted risk for increased Legionella growth; however, less is known about the impact of flow dynamic on Legionella in potable water systems. This is especially complex due to the inherent intermittent and variable usage observed within outlets of a potable water system. This systematic literature review examines the role of fluid dynamics and stagnation on the colonization and growth of Legionella in potable water systems. Twenty two of 24 identified studies show a positive association between stagnation zones and increased colonization of Legionella. These zones included dead legs, dead ends, storage tanks, and obstructed water flow (such as intermittent usage or flow restriction). Prolonged stagnation in building plumbing systems also deteriorates the quality of thermally or chemically treated potable water. This stimulates the colonization of Legionella established biofilms. Such biofilms are intrinsically resistant to disinfection procedures and accelerate the rate of decay of chemical disinfectants. Sub-lethal doses of disinfectants and the presence of protozoan hosts in stationary water promote generation of viable but non-culturable Legionella cells. This results in false negatives in surveillance methods that use culture methodology. In conclusion, elimination of temporal and permanent stagnation points can improve the quality of potable water, efficacy of disinfectants, and reduce the risk of legionellosis. Current guidelines and water safety plans recognize the risks associated with permanent stagnation point (dead ends and dead legs); however, there is a need for greater emphasis on controlling temporal stagnation arising from intermittent usage.


INTRODUCTION
The importance of Legionella as an opportunistic waterborne intracellular human pathogen is well-documented (Berjeaud et al., 2016). It is frequently associated with nosocomial and community-acquired legionellosis (Pontiac fever and Legionnaires' disease [LD]) in immunocompromised and immunosuppressed individuals (Fields et al., 2002). Legionella is ubiquitously present in constructed water systems. It is frequently detected in domestic and hospital water systems, cooling towers, humidifiers, fountains, and pools, (Bartram et al., 2007;Fitzhenry et al., 2017).
Despite recent advancements in disease surveillance and management systems, the legionellosis incidence and outbreak rate remains high. In the USA, there were 290 reported outbreaks of legionellosis from 2009 to 2017 (Centers for Disease Control Prevention, 2020) and in 2017, 28 outbreaks were reported in nine European Union countries. In Europe, the legionellosis notification rate also increased from 12 cases/million in 2013 to 18 cases/million in 2017 (European Centre for Disease Prevention and Control, 2019).
According to the USA Centers for Disease Control and Prevention, from 2013 to 2015 ∼67% potable waterborne outbreaks were caused by Legionella pneumophila (Shah et al., 2018). Plumbing networks and potable water distribution systems of hotels, healthcare facilities, and residential buildings were identified as Legionella hotspots (Benedict et al., 2017). Clearly, understanding the environmental factors that affect the survival and growth of Legionella in potable water is essential to inform improved management and control strategies.
The growth of Legionella in potable water systems is primarily associated with biofilms and the free-living amoeba hosts feeding on biofilms (Ashbolt, 2015). This provides Legionella protection from severe physiochemical stresses (Thomas et al., 2004;Dupuy et al., 2011). The environmental conditions that influence the formation of biofilms are essential for the control of Legionella. These include stagnation and the presence of plumbing dead legs and dead ends that provide favorable conditions for microbial growth (Bartram et al., 2007). Dead legs (isolated pipes with limited or no water flow) and dead ends (redundant capped pipe which completely obstruct water flow) are well-established as factors contributing to stagnation (National Academies of Sciences Engineering Medicine, 2019). Stagnation or areas with inappropriate hydraulic dynamics are likely to result in the failure of disinfection procedures (Ling et al., 2018). Due to the complex relationship between different environmental variables it can be challenging to study role of stagnation in Legionella growth. This systematic literature review examines the specific role that water stagnation plays in Legionella colonization and growth within potable water systems. Studies on potable water systems including surveillance and laboratorybased studies that examine the relationship between stagnation and fluid dynamics on water quality and Legionella colonization are explored.

MATERIALS AND METHODS
The protocol of this systematic literature review was designed according to the instructions of the PRISMA statement (Figure 1). The databases Scopus and Web of Science were searched for all articles written in English and published prior to April 2020. In the databases the search terms ("Legionella pneumophila" OR "L. pneumophila" OR Legionella OR legionellosis OR "Legionnaires' disease" OR "Pontiac fever") AND (flow * OR "dead end * " OR "dead leg * " OR "water circulation" OR "water recirculation" OR "Reynolds number" OR stagnation OR stagnant OR "stationary water" OR turbid OR turbidity OR usage * ) AND (plumbing OR pipework OR pipe OR artificial OR piping OR building OR manufactured OR engineered OR potable OR manmade * OR biofilm OR water) were applied. Initially, all identified records were combined and duplicates removed. Subsequently, articles were manually screened by reading the titles and abstracts and excluding those that discussed engineering works, bacterial colonization or biofilms without specifically referring to Legionella. Papers were also excluded that referred to stagnation in municipal water supplies as this is a different issue compared with building water supplies. Finally, the remaining articles were analyzed in full and excluded if they did not examine or describe the relationship between Legionella colonization, growth, survival, and water stagnation, or recirculation in building water systems.

RESULTS
A total of 395 abstracts were obtained from the Web of Science and Scopus. After applying the described inclusion and exclusion criteria (as presented in Figure 1), 24 research articles describing a relationship between water stagnation and Legionella or L. pneumophila survival and colonization were identified as suitable for inclusion in this review (Table 1).

Study Sites
Seventeen of 24 studies (Table 1) investigated hot or cold water storage tanks and building water distribution systems; 7/24 studies were in vitro modeling experiments. The majority of the real-world studies were from investigations of hospital water systems (12/17) and almost half of these were conducted in response to a legionellosis outbreak (5/12). Half of the studies (14/24) discussed the relationship between Legionella contamination or colonization and permanent stagnation points (dead ends 6/24, dead legs 8/24). In the majority of studies, there was limited information provided describing fluid or water dynamics.
Sixteen out of 17 studies examining building plumbing systems showed a positive association between stagnation and increased Legionella colonization/persistence. Ten out of 17 studies of the building plumbing systems demonstrated that permanent stagnation points (dead ends 5/17, dead legs 5/17) were positively associated

L. pneumophila sg1
Post nosocomial LD outbreak prevention measures adopted. Prevention of stagnation in hot water (45.5-47.5 • C) tanks reduced L. pneumophila population.
Colorado, USA Ciesielski et al., 1984 Residential and institutional cold and hot water distribution system L. pneumophila sg1 Stagnation, low hydraulic flow and low chlorine concentration (0-0.7 mg/L) resulted in L. pneumophila contamination.
Brussels, Belgium Ezzeddine et al., 1989 Hospital hot and cold water storage distribution system Post fatal nosocomial LD outbreak prevention measures adopted. Removal of fire hydrant spurs (dead legs) connected to storage tanks reduced the bacterial contamination.
Glasgow, Scotland Patterson et al., 1994 Hospital hot water storage (15,000 L) and distribution system L. pneumophila sg1 10 year surveillance program designed to control nosocomial LD. Circulation of hot water (>55 • C) identified as the best way to reduce risk of nosocomial LD.

Legionella
Removal of dead legs and plumbing system repair unable to produce any profound effect on Legionella control. Disinfection of water with ClO 2 (0.3-0.5 mg/L) best way to control Legionella.
Pennsylvania, USA Sidari et al., 2004 Hospital water distribution system L. pneumophila sg1 Probing of nosocomial legionellosis infections demonstrated that dead end pipe source of Legionella.
Jesenice, Slovenia Tercelj-Zorman et al., 2004 Hotel, office, school, store, and assembly hall hot water distribution system L. pneumophila sg1 and sg4 L. anisa Nonculturable Legionella Hot water system of 40% buildings found contaminated with Legionella. Buildings with central hot water storage system (66.7%) showed higher prevalence of Legionella than buildings with central hot water circulation system (38.5%) or local instantaneous hot water production system (20%).
Osaka, Japan Edagawa et al., 2008 Hospital hot water distribution system  In vitro model plumbing studies Pilot scale domestic water system of loops and dead legs L. pneumophila sg1 Continuous flow of ClO 2 (0.5 mg/L) and chlorine (2.5 mg/L) treated water reduces the bacterial population in loops, whereas continuous flow (7 days) of ClO 2 treated water is only effective way to reduce bacterial population in dead legs. Copper dead legs possess intrinsic antibacterial property and inhibits proper colonization and growth of Legionella.
France Thomas et al., 2004 Pilot scale domestic water distribution model Legionella Regular 20% renewal of dead legs water unable to produce any impact on Legionella population. However, regular complete replacement of dead legs water eliminates culturable Legionella within 7 days.
Frontiers in Environmental Science | www.frontiersin.org with increased Legionella contamination. Three studies presented in Table 1 specifically demonstrated that restricted water circulation resulted in persistence of Legionella in hospital hot water storage tanks with capacities of ≈7,600-15,000 L (Ciesielski et al., 1984;Ezzeddine et al., 1989;Darelid et al., 2002). Ten building water system studies reported a reduction in Legionella numbers after increasing water flow rates and removal of dead legs or dead ends. Some studies specifically used hot water recirculation (Darelid et al., 2002;Gavalda et al., 2019) or an increase in the flow rate of hot water (Ezzeddine et al., 1989;Boppe et al., 2016) to reduce Legionella contamination. In Italy (2018-2020), installation of time flow taps with 64-192 L/day flow rate in the vicinity of dead legs and dead ends effectively reduced L. pneumophila contamination from hot water supplies of hospitals (Totaro et al., , 2020. The one study that did not find any positive association between stagnation and Legionella growth was by Sidari et al. (2004). This study examined water storage and distribution in a 437 bed hospital in Pennsylvania after numerous cases of nosocomial LD over a 5 year period (1994)(1995)(1996)(1997)(1998)(1999). Flushing of hot water temporarily reduced the concentration of Legionella; however, this study found that removal of dead legs and routine chlorination was not sufficient to reduce Legionella contamination. Finally, ClO 2 (0.3-0.5 mg/L) treatment for 21 months resulted in complete removal of Legionella (culture negative).

In vitro Model Plumbing Studies
As with the real-world studies, in 6/7 in vitro model plumbing studies a positive relationship between water stagnation and Legionella colonization was identified ( Table 1). Several model systems demonstrated that stagnation promoted genesis of stable Legionella-eukaryotic microbe biofilms (Vervaeren et al., 2006;Biyela et al., 2012). Farhat et al. (2010) constructed a pilot scale hot water system consisting of both dynamic water loops (flow rate: 15 L/min) and dead leg stagnation areas. They estimated the Legionella concentration (using culture, qPCR, and epifluorescent microscopic counts) was 1-2 log greater in stagnant points. Two studies also investigated the role of ClO 2 treated water in elimination of Legionella contamination from dead legs (Thomas et al., 2004;Loret et al., 2005). It was found that 20% replacement of dead leg water with ClO 2 treated water was not sufficient to completely eliminate culturable Legionella. However, regular complete replacement of ClO 2 treated water in dead legs was able to eliminate culturable Legionella within 7 days (Loret et al., 2005). Another pilot scale domestic water model consisting of both dynamic and stagnant water channels demonstrated that continuous flow of ClO 2 treated water for 7 days significantly reduced L. pneumophila from copper dead legs (Thomas et al., 2004).
The one in situ model study that did not find any positive association with stagnation and Legionella colonization was conducted by Liu et al. (2006). This model consisted of three parallel pipes within a partially open system (5% of water continuously flowed through the system while 95% of the water was recirculated). The water temperature was maintained at 24 • C, one pipe had laminar flow (Re: 355-2,000), one turbulent flow (Re: 10,000-40,000), and one was stagnant. Significantly higher concentrations of culturable Legionella were recovered from the biofilm in the pipe with turbulent flow followed by laminar flow with the lowest Legionella concentrations observed in the biofilm in the pipe with stagnant flow. However, a significant difference in Legionella concentrations in biofilm was observed after 1 week. These concentrations remained static for the remaining 5 weeks of sampling. The authors attributed this result to the turbulent flow increasing the mass transfer of the nutrients and oxygen from the water in the seeding tank. They also noted that the turbulence created may have been insufficient to detach the biofilm. There was no significant difference in planktonic Legionella between any of the treatments. This study concluded that intermittently used pipe work (dead legs) sustained aerobic microbial populations than dead ends (where nutrient depletion is more likely).

DISCUSSION
Legionella is a typical premise plumbing pathogen that can tolerate disinfectants, develop biofilms, survive in waterborne protozoa, and thrive in low levels of nutrients (Fields et al., 2002). Prolonged water stagnation can result in accumulation of nutrients and compromises disinfection, promoting colonization of premise plumbing pathogens in potable water systems (Gauthier et al., 1999;Rhoads et al., 2016;Ling et al., 2018;Xu et al., 2018).

Influence of Water Stagnation on Bacterial Colonization
Microbial ecology of potable water is very complex. From water source to point of use, the varying environmental conditions modulate growth and composition of microbial communities. In building plumbing systems, water can stagnate permanently or temporarily (Prévost et al., 1997). Dead ends permanently stagnate water (Tercelj-Zorman et al., 2004); whereas, water storage tanks and temporal water usage can result in intermittent stagnation (Peter and Routledge, 2018). Prior to consumption, water can stagnate from a few hours to weeks in a building piping system (Manuel et al., 2009). This may result in the deterioration of water quality, and promote accumulation of biomass thereby increasing the concentration of intact and cultivable cells.

Temporary Stagnation
Water in municipal distribution pipelines rarely gets obstructed. As it enters domestic building plumbing systems it stagnates and changes the diversity of microbial communities (Pepper et al., 2004). It has been observed that overnight stagnation resulted in a 4 to 580-fold increase in microbial load (using heterotrophic plate counts), a 2 to 3-fold increase in total cell concentration (using flow cytometry), and a 2 to 8-fold increase in microbial activity (ATP analysis) (Lautenschlager et al., 2010). This was supported by a study which showed that a 2 week stagnation of a domestic buildings plumbing system resulted in a 2 log increase in microbial load (using flow cytometry) (Lipphaus et al., 2014).
Building design also impacts on stagnation and microbial water quality. A recent study demonstrated that water from the plumbing system of a net-zero energy residential building contained 5 log higher amount of bacterial 16S rDNA (6.46 × 10 7 gene unit/mL) and Legionella 16S rDNA (8.91 × 10 4 gene copies/mL) compared with a typical residential building plumbing system (400 gene unit/mL and 100-2.3 × 10 3 gene copies/mL, respectively). This variation, due to building structure and design, could be attributed to increased stagnation prior to consumption as the average stagnation time for a netzero energy residential building is 2.7 days compared with 1 day in a typical residential building (Rhoads et al., 2016). In vitro green building hot water plumbing model studies also suggested limited water flushing and inadequate temperature (51 • C) support colonization of Legionella and host protozoa (Rhoads et al., 2015). In conclusion, stagnation and aging of water in plumbing system of green energy buildings resulted in persistence of microbial contamination.

Permanent Stagnation
Dead legs and dead ends have been implicated in several nosocomial outbreaks of LD (Patterson et al., 1994;Tercelj-Zorman et al., 2004;Bartley et al., 2016). These permanent stagnation points should be avoided or managed during building construction or modifications (Bartram et al., 2007). However, it is impractical to remove them all. Recently, two studies of Italian hospitals demonstrated that installation of time flow taps (flow rate 64-192 L/day) in the vicinity of dead legs was successful in reducing L. pneumophila contamination in the hot water system (Totaro et al., , 2020. In the first study, L. pneumophila from various serogroups including sg3, sg2-4, sg6, and sg10-14, with concentrations ranging from 1 × 10 2 to 1.3 × 10 5 CFU/L, was present in hospital hot water (temperature: 38.2 ± 2.1-46.8 ± 2.1 • C and free chlorine: 0.05 ± 0.007-0.30 ± 0.01 mg/L). Instead of removing dead legs, time flow taps were installed in the outlet closest to each identified dead leg. This strategy successfully eliminated all culturable L. pneumophila . In the second study, it was found that within 15 days of installation of time flow taps (flow rates 192 L/day) in the proximity of dead ends, L. pneumophila sg1, sg2-14, and sg2-16 contamination in hospital hot water (temperature: 44.8-53.2 • C and free chlorine: 0.23-0.36 mg/L) was reduced from 2 × 10 2 to 1.4 × 10 5 CFU/L to no culturability (Totaro et al., 2020). These studies demonstrate that reducing temporary stagnation and increasing flow, even in the presence of permanent stagnation points, reduces the risk of Legionella contamination.
Several model studies have explored the impact of permanent stagnation on Legionella colonization. In a pilot scale model, bacterial biofilms in dead legs (pre-treatment population density 10 7 CFU/L and 10 8 genomic unit/L) survived thermal shock treatment and promoted rapid recolonization within 48 h (Farhat et al., 2010). Thus, permanent stagnation sites act as a reservoir of Legionella biofilms that play an important role in recontamination of water.

Water Stagnation and Failure of Disinfection Procedures
The accelerated decay of residual disinfectant significantly increases the risk of LD (Voss et al., 1985). Prolonged water stagnation, microbial communities, and organic nutrients accelerate the decay of disinfectants (Rhoads et al., 2016;Ling et al., 2018;Xu et al., 2018). According to the US CDC, from 2000 to 2014 70% of LD outbreaks were due to inadequate disinfectant concentrations in water supplies (Garrison et al., 2016). Currently in the USA, six disinfection procedures: Cu-Ag ionization, chlorine, chlorine dioxide, monochloramine, ozonisation, and ultraviolet disinfection, are used to control Legionella (Environmental Protection Agency, 2016b). In the USA, the concentration of chemical disinfectant is maintained in 95% potable water delivered to consumers (Environmental Protection Agency, 2016a). In studies from Austria, Germany, Netherlands, and Switzerland, the concentration of residual disinfectant in potable water (delivered to consumers) is not maintained (Rosario-Ortiz et al., 2016).

Decay of Disinfectant
Different chemical disinfectants are widely used to disinfectant potable water supplies (Pontius, 1990;Kim et al., 2002;World Health Organization, 2004). The residual disinfectant concentration in treated water does not remain constant and it gradually decreases within building plumbing systems. This continuous process of decay may result in complete degradation of chemical disinfectants, thereby increasing the chances of persistence of microbial contamination in building plumbing systems (Vieira et al., 2004). Rates of decay have been associated with temporary or permanent stagnation, aging of plumbing systems, nature of plumbing material, and microbial biomass (Vieira et al., 2004;Patrick et al., 2012;Ling et al., 2018). Goyal and Patel (2015) reported continuous temporal stagnation in storage tanks for 22 h decreased residual chlorine concentration (from 0.2 to ∼0.12 mg/L) in domestic buildings. Similarly, Barbeau et al. (2005) reported that 24 h of temporary stagnation decreased chlorine content of water (from 0.6 to ∼0.3 mg/L in cement line ductile and 0.4-0.05 mg/L in gray cast iron pipe dead end) and increased the microbial count. Galvin (2011) demonstrated that the lower flow velocity observed in dead ends can result in a decrease in concentration of both chlorine and chloramine within 200 h. Laboratory model based experiments have also demonstrated that an increase in residence time of water results in decreased concentrations of chlorine, chloramine, and ozone and increased microbial contamination (Clark et al., 1994). Another model system showed that complete renewal of ClO 2 (0.5 mg/L) disinfected water in dead legs eliminated culturable Legionella (Loret et al., 2005).

Persistence of Intrinsically Disinfectant Resistant Legionella
Intrinsic resistance is a natural tolerance and resistance attributed to Legionella against physical and chemical water treatments (Steinert et al., 1998;Cooper and Hanlon, 2010). Biofilm in areas of stagnation within plumbing systems may harbor resistant populations and act as a continuous source of microbial contamination (Bartley et al., 2016). Studies have shown that currently available water disinfection methods (i.e., chlorine, chlorine dioxide, chloramines, hydrogen peroxide, ozone, copper, and silver ions) are only successful in reducing or eliminating Legionella populations transiently. Sidari et al. (2004) and Totaro et al. (2018) (see Table 1) demonstrated the inability of chlorine to eliminate all culturable Legionella. According to Totaro et al. (2018), circulation of hot chlorinated water in dead legs eliminated all chlorine sensitive culturable L. pneumophila sg2-4 and sg10-14 serotypes, though chlorine resistant and thermostable L. pneumophila sg3 and sg6 serotypes persisted in low concentrations.
A study conducted in a century old hospital in Italy found that continuous hyperchlorination (0.5-1 mg/L) for 5 years was not sufficient to completely eliminate Legionella contamination. Multiple factors including outdated piping, dead legs, improper water circulation and corrosion of plumbing materials were proposed as causes of Legionella persistence (Orsi et al., 2014). Similarly, a study conducted in another Italian hospital plumbing system demonstrated persistence of the same strain of L. pneumophila for a period of 15 years (1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004) despite thermal treatment, chlorination, and chlorine dioxide treatment (Scaturro et al., 2007). This demonstrates the role permanent stagnation (dead ends and dead legs) has in harboring and potentially selecting for more resistant strains through constant exposure to sub-lethal concentration of disinfectants (Cooper and Hanlon, 2010;Dupuy et al., 2011).

Stagnation and Microbial Biofilms
In water storage and distribution systems, 95% of the microbial population exists as biofilms attached to the inner surfaces and only 5% in the water (Flemming et al., 2002). According to an in vitro simulation experiment, plumbing coated with Legionella biofilms decayed free chlorine in stagnant water (48 h stagnation period) and increased the risk of legionellosis up to six times (Huang et al., 2020). Available literature shows that in potable water and building plumbing systems complex biofilms and amoebae hosts protect Legionella (Kilvington and Price, 1990;Thomas et al., 2004). Cargill et al. (1992) reported that in contrast to free living cells, L. pneumophila existing in complex biofilms can tolerate high doses of disinfectant. Specifically, the amoeba Acanthamoeba and Vermamoeba provide additional protection from prolonged and persistent water treatment processes (Kilvington and Price, 1990;Dupuy et al., 2011;Cervero-Arago et al., 2014;Dobrowsky et al., 2016). Donlan et al. (2005) observed that for L. pneumophila-amoebae complex biofilms, monochloramine (0.5 mg/L) was a more effective disinfectant compared to free chlorine (0.5 mg/L). Stagnant potable water allows formation of thick, dense, complex, and adherent biofilms, which accelerate decay of disinfectants (Tsagkari and Sloan, 2018). More importantly, chemical disinfectants are unable to penetrate such multispecies biofilms (Bridier et al., 2011).

Stagnation in Building Hot Water System
The studies presented in Table 1 demonstrate that due to stagnation points and low water consumption, building hot water systems will develop Legionella contamination. Unlike cold water, hot water systems contained diverse species and serotypes of Legionella (Voss et al., 1985;Trop Skaza et al., 2012;Totaro et al., 2018;Girolamini et al., 2019). An in vitro study demonstrated colonization of diverse species of Legionella i.e., L. pneumophila, L. anisa, L. taurinensis, and L. drancourtii, and protozoan hosts i.e., alveolata members, Bodonidae, Euglenozoa, Neobodo curvifilu, Thecamoebae, Vannella, and Vermamoeba vermiformis in a hot water plumbing model. Moreover, it was also noticed that pathogenic L. pneumophila and L. anisa developed stable biofilms with protozoa (Thecamoebae, Vannella, and V. vermiformis) in storage tanks and survived during thermal (70 • C/30 min) and chemical (biodispersant: tensio-active Ferrofos R 5260 and biocide: H 2 O 2 -peracetic acid) treatments. After such treatments, these stable biofilms re-contaminated hot water within the entire plumbing model (Farhat et al., 2012). Saby et al. (2005) analyzed hot water plumbing system models (materials: steel, galvanized steel, and chlorinated polyvinyl chloride) and concluded that chemical disinfection procedures (H 2 O 2 , continuous chlorination, hyperchlorination, and peracetic acid treatment) temporarily eradicated established Legionella biofilm. The only solution to eradicate Legionella biofilms was to maintain water temperature at >55 • C at all points, which required continuous circulation of hot water. In another pilot study (material: stainless steel), it was noticed that thermal treatment (70 • C/30 min) of the plumbing system containing a well-established Legionella (10 3 CFU/cm 2 ) biofilm decreased culturable Legionella. It was also observed that existence of dead legs in plumbing system promoted rapid recontamination of remaining water (Farhat et al., 2010). All studies discussing hot water supplies to buildings showed that stagnation of hot water in storage tanks and dead legs or dead ends led to a reduction in water temperature (<45 • C), and rapid decay of disinfectants promoting colonization of Legionella. This was exacerbated in warm water systems as dissipation of the residual chemical disinfectant was further accelerated at high temperature (Vasconcelos et al., 1997;Ndiongue et al., 2005).

Nutrient and Oxygen Supply Hypothesis
One argument presented by Liu et al. (2006) that counters the positive association of stagnation with Legionella contamination of biofilms is the reduced oxygen and nutrient content present in areas of stagnation compared with areas of turbulent flow. However, L. pneumophila sg1 has been shown to survive in drinking water under low nutrient availability for more than 2 years (Paszko-Kolva et al., 1992). L. pneumophila can also proliferate efficiently at 6-6.7 mg/L concentration of dissolved oxygen. However, at oxygen concentration <2.2 mg/L it can survive but stops multiplying (Wadowsky et al., 1985). L. pneumophila (sg1, sg2, sg3, sg4, and sg6) has also been isolated from different water bodies (i.e., lakes and rivers) with dissolved oxygen concentrations of 0.3-9.6 mg/L (Fliermans et al., 1981). In water at temperatures of 20-45 • C, the concentration of dissolved oxygen ranges from 9.06 to 5.94 mg/L (Environmental Protection Agency, 1989), which is sufficient for Legionella survival and growth. Furthermore, upon environmental stresses (nutrient starvation, low oxygen, osmolarity alterations, pH, and temperature fluctuations) bacteria in biofilms activate stress tolerance mechanisms, which can lead to genesis of viable but non-culturable (VBNC) cells (Yamamoto et al., 1996;Fux et al., 2005). Previous research has also shown replication within protozoan hosts (Buse et al., 2013) and presence of disinfectants (Allegra et al., 2008;Turetgen, 2008;Mustapha et al., 2015;Whiley et al., 2017;Casini et al., 2018) promotes Legionella transformation into VBNC cells. As such, studies that rely solely on culture methods of detection [such as the study by Liu et al. (2006)] will not detect VBNC Legionella and underestimate the numbers present (Whiley, 2016).

Current Methods of Legionella Screening
In stagnation areas, limited nutrient availability and sublethal doses of disinfectant promote generation of VBNC Legionella (Li et al., 2014). According to Farhat et al. (2010) (Farhat et al., 2011). As mentioned earlier, at stagnation points Legionella also exists in complex biofilms, and amoebae hosts dwell in such microbial communities as well. It is well-known that amoebae hosts (V. vermiformis) are capable to transform culturable Legionella into VBNC (Buse et al., 2013). VBNC are pathogenic in nature and infect amoebae and human cell lines (Cervero-Arago et al., 2019). Proteomic profiling suggests VBNC are able to synthesize some virulence factors and proteins involved in different metabolic pathways (Alleron et al., 2013). Acanthamoeba castellanii and A. polyphaga resuscitate VBNC synthesized by starvation (Steinert et al., 1997) and disinfectant treatment (Garcia et al., 2007), respectively.
In vitro studies showed A. polyphaga resuscitated Legionella are infectious for alveolar epithelial and macrophage like cells . So far the underlying mechanisms of VBNC biogenesis and resuscitation are not yet well-understood. Multiple methods are available to screen and detect Legionella contamination in potable water and building plumbing systems ( Table 2). Any method that can identify and estimate both culturable and VBNC Legionella is most appropriate to monitor building plumbing system and potable water. Microbiological culturing (International Organization for Standardization, 2017) and qPCR (International Organization for Standardization, 2019) are approaches recommended by ISO (international organization for standardization), however both methods are unable to provide information about VBNC Legionella. Fluorescence in situ hybridization (Delgado-Viscogliosi et al., 2005) and viability qPCR (Ditommaso et al., 2014) are two techniques that can be used to estimate populations of VBNC Legionella, however validity of both assays is challenged by high density of sample background bacteria other than Legionella. Differential live/dead stain flow cytometry is widely used for in vitro disinfectant efficacy and plumbing model experiments to estimate the population of VBNC Legionella (Allegra et al., 2008(Allegra et al., , 2011Wang et al., 2010;Mustapha et al., 2015). Some researchers have tried to develop and use dye labeled Legionella specific antibodies for detection of VBNC, however these antibodies are highly specific and can only detect specific strain/serogroup/serotypes (Füchslin et al., 2010;Keserue et al., 2012). A universal probe, which can cover all members of genus Legionella, is required to be effective to detect and estimate VBNC contamination from building plumbing systems. Legionellaamoebae co-culturing is one of the best techniques to identify VBNC contamination. In this assay, suspected samples are cultivated on amoebae hosts (preferably Acanthamoeba) and VBNC resuscitation is monitored microscopically (Conza et al., 2013). Garcia et al. (2007) artificially generated VBNC and then resuscitated them back into culturable Legionella using an Acanthamoeba host. Using an Acanthamoeba co-culture assay, Conza et al. (2013) estimated the quantity of VBNC Legionella (10 2 -10 6 cells/g) from a composting facility. A major drawback of Legionella-amoebae co-culturing is the requirement of incubation for prolonged time. However, to our knowledge available literature has not discussed the application of coculture assays to estimate VBNC contamination from building plumbing system.

CONCLUSION
Restricted water circulation and temporary or permanent water stagnation allows Legionella to colonize in building plumbing systems and water storage facilities. Aging of water, stagnation, and microbial biofilms accelerate decay of residual disinfectants and deteriorates water quality in buildings. Further research is required to better understand role of complex Legionellaprotozoa biofilms in degradation of disinfectant in stationary water. To achieve long term disinfection of potable water continuous circulation of thermally or chemically treated water in buildings is the only solution to prevent outbreaks of legionellosis.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

AUTHOR CONTRIBUTIONS
MN and HW drafted and edited the manuscript, HW, KR, MB, and RB corrected and contributed to the manuscript. All authors approved of the final manuscript.