Laying the groundwork for a Legionella pneumophila risk management program for public drinking water systems

According to the National Academies of Science, Engineering, and Medicine (NASEM 2019), water from drinking water systems has been linked to legionellosis outbreaks, and Legionella pneumophila is the species of primary concern. However, while L. pneumophila can be detected in water samples from distribution systems, the levels are typically low and below the levels of public health concern (Bartrand et al. 2024a). L. pneumophila is different from traditional drinking water contaminants because it presents a latent public health risk for public and private drinking water systems and for the building water systems they supply. That is, whereas most contaminants of concern in drinking water are present in source water and risks are reduced through centralized treatment (e.g., pesticides, herbicides, arsenic, Cryptosporidium), L. pneumophila and other opportunistic pathogens are common members of the biofilm consortium in distribution systems and building water systems wherein they amplify to levels of concern under favorable conditions (i.e., if barriers and controls are deficient or inconsistent).

The United States Environmental Protection Agency's (US EPA) National Drinking Water Advisory Council's Working Group agreed with this perspective (NDWAC 2023):

An interesting situation with managing public health risks due to L. pneumophila in water is that all risks, infections, illnesses, and outbreaks have been from exposures that occurred after water has left a public water distribution system and has passed through a building water system. L. pneumophila bacteria that are latent in building water system biofilm are the proximal source of organisms that cause illness, even though their progenitors originated in the water supply (i.e., a drinking water distribution system). Whereas well-managed water supplies seldom if ever have detected L. pneumophila concentrations at levels of direct public health concern (Bartrand et al. 2024a), organisms in building water system biofilm have been documented to amplify to a level of significant health risk (e.g., Völker et al. 2016). The level of L. pneumophila in drinking water from a distribution system remains a data gap, particularly for small systems, consecutive systems, or systems with significant distribution system deficiencies.

While there is much that we do not know about the occurrence of L. pneumophila in engineered water systems, there is sufficient information today with which to initiate water system risk management programs (Burlingame & Bartrand 2023), much the same as some building water systems practice Legionella risk management via water management programs or water safety plans. For example, research has observed that well-operated public water systems can maintain a low level of L. pneumophila such that a public health concern through direct exposure to the drinking water supply would be negligible (Bartrand et al. 2024a). The detection of L. pneumophila in a drinking water distribution system does not, by its presence alone, signal a public health risk. Therefore, a reasonable protocol that could be implemented under the US EPA's drinking water regulations has been proposed for the response to positive detections of L. pneumophila (Bartrand et al. 2024b).

This paper's intent is to present current information on the likelihood of occurrence of L. pneumophila in public water systems to lay a foundation for public drinking water systems, as a stakeholder in public health risk management, to move forward in better managing the public health risk that can result from the amplified occurrence of L. pneumophila.

Better management of the public health risk caused by L. pneumophila in water is enabled by a risk assessment that:

• Identifies hazardous conditions that can lead to intrusion of L. pneumophila into distribution systems or amplification of bacteria latent in the systems;

• Identifies the locations in water systems wherein hazardous conditions develop;

• Identifies the controls that prevent or mitigate those conditions from developing or having an impact; and

• Incorporates this information into a risk management plan.

There are many aspects of a public water system (PWS) that present potentially hazardous conditions, although there is a lack of sufficient data to verify that such conditions significantly impact the occurrence or abundance of L. pneumophila in distribution systems. These conditions, largely based on expert judgement and experience with other drinking water concerns, can be inventoried, and there is value in understanding and managing the likelihood of conditions that can increase the potential risk of exposure to L. pneumophila. There is questionable value in trying to parse out varying levels of the severity of consequences should conditions develop that favor the amplification of L. pneumophila. The preventive and mitigative controls that need to be in place must focus on managing the conditions that favor the amplification and the latent risk.

Our current understanding is that the primary sources for Legionella bacteria entering the PWS distribution system are as follows:

• Treatment breakthrough (e.g., within encysted amebae).

• Intrusion during main breaks.

• Residual organisms from insufficiently sanitized new mains or main replacements.

• Intrusion in storage facilities.

• Intrusion from leaks during negative pressure transients.

• Entrainment from service lines with deficient backflow prevention.

As discussed in the following, knowledge of the relative importance of these sources, the loads of Legionella that are introduced into the system via these sources, and the frequency with which the sources introduce L. pneumophila remain knowledge gaps and important research areas.

Important distribution system processes relevant to Legionella occurrence in the distribution system or in connected buildings include the following:

• Management of disinfectant decay and demand.

• Management of corrosion byproducts and sediment accumulation.

• Minimization of biofilm development and growth.

• Minimization of L. pneumophila establishment within microbial consortia in part(s) of the system.

• Minimization of the occurrence or prevention of conditions favorable to L. pneumophila amplification.

• Active water quality management for meeting regulatory requirements and water quality goals.

Data collected to date indicate that L. pneumophila occurs infrequently in distribution system water samples and at concentrations far below levels of public health significance (Bartrand et al. 2024a). Even so, distribution systems inevitably seed buildings with L. pneumophila that can, in turn, become established in building water systems and pose a latent risk of amplifying to levels of public health significance under the right conditions.

Considering that L. pneumophila presents a latent risk, there are two primary aspects of a PWS that can increase that risk: (1) when a condition exists in the water supply that could exacerbate the amplification of L. pneumophila within a building water system or interfere with the controls in place; (2) when an actual or potential spike in the abundance or occurrence of L. pneumophila has occurred in the water supply that increases the seeding of building water systems. These could be identified by actual testing for L. pneumophila, or they could be suggested by operations or treatment information and/or supporting water quality data. In many cases, the operations information with supporting water quality data would be used to identify conditions that could favor an increase in the occurrence or abundance of L. pneumophila because there is currently a lack of routine monitoring data to support that determination.

Providing water with good quality and adequate pressure to building water systems helps building water quality management maintain control. A loss in water quality or supply (e.g., low or no disinfectant residual, high water temperature, high suspended solids concentration, suboptimal corrosion control, negative water pressure) can promote the survival and amplification of L. pneumophila within the building water system or negate its controls. It can also support the growth and survival of other microorganisms and biofilm that in turn support L. pneumophila. These aspects of the water supply can be summarized in the following four ways:

• L. pneumophila occurrence and abundance lie within a PWS-specific baseline level (not at levels that indicate an acute public health concern), but water quality conditions exacerbate the challenges that building water system managers face in preventing amplification of the L. pneumophila that seeds building water systems.

• An actual spike in L. pneumophila, as determined and confirmed by lab testing of water samples, suggests that amplification in the distribution system has occurred and that those conditions could exacerbate the challenges that building water system managers face in preventing amplification of L. pneumophila, or that overcome the controls in place in building water systems.

• A potential, but not verified spike in L. pneumophila, as suggested by water system treatment/operations information (e.g., a temporary loss of primary disinfection or a water main break that did not have a controlled shutdown) results in a suspected increase in L. pneumophila abundance and/or occurrence that increases the likelihood that building systems are seeded with L. pneumophila.

• A potential, but not verified, spike in L. pneumophila as suggested by water system treatment/operations information along with supporting water quality or supply data that verifies that water quality or supply was affected and conditions favoring L. pneumophila amplification or intrusion were known to exist (e.g., an exceedance of water treatment turbidity goals along with an increase in turbidity as measured at the entry point to distribution, or a failure in the turnover of water in a storage tank along with a loss in disinfectant residual and an increase in heterotrophic bacteria in the affected distribution system). This results in a suspected increase in L. pneumophila abundance and/or occurrence along with the exacerbation of challenges that building water system managers face in preventing the amplification of L. pneumophila.

Legionella and free-living ameba are abundant in source waters and in other environmental waters (i.e., runoff, trench water) and in soil. The treatment of such water and the prevention of such water or soil from entering the drinking water system protects against the potential intrusion of L. pneumophila. Water management practices that reduce the buildup of sediment and biofilm within the drinking water system minimize the opportunity for L. pneumophila and free-living ameba to be retained and proliferate. Although all water systems differ, there are conditions (e.g., extreme rainfall events in the watershed, corrosion byproducts from unlined cast iron mains, dead volume in storage facilities) that are common. These aspects of the public drinking water system (increase in L. pneumophila occurrence or abundance and water quality and supply disturbances) drive the latent risk, largely presented with building water systems, and the effectiveness of controls to manage the risk.

The above discussion lays out a general perspective of a PWS and how its operation and maintenance may affect the ability of a building water system to maintain a minimal risk of L. pneumophila amplification. The following is a review of the literature to identify an inventory of possible, more specific conditions that could potentially promote L. pneumophila being present in drinking water at either an elevated abundance or at an increased frequency of occurrence, and/or water quality and supply conditions that would contribute to its amplification. These conditions are categorized as being associated with the source water, water treatment processes, post-treatment conditions prior to distribution, and within the distribution system:

• Source water: Source water that is impacted by untreated wastewater and stormwater, such as during flood conditions in the watershed, can have increased levels of pathogens, which could include L. pneumophila and associated ameba. These can enter a water treatment plant at higher levels than usual, which increases the likelihood of a breakthrough into the drinking water system. Actual confirmation has not yet been provided for L. pneumophila, but the increased risk seems evident (Jagai et al. 2012; Schalk et al. 2012; Van Heijnsbergen et al. 2015; Caicedo et al. 2019).

• Source water: Source water that has a high level of turbidity, during flood conditions in the watershed, can have increased levels of pathogens, which could include L. pneumophila and associated ameba, which can enter a water treatment plant and increase breakthrough into the drinking water system. Actual confirmation has not yet been provided for L. pneumophila; however, evidence has been provided for pathogens in general as associated with turbidity and rainfall (Curriero et al. 2001; Fisman et al. 2005; Tinker et al. 2010; De Roos et al. 2016, 2017).

• Source water: Groundwater, especially when under the influence of surface water, is also subjected to the presence of L. pneumophila, since it is an inhabitant of soil and environmental water (Schalk et al. 2012; Van Heijnsbergen et al. 2015). L. pneumophila and ameba can enter water systems from groundwater (Atkinson et al. 2022). Groundwater systems have been associated with legionellosis outbreaks (Holsinger et al. 2022).

• Water treatment: Water treatment plants, using water sources that contain L. pneumophila, experience ongoing breakthroughs at very low levels (Thomas et al. 2008, 2010; Thomas & Ashbolt 2011; Wang et al. 2012b; Bartrand et al. 2024a; Donohue et al. 2014; Lu et al. 2016; Donohue et al. 2019; LeChevallier 2019a; NASEM 2019; Isaac & Sherchan 2020; LeChevallier 2020; Martin et al. 2020; Malinowski et al. 2022) as evident by the detection of L. pneumophila in the finished drinking water. Such breakthrough is greater during the summer or warmer months (Perrin et al. 2019; Gleason et al. 2023) such as at water temperatures >18 °C or at the highest levels for a water supply (LeChevallier 2019b; LeChevallier 2020) when the survival and growth of L. pneumophila and associated ameba are greater.

• Post treatment: Water systems that carry a free chlorine residual in comparison to a chloramine residual experience greater survival and occurrence of L. pneumophila in the drinking water system. The observation that a chloramine residual is associated with a lower occurrence of L. pneumophila has been reported (Kool et al. 1999; Flannery et al. 2006; Moore et al. 2006; Buse et al. 2012; Donohue et al. 2019; LeChevallier 2019b; Holsinger et al. 2022).

• Distribution system: When a water main or appurtenance is replaced or added to the drinking water system, there is an opportunity for environmental contamination of the materials and the drinking water. Sanitary procedures must therefore be followed. Since L. pneumophila is a common soil and environmental water inhabitant, L. pneumophila and ameba can gain entry depending on the cleanliness of the work. However, while this has been documented for other environmental microorganisms, it has not yet been documented for L. pneumophila (Besner et al. 2011; AWWA 2014; Kirmeyer et al. 2014; Van Heijnsbergen et al. 2015; Omoregie et al. 2022).

• Distribution system: When there is a break in a water main, there is an opportunity for environmental contamination to enter the drinking water and contaminate the main environment (such as with sediment and trench water). The opportunity can be assessed by risk factors and appropriate mitigation procedures can be used to offset the risk. Since L. pneumophila is a common soil and environmental water inhabitant, L. pneumophila and ameba can gain entry depending on how clean and sanitary the work done was. However, while this has been documented for other environmental microorganisms, it has not yet been documented for L. pneumophila (Besner et al. 2011; AWWA 2014; Kirmeyer et al. 2014; Van Heijnsbergen et al. 2015; Rhoads et al. 2017).

• Distribution system: Pressure transients occur in transmission and distribution systems for a variety of normal operating and emergency response reasons. If negative pressure transients occur, they can promote intrusion or pull in environmental contamination, including non-potable water, to varying degrees depending on the extent of the transients. Since L. pneumophila is an inhabitant of soil and environmental water, L. pneumophila and ameba can gain entry; however, while this potential has been documented for other environmental microorganisms, it has not yet been documented for L. pneumophila (Shen et al. 2017; Atkinson et al. 2022).

• Distribution system: Flooded meter pits can allow environmental contaminants to enter the system during pressure transients. Since L. pneumophila is a common soil and environmental water inhabitant, L. pneumophila and ameba can gain entry in this way. However, while this potential has been documented for other environmental microorganisms, it has not yet been documented for L. pneumophila (Besner et al. 2011; Van Heijnsbergen et al. 2015).

• Distribution system: Sections of a distribution system where the disinfectant residual is low, the water is seasonally warm (e.g., > 18 °C), biofilm grows, and/or sediment collects are opportune sites for bacterial regrowth, which can include L. pneumophila and associated ameba wherein L. pneumophila can be protected and amplify. The potential for this to occur has been documented (Moore et al. 2006; Trolio et al. 2008; Buse et al. 2012; Wang et al. 2012a, 2012b; Delafont et al. 2013; Cohn et al. 2015; Lu et al. 2015; Rodriguez-Martinez et al. 2015; Rhoads et al. 2017; Mraz & Weir 2018; Shaheen & Ashbolt 2018; LeChevallier 2019b; NASEM 2019; Falkinham 2020; LeChevallier 2020; Martin et al. 2020; Rhoads et al. 2020; Biyela et al. 2012; Del Olmo et al. 2021; Zhang et al. 2021; Dowdell et al. 2023; Xin et al. 2023).

• Distribution system: Sections of a distribution system where biofilm is not well controlled, and changes in water demand and flow disrupt it, can experience microbial contributions to the drinking water from the biofilm. It is possible that biofilm containing L. pneumophila or associated ameba can be dislodged into the water system, although actual proof has not yet been provided for a real water system (Moore et al. 2006; Biyela et al. 2012; Buse et al. 2012; Wang et al. 2012a, 2012b; Cohn et al. 2015; Rodriguez-Martinez et al. 2015; Shen et al. 2015, 2017; Waak et al. 2018; Falkinham 2020; Del Olmo et al. 2021).

• Distribution system: Sections of the water distribution system that experience increased iron corrosion can also experience an increase in L. pneumophila since iron is a nutrient for its growth and corrosion can provide a protective environment for growth. This has not yet been verified for real water systems, but the potential has been shown (Wang et al. 2012a; Rhoads et al. 2017, 2020; LeChevallier 2020).

• Distribution system: Water in a storage facility that has a low disinfectant residual, high water age, is seasonably warm (e.g., > 18 °C) and is in contact with sediment could be a source of L. pneumophila and associated ameba (EPA 2002; Biyela et al. 2012; Buse et al. 2012; Cohn et al. 2015; Lu et al. 2015, 2016; King et al. 2016; Rhoads et al. 2017; Del Olmo et al. 2021). Studies documenting the presence of L. pneumophila in storage facility sediments used molecular methods (real-time, quantitative polymerase chain reaction (qPCR)), although the concentration range of culturable, viable L. pneumophila in sediments remains unknown.

• Distribution system: Stagnant water in a storage facility, supporting the regrowth of microorganisms that could also contain L. pneumophila and associated ameba, could be drawn out into the water system during high demand, although actual proof that this happens has not yet been provided (EPA 2002; Biyela et al. 2012; Wang et al. 2012b; Cohn et al. 2015; Falkinham 2020).

• Distribution system: Water storage facilities that have openings to the environment (e.g., hatches, vents, holes, cracks) can allow various contaminants to enter and such contamination could include L. pneumophila and associated ameba. This has been shown for pathogens in general but has not yet been documented for L. pneumophila (EPA 2002).

• Distribution system: Backflow or back-siphonage through unprotected cross-connections are highly under-reported and are, therefore, a regular risk for drinking water contamination, which could include L. pneumophila if the non-potable water source (e.g., recirculating hot water system, cooling tower, irrigation pond) contained L. pneumophila. Such events can vary greatly in extent and be due to normal operating or emergency conditions in the PWS. Therefore, a water system's provision of supply and pressure is important for the management of L. pneumophila risk. Actual data on L. pneumophila in backflow water has not been reported or researched, although such non-potable waters have been shown to be sources of L. pneumophila (EPA 2001; Omoregie et al. 2022).

• Distribution system: A water system that is not in control of coliform regrowth and occurrence could be more likely to also have a greater occurrence of L. pneumophila as the conditions for coliform growth can also favor the growth of L. pneumophila, especially when the disinfectant residual is unstable and low, and the water temperature is greater than 18 °C (or at its warmest). Actual studies of such situations do not yet exist (Kuchta et al. 1983).

• Distribution system: A water system that is not in control of its heterotrophic plate count (HPC) bacteria, either system-wide or in sections of the system, shows signs of conditions that could favor L. pneumophila survival and growth along with the growth of ameba (such as stagnant water, a higher level of nutrients, loss of disinfectant residual, warm water). Long water age, stagnant water conditions, and loss of a disinfectant residual are known to promote microbiological growth and biofilm, such as in dead-end water mains and storage facilities as well as in building water systems (WHO 2003; Moore et al. 2006; Falkinham 2020). However, actual documentation of these conditions promoting the growth of L. pneumophila in a distribution system does not yet exist, although the connection seems logical.

The following three tables are based on the above review of the literature in the context of the basic ways that conditions can develop from the water supply and increase the public health risk. The conditions presented in Tables 1–3 are common to many utilities and are presented as a suggested starting point for utilities developing their utility-specific hazards inventory and assessing the sufficiency of controls in place. Table 1 summarizes conditions that could provide a pathway for the entry of L. pneumophila into a drinking water system and increase the seeding of building water systems. Table 2 summarizes specific distribution system conditions that might harbor L. pneumophila and allow it to grow, thereby increasing the seeding of building water systems. Table 3 summarizes specific conditions that might challenge a building water system's operations and maintenance of controls, thus making it harder for building water system managers to prevent the amplification of L. pneumophila. Note that a water temperature of 18 °C was chosen to represent the warmest months in a supply that experiences seasonal water temperature changes and was based on findings by LeChevallier (2019b) in his survey of distribution system occurrence of L. pneumophila. This water temperature can be exchanged with the three warmest water temperature months that a specific water utility experiences, which would represent conditions more likely to promote bacterial survival and growth.

The likelihood that conditions such as those in Tables 1–3 exist and their frequency of occurrence in a specific PWS would be evaluated based on existing technical knowledge of the specific system as provided by the water system's operators. The likelihood could be based on actual events or near misses for experiencing the conditions, on an expert judgement about the system, on research, or on experience from nearby or similar water systems. The likelihood of a public health risk from L. pneumophila presenting itself directly from a PWS still needs research and depends on each PWS's characteristics, operations, and treatments.

Once hazardous conditions are listed or inventoried and the likelihood of occurrence determined or estimated, then preventive and mitigative controls can be put in place or better assessed for their reliability and effectiveness. Preventive controls will deter hazardous conditions from developing, whereas mitigative controls are actions taken to remove or reduce the conditions once they have developed. Controls require monitoring to confirm that they are working, along with critical control limits, that when exceeded, drive actions such as mitigative treatment and communications. Many of the controls and management programs are already in place for other water quality management purposes (e.g., compliance with drinking water regulations).

In addition, the estimated likelihood of occurrence of conditions based on expert judgement alone may need the support of research to determine with greater reliability whether the conditions are likely to occur in a specific system. In addition, even if certain conditions are likely to occur, such as a loss of disinfectant residual in a storage tank, water monitoring would provide important data to determine whether such conditions actually increase the occurrence or abundance of L. pneumophila and thereby increase the latent risk.

The above summary of existing knowledge lays down a foundation for PWSs to develop risk management and a standard of care for L. pneumophila that can be fine-tuned over time as new research and other information comes to light. Examples of hazardous conditions have been presented from which a water system can make a hazardous conditions inventory specific to its system's characteristics and operations.

A deliberate and targeted approach to risk management is needed to promote the ongoing reduction of public health risks due to L. pneumophila with PWSs as important stakeholders. This paper presents summary information for such an approach. This approach is based on many assumptions, or expert opinions because there is a lack of data to identify sources and causes of elevated L. pneumophila in drinking water distribution systems. Bartrand et al. (2024a) have provided data on which to estimate baseline occurrence and abundance, but studies are needed to identify cases associated with the amplification of L. pneumophila or with water quality and supply conditions that exacerbate the challenges that building water system managers face in preventing its amplification in their water systems.

Sector-wide development and implementation of effective L. pneumophila management programs are hampered by significant data and knowledge gaps. Critical research that will advance the water system's assessments and management of L. pneumophila includes the following:

• Continued collection and analysis of L. pneumophila occurrence and concentration data, along with water quality and contextual data (system characteristics and sample location characteristics). Central collection and sharing of these data, such as in a quality-assured database, would greatly help with a better understanding of the baseline variability of L. pneumophila across water systems and water supply sources.

• Collection and analysis of L. pneumophila data during distribution system disturbances and maintenance. Such as the collection of samples when main repairs and replacements are performed and when water main breaks occur.

• Intensive, large-volume sample monitoring at distribution system entry points with sample assays for Legionella spp., L. pneumophila, free-living amebae, and potentially other organisms. Large-volume samples and frequent sample collection are required given the infrequent occurrence and low concentrations anticipated for finished drinking waters.

• Additional collection and analysis of sediment and biofilm samples from various water mains and storage facilities where L. pneumophila may be harbored to better understand its contribution to the microbiome of water systems.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.