Microbial and Viral Indicators of Pathogens and Human Health Risks from Recreational Exposure to Waters Impaired by Fecal Contamination

McKee, Anna M.; Cruz, Marcella A.

doi:10.1061/JSWBAY.0000936

Open access

State-of-the-Art Reviews

Jan 31, 2021

Microbial and Viral Indicators of Pathogens and Human Health Risks from Recreational Exposure to Waters Impaired by Fecal Contamination

Authors: Anna M. McKee https://orcid.org/0000-0003-2790-5320 [email protected] and Marcella A. CruzAuthor Affiliations

Publication: Journal of Sustainable Water in the Built Environment

Volume 7, Issue 2

https://doi.org/10.1061/JSWBAY.0000936

PDF

Abstract

Fecal indicator bacteria (FIB) (e.g., fecal coliforms, Escherichia coli, and enterococci) have been used for decades to monitor for and protect the public from waterborne pathogens from fecal contamination. However, FIB may not perform well at predicting the presence of waterborne pathogens or human health outcomes from recreational exposure to fecal-contaminated surface waters. Numerous factors can influence the relationship between FIB and pathogens or human health outcomes, including the source(s) of contamination, the type of pathogen(s) present, differences in the survival and behavior of FIB and pathogens in the wastewater conveyance and treatment process, and varying environmental conditions. As a result, different indicators, such as source-specific microbial source tracking (MST) markers and viral fecal indicators, have been used as possible surrogates to better approximate pathogen abundance and human health risks in recreational waters. The performance of these alternative indicators has been mixed, with some promise of viral indicators better approximating viral pathogens than bacterial fecal indicators, and FIB generally more closely associated with bacterial and protozoal pathogen presence than human MST markers. Many of the assays to detect and quantify fecal indicators and pathogens are polymerase chain reaction-based assays, which detect and quantify nucleic acid [deoxyribonucleic acid (DNA) and ribonucleic acid (RNA)] sequences specific to a target of interest. Recent advances in DNA and RNA sequencing technologies may push the field toward metabarcoding approaches, where multiple targets can be detected and quantified simultaneously. Metabarcoding is currently more applicable to bacterial and protozoal assessments than viral assessments based on a lack of universal metabarcoding markers for viruses. Innovative technologies, such as biosensors and nanotechnologies, may provide more sensitive and accurate tools to detect and quantify pathogens. When a specific pathogen is of concern for a recreational water body, a practical approach in estimating the likelihood of human health outcomes is the application of quantitative microbial risk assessments (QMRAs). Quantitative microbial risk assessments can be used to model the likelihood of pathogen-specific human health outcomes from recreational exposure as a function of a surrogate indicator. Inputs for QMRAs include the ratio between the indicator to be monitored and the pathogen of interest, the concentration of the indicator, the amount of water ingested, and the likelihood of the health outcome based on the estimated amount of pathogen consumed. There are numerous unknowns about the behavior and survival of fecal indicators and pathogens in environmental waters. Developing accurate models to predict pathogen concentrations from fecal indicators in recreational waters will require a better understanding of these unknowns. Current methods and technologies for detecting and quantifying fecal indicators and pathogens are limited due to the rare and patchy nature of pathogens. Technological advances may help improve sensitivity for detecting and quantifying pathogens.

Introduction

Exposure to water-borne pathogens is a common human health risk in recreational surface waters. Urban stormwater runoff can contain high levels of fecal contamination and water-borne pathogens (Sidhu et al. 2012, 2013), suggesting that it may be an important source of contamination to surface waters. The purpose of this review is to provide an overview of the state of the science in the monitoring of fecal-associated waterborne pathogens and microbial and viral human health risk indicators in surface waters impaired by fecal contamination. The general focus of this review is pathogens transmitted via the fecal-oral route and associated with gastrointestinal (GI) illnesses versus other recreational waterborne illnesses, including those caused by inhalation, dermal, and ocular exposure routes (e.g., respiratory illnesses, rash, and eye ailments). Studies have found GI illnesses generally occur at higher rates than other illnesses caused by water-based recreational exposure (EPA 2012). Further, the USEPA has suggested that water quality criteria aimed at protecting the public from GI illnesses will prevent most other waterborne diseases (EPA 2012).

We review traditional methods for assessing the human health risk of recreating in freshwater systems with fecal pollution and how the science of microbial and viral monitoring for human health risks in recreational waters has evolved. We compare the benefits and limitations of different indicators. We discuss direct monitoring of pathogens, including the status of monitoring efforts, current methodologies, limitations of these methodologies, and statistical approaches to estimating human health risks quantitatively. Finally, we discuss research gaps to be addressed in order to better monitor and predict human health risks from exposure to fecal-associated pathogens in recreational waters.

Methods

Google Scholar search terms included but were not limited to pathogens, viruses, fecal indicator, fecal contamination, microbial source tracking, sewage pollution, freshwater, recreational water, environmental water, quantitative microbial risk assessment, metabarcoding, bacteriophage, coliphage, and crAssphage. Google Scholar Alerts was used to receive notification of new scientific publications or preprints with select keywords in the title as they become available; selected keywords included microbial source tracking, environmental deoxyribonucleic acid (DNA), and metabarcoding. Review papers and articles published in 2015 or later were prioritized initially and were used to identify seminal, foundational, and other relevant publications for more in-depth information. If a publication was particularly pertinent, the Google Scholar feature cited by… was used to look for more recent publications on the topic. Additional pertinent keywords were searched in Google Scholar as they were encountered in papers. Articles were generally restricted to freshwater systems, although some papers on marine systems and wastewater treatment plants (WWTP) were included.

Background

Human fecal contamination is a primary source of infectious enteric pathogens in many recreational waters (McBride et al. 2013; Wade et al. 2006; WHO 2017). These water-borne pathogens are transferred via the fecal-oral route in which infected individuals have high shedding rates, and low doses of pathogens are required for illness, such that accidental ingestion or inhalation can result in transmission. The Clean Water Act (Pub. L. No. 92-500), as amended by the Beaches Environmental Assessment and Coastal Health Act in 2000, requires the EPA to recommend research-based guidelines for monitoring indicators of pathogens. Direct screening of water for pathogenic microorganisms and viruses would be ideal. However, the number of possible pathogens that could be present—including viruses, bacteria, and protozoa (Table 1)—makes this cost-prohibitive and would be impractical and time-consuming.

Table 1. Waterborne pathogens transmitted through drinking water supplies

Pathogen^a	Primary associated disease(s)	Health significance^b	Persistence in water supplies^c	Resistance to chlorine^d	Relative infectivity^e	Important animal source^f
Bacteria
Burkholderia pseudomallei	Melioidosis	High	May multiply	Low	Low	No
Campylobacter jejuni, C. coli	Gastroenteritis	High	Moderate	Low	Moderate	Yes
Escherichia coli—pathogenic^g	Gastroenteritis	High	Moderate	Low	Low	Yes
E. coli—enterohaemorrhagic	Gastroenteritis, hemolyticuremia	High	Moderate	Low	High	Yes
Legionella spp.	Legionnaires’ disease	High	May multiply	Low	Moderate	No
Nontuberculous mycobacteria	Pulmonary disease, skin infection	Low	May multiply	High	Low	No
Pseudomonas aeruginosa	Pulmonary disease, skin infection	Moderate	May multiply	Moderate	Low	ND
Salmonella typhi	Typhoid fever	High	Moderate	Low	Low	No
Salmonella enterica	Salmonellosis	High	May multiply	Low	Low	Yes
Shigella spp.	Shigellosis	High	Short	Low	High	No
Staphylococcus aureus	Gastroenteritis, skin infection	ND	ND	ND	ND	ND
Vibrio cholerae^h	Cholera	High	Short to long	Low	Low	No
Yersinia enterocolitica	Gastroenteritis	Moderate	Long	Low	Low	ND
Viruses
Adenoviruses	Gastroenteritis, respiratory infection	Moderate	Long	Moderate	High	No
Enteroviruses	Gastroenteritis	High	Long	Moderate	High	No
Astroviruses	Gastroenteritis	Moderate	Long	Moderate	High	No
Hepatitis virus A, E	Hepatitis	High	Long	Moderate	High	Potentially
Noroviruses	Gastroenteritis	High	Long	Moderate	High	Potentially
Sapoviruses	Gastroenteritis	High	Long	Moderate	High	Potentially
Rotaviruses	Gastroenteritis	High	Long	Moderate	High	No
Protozoa
Acanthamoeba spp.	Keratitis, encephalitis	High	May multiply	Low	High	No
Cryptosporidium parvum	Cryptosporidiosis	High	Long	High	High	Yes
Cyclospora cayetanensis	Gastroenteritis	High	Long	High	High	No
Entamoeba histolytica	Amoebic dysentery	High	Moderate	High	High	No
Giardia intestinalis	Giardiasis (Beaver fever)	High	Moderate	High	High	Yes
Naegleria fowleri	Primary amoebic meningoencephalitis	High	May multiply	Low	Moderate	No
Toxoplasma gondii	Toxoplasmosis	High	Long	High	High	ND
Helminths
Dracunculus medinensis	Dracunculiasis (Guinea worm disease)	High	Moderate	Moderate	High	No
Schistosoma spp.	Schistosomiasis	High	Short	Moderate	High	ND

Source: Adapted from Kumar et al. (2018) and WHO (2017).

a

This table contains pathogens for which there is some evidence of health significance related to their occurrence in drinking-water supplies and is not an exhaustive list of waterborne pathogens.

b

Health significance relates to the incidence and severity of the disease, including association with outbreaks.

c

Detection period for infective stage in water at 20°C: short, up to 1 week; moderate, 1 week to 1 month; and long, over 1 month. Within pathogen species and groups, there are likely to be variations in resistance, which could be further impacted by characteristics of the water supply and operating conditions.

d

Resistance is based on 99% inactivation at 20°C where, generally, low represents a

C t_{99}

of

< 1 \min \cdot mg / L

, moderate

1 - 30 \min \cdot mg / L

, and high

> 30 \min \cdot mg / L

(where

C

= concentration of free chlorine in

mg / L

; and

t

= contact time in minutes) under the following conditions: the infective stage is freely suspended in water treated at conventional doses and contact times, and the pH is between 7 and 8. It should be noted that organisms that survive and grow in biofilms, such as Legionella and mycobacteria, will be protected from chlorination.

e

From experiments with human volunteers, from epidemiological evidence, and from experimental animal studies. High means infective doses can be

1 - 10^{2}

organisms or particles, moderate

10^{2} - 10^{4}

, and low

> 10^{4}

.

f

Wild, domestic, or farm animal play that plays a role in the chain of pathogen transmission typically through shedding of the pathogenic organism via feces.

g

Includes enteropathogenic, enterotoxigenic, enteroinvasive, diffusely adherent, and enteroaggregative.

h

Vibrio cholerae may persist for long periods in association with copepods and other aquatic organisms. ND indicates the information was not defined by WHO (2017).

Instead, indicators of fecal contamination are used as surrogates to estimate the health risk from exposure to fecal-associated pathogens. Characteristics of ideal surrogates for fecal contamination include the following (EPA 2015):

•

Specificity to a host species or indicative of a specific source of origin;

•

High abundance in feces of host species;

•

Cooccurrence with pathogens (present when pathogens are present, and absent when pathogens are absent);

•

Occurrence at a higher concentration than pathogens and relatively easy and inexpensive to monitor and detect;

•

Similar survival, fate, and transport as pathogens in WWTP and the environment;

•

Does not multiply outside of the host;

•

Nonpathogenic (to avoid the risk of illness for those who are monitoring); and

•

Correlated with human illness.

Some surrogates are more likely to meet particular criteria than others (Payment et al. 2010); however, there is currently no known single surrogate that meets all criteria.

Fecal Indicator Bacteria

Fecal coliforms were the first fecal indicator bacteria (FIB) recommended by the EPA (1976) to be used for indicating the possible presence of pathogens in designated recreational waters. Although fecal coliforms are abundant in the intestinal tract of humans and can be indicative of fecal contamination, they can also occur across a broad range of hosts and are able to replicate in environmental settings (McLellan et al. 2001), resulting in inconsistent relationships between FIB concentrations and fecal-associated pathogens. In the late 1970s and early 1980s, the EPA conducted epidemiological studies that compared several fecal indicators, including fecal coliforms, Escherichia coli (E. coli), and enterococci (Cabelli et al. 1983; Dufour 1984). The results from these studies led the EPA to change their recommended fecal indicator for freshwater from fecal coliforms to E. coli and enterococci in their 1986 Recreational Water Quality Criteria (RWQC) (EPA 1986). Both E. coli and enterococci are also abundant in the intestinal tract of humans and a variety of other host species but, in general, do not replicate outside the gut system (Ishii and Sadowsky 2008; Rivera et al. 1988; Staley et al. 2014).

Although not used in the US for regulatory purposes, another bacterium that is occasionally used for monitoring fecal contamination and water quality is Clostridium perfringens (C. perfringens) (Wu et al. 2011). C. perfringens is a bacterium that can be pathogenic and occurs in feces in both an environmentally sensitive state and environmentally resistant spore form. Like E. coli and enterococci, C. perfringens occurs in all warm-blooded animals and, therefore, does not provide information about the source of fecal contamination. Because C. perfringens is not as widely used as fecal coliforms, E. coli, and enterococci for monitoring water quality, it is considered an alternative fecal microorganism as opposed to a general FIB (Korajkic et al. 2018). C. perfringens is also used as an alternative fecal microorganism in tropical and subtropical waters, where general FIB may replicate (Ferguson and Signoretto 2011).

The studies used to develop the RWQC investigated human health outcomes in swimmers recreating downstream of a WWTP, which was assumed to be the primary source of fecal contamination and FIB (EPA 2012). Although the EPA studies used for the RWQC found E. coli and enterococci to be a better predictor of illness rates than fecal coliforms, other studies have failed to detect a correlation between these FIB and a pathogen presence or human health outcomes (for a review, see Korajkic et al. 2018). One factor for these discrepancies could be the differences in the health risk from exposure to contamination from different sources (Schoen and Ashbolt 2010; Soller et al. 2010b, 2015). The 2012 EPA RWQC document acknowledges possible differences in infectious disease potential depending on the sources of fecal contamination. However, the difficulty of conducting experiments to determine health risks associated with exposure to specific sources has precluded the EPA from developing source-specific RWQC (EPA 2012). In lieu, they have included recommendations for tools that states can use to develop alternative site-specific water quality criteria for locations impacted by fecal contamination from nonhuman sources.

Microbial Source Tracking

In addition to potential differences in human health risks from exposure to fecal contamination from different sources, source information is also critical for developing mitigation plans for areas with high levels of fecal contamination. Microbial source tracking became prominent at the end of the 20th century with the objective of determining dominant sources of fecal contamination based on the detection of microbiota associated with particular host species (Field and Samadpour 2007; Scott et al. 2005; Stoeckel and Harwood 2007). Many early MST studies relied on comparisons of microbial fingerprints between environmental samples and local reference libraries (Harwood et al. 2000; Parveen et al. 1999; Wiggins 1996). Methods that relied on the comparison of environmental samples to reference samples to infer sources of fecal contamination are known as library-dependent methods. Depending on their size, reference libraries used for MST studies are often unique to the particular study area (Hartel et al. 2002) and can vary over time (Wiggins et al. 2003), which limits their broadscale utility and necessitates a large initial investment for development (Stewart et al. 2003). Technological advances have facilitated the development of library-independent methods (which do not require sample comparison with a reference library) for MST, including the development of quantitative polymerase chain reaction (qPCR) MST assays.

Many of the library-independent MST qPCR markers target the bacterial order Bacteroidales, and particularly the genus Bacteroides, which are obligate anaerobes and cannot survive in the environment (but see Yamahara et al. 2020). Because they are unable to survive under aerobic conditions in the environment, Bacteroides are useful as indicators of recent contamination. They are restricted to warm-blooded animals and, unlike fecal coliforms or E. coli, make up a substantial portion of the bacteria present in mammalian feces. Additionally, Bacteroides have coevolved with their host species. Together, these characteristics make them good candidates for MST markers because of their sensitivity and specificity. Marker sensitivity reflects the ability to be confident that a lack of marker detection is indicative of the absence of target host contamination; an MST marker with perfect sensitivity would be detectable when host contamination is present and, therefore, if not detected, it would indicate the absence. Marker specificity reflects the ability to be confident that its detection is indicative of contamination from the target host; an MST marker with perfect specificity would only be detectable when host contamination is present and, therefore, detection would indicate host contamination is present. In practice, MST assays generally do not exhibit perfect sensitivity and specificity (Boehm et al. 2013), and compounds in environmental samples can further reduce the sensitivity of assays by reducing the efficiency of PCR (referred to as PCR inhibitors). The focus of MST qPCR marker development has been primarily focused on human, ruminant, porcine, and poultry sources (García-Aljaro et al. 2019). However, markers for other host species have been developed, including a general bird marker (e.g., GFD; Green et al. 2012), dogs (e.g., DG3 and DG37; Green et al. 2014b), horses (Dick et al. 2005), and others. Microbial source tracking markers exhibit a range of sensitivities and specificities (Boehm et al. 2013; Harwood et al. 2014). For humans, the most widely tested MST marker is arguably HF183, which is used in the EPA-validated method 1696 for detecting human contamination (EPA 2019) and has several primer-set variations (e.g., Bernhard and Field 2000; Green et al. 2014a; Haugland et al. 2010). The EPA-validated method for HF183 is based on the marker set described by Green et al. (2014a) and was developed in response to stakeholder need for a validated human MST assay for purposes, such as the national pollutant discharge elimination system permitting (EPA 2019). HF183 is generally a highly sensitive marker for sewage; however, it has been detected in nonhuman sources (Boehm et al. 2013), such as deer (Nguyen et al. 2018), albeit at lower levels than from human sources.

In a recent metaanalysis of relationships between microbial indicators and pathogens in recreational waters, Korajkic et al. (2018) found tenuous relationships between human MST markers and pathogens and human MST markers and illness. Another metaanalysis by Harwood et al. (2014) similarly found a lack of association between human Bacteroidales MST markers in environmental waters and various types of waterborne illnesses, including GI and respiratory illnesses, and skin, eye, and ear infections. However, in a study performed in the Hudson River tributaries in New York during peak recreational water use, Brooks et al. (2020) found correlations between concentrations of a human-specific MST marker (HumM2), enteropathogenic E. coli (E. coli gene), and a general E. coli gene (Applied Biosystems part number 4460366). Brooks et al. (2020) also found correlations between a bovine-specific MST marker (CowM3), shiga-toxin producing E. coli (E. coli stx1 gene), and E. coli O157 (E. coli rfbE gene), suggesting relationships between human MST markers and pathogens as well as nonhuman sources and pathogens. Several studies suggest that the geographical location and environmental conditions play a role in the success of assays and suggest assessing performance in the laboratory prior to application in environmental settings (Harwood et al. 2014; Boehm et al. 2013; Ahmed et al. 2020). Tenuous relationships between human and nonhuman MST markers and pathogenic bacteria and viruses warrant further investigation to understand better the relationships between source and pathogen markers and their application to water quality management.

Metabarcoding and Community-Based Microbial Source Tracking

Most of the MST research over the last decade has used qPCR methodology to detect specific host-associated bacteria. However, advances in high throughput DNA and ribonucleic acid (RNA) sequencing (HTS), also commonly referred to as next-generation sequencing, have made the technology more accessible for a broad range of applications, including MST (Ahmed et al. 2015; Ahmed et al. 2017). Source tracking via HTS and bacterial community-based analysis can help address some of the issues with sensitivity and specificity that can plague qPCR MST markers (Ebentier et al. 2013; Shanks et al. 2010). Community-based MST is a library-dependent method in which the library is created from operational taxonomic units (OTU; distinct DNA sequences for a given gene) present in feces from different sources. Fecal-associated bacterial OTUs can be categorized as cosmopolitan (found across all tested host species), shared (occur in some host species), or specific (only found in a single host species). The composition of OTUs in feces from a species thus creates a source-specific bacterial profile. The resulting compilation of these microbial profiles from potential sources is referred to as a fecal taxon library (FTL). Microbial community profiles from environmental samples are then compared to the FTL to find matching OTUs. The proportion and probability that particular sources contribute to the total bacterial community can be estimated with Bayesian statistics based on community profile similarities, such as with the software Source Tracker (Knights et al. 2011). Limitations to metabarcoding-based MST include primer bias (i.e., some DNA sequences are detected better than others) and the optimization of sequence processing parameters (e.g., sequence quality filtering and clustering algorithms). Some of the same issues that affect qPCR-based MST also affect metabarcoding-based MST, including differential decay rates of fecal material from different species and the spatiotemporal variability in fecal bacteria community composition (Unno et al. 2018).

Management Applications of Microbial Source Tracking

Microbial source tracking can be particularly helpful for determining sources of fecal contamination when FIB levels are high, but the likely source(s) is/are unknown, such as when the site is unaffected, or unlikely to be affected, by point sources of contamination. For example, McKee et al. (2020) determined that E. coli levels were most strongly correlated with levels of a dog MST marker in the reach of the Chattahoochee River encompassed in the Chattahoochee River National Recreation Area in Georgia. McKee et al. (2020) also found that samples with E. coli levels that exceeded the EPA beach action value of 235 most probable number (MPN) per 100 mL were best predicted by samples with dog MST marker concentrations of 300 gene copies per 100 mL or greater. Results from this study led to a joint effort between the National Park Service and a local nonprofit to install dog waste stations throughout the park. Similar results were found at a California beach, where dogs were found to be more important contributors to fecal contamination than humans or birds (Ervin et al. 2014). Levels of dog contamination decreased after a targeted outreach effort, demonstrating the value of knowing the source and spatial origins of fecal contamination. Also, in California, correlations between FIB and a gull MST marker at a beach led to the implementation of gull-abatement best management practices (Goodwin et al. 2016), which resulted in significant decreases in FIB. In Florida, Nguyen et al. (2018) used MST to demonstrate that wildlife, as opposed to humans, were the sources of high levels of FIB in a stream in a wildlife refuge resulting in the delisting of that stream reach from Florida’s 303(d) list of impaired waters.

Viral Source Tracking

Most MST work has focused on bacterial source tracking markers. However, enteric viruses have shown promise as specific source tracking markers because of the host specificity of many viruses. Human adenovirus and human polyomavirus were found to be promising human source tracking markers (Wong et al. 2012). Other studies have found these markers to have high specificities and sensitivities (Ahmed et al. 2010; Rachmadi et al. 2016). Malla et al. (2019) found human adenovirus and two types of human polyomavirus to have high specificities; however, they found sensitivities of a human adenovirus to be 40% (4 of 10 samples) and sensitivities of the two human polyomaviruses to be 0%–30%. Potential explanations for the discrepancies among studies include geographic differences in gut bacterial communities or methodological differences. Improvements in viral concentration methodologies may help increase the development and usage of viruses as source tracking markers.

Viral Indicators

Exposure to human contamination is theorized to be a greater health risk than exposure to nonhuman contamination in-part due to the specificity of many viruses (e.g., Medina and García-Sastre 2011). Studies have indicated enteric viruses are responsible for the majority of waterborne disease outbreaks from recreational exposure (Boehm et al. 2015; Fewtrell and Kay 2015; Graciaa et al. 2018; McBride et al. 2013; Soller et al. 2010a). In addition to the diseases caused by viruses, they are of a particular public health concern due to the high rate of viral shedding in infected individuals and the low dose necessary to result in illness. However, over 150 human pathogenic viruses have been identified in aquatic environments (Rodriguez-Lazaro et al. 2012; Tran et al. 2015), and monitoring for all viral pathogens is currently not feasible.

Differences between bacteria and viruses include factors such as survival, fate, and transport in WWTP processes and the environment. Because most human health illnesses from exposure to fecal-contaminated recreational waters are caused by viruses, the use of bacteria as surrogates has been called into question (Noble and Fuhrman 2001; Wyer et al. 1995). Although the use of bacterial indicators of fecal contamination is standard practice and used for regulatory purposes in the US, viral fecal indicators have been used for research for the last several decades (Havelaar et al. 1986; Wyer et al. 1995). The EPA developed methodologies for detecting viral indicators in the early 2000s (EPA 2001a, b). Additionally, because of their general host specificity, viruses have also shown promise as source tracking markers (Ley et al. 2002; Noble and Fuhrman 2001). Common viral fecal indicators are described subsequently.

Bacteriophages

In recognition of the different responses between bacteria and viruses to intrinsic and extrinsic factors, the EPA is in the process of developing recreational water quality guidelines based on a viral fecal indicator (EPA 2015, 2017). Viruses that infect bacteria (bacteriophages or informally referred to as phages) can occur in human feces at concentrations similar to FIB (Ballesté et al. 2019). Bacteriophages that infect intestinal bacteria have gained support as fecal indicators because of their persistence throughout the wastewater conveyance and treatment process. The survival, fate, and transport of bacteriophages in the environment may serve as more appropriate indicators of enteric viruses than FIB due to factors such as rates of attachment to particulate matter and persistence in the environment (McMinn et al. 2017).

Coliphages

The subset of phages that infect E. coli are termed coliphages and are the most well-studied phages for consideration as an alternate indicator of fecal contamination by the EPA for regulatory purposes (EPA 2015, 2017). Coliphages have already been included or are proposed in some water quality guidelines in other countries, including Australia, Colombia, and the European Union (EPA 2017), and may be better suited as an indicator of viral pathogens than FIB when assessing the presence or removal of viruses in constructed or natural environments (McMinn et al. 2017). Coliphages are organized into two groups based on the route by which they infect E. coli, which can affect the method used for their detection and quantification. Somatic coliphages infect E. coli through their cell walls, whereas F-specific coliphages (also known as male-specific or F+ coliphages) infect E. coli through the sexual pili. Together, somatic and F+ coliphages constitute the total coliphage group. Similar to FIB, both groups of coliphages occur in human and animal hosts. The presence of bacterial fecal indicators is not necessarily indicative of the presence of bacterial pathogens. Similarly, the presence of coliphages is not necessarily indicative of a human viral pathogen presence. However, compared to bacterial indicators, the characteristics of coliphages may more closely resemble those of viral pathogens, and therefore, coliphages could be a more suitable indicator of viral pathogen survival in recreational waters if viral pathogens are present. Standardized and cost-effective quantitative and presence/absence methods are available for each group (EPA 2001a, b), with recently developed assays that can provide results in 1.5–4 h (Rames and Macdonald 2019; Toribio-Avedillo et al. 2019), although the maximum sample volumes for some of these assays (e.g., 10-mL for the QuantiPhage assay) may make them less applicable to recreational surface waters for which fecal pollution is diluted by ambient waters.

Expert opinions vary on which group of coliphages, or both, would be better suited for a variety of applications (EPA 2016). The two groups of coliphages each have different advantages over the other for use in recreational water monitoring. Both coliphage groups have demonstrated relationships with GI illness for recreational waters (EPA 2015). However, a more recent review by Korajkic et al. (2018) suggested the percent of studies demonstrating relationships between F+ coliphages and illness was greater than those demonstrating relationships between somatic coliphages and illness (Fig. 1). Somatic coliphages are more numerous in wastewater and, therefore, may be more sensitive indicators of contamination and enable a wider range for monitoring virus removal through the wastewater treatment process. Somatic coliphages may represent a broader range of human enteric viruses and, under some conditions, may persist longer in the environment compared to F+ coliphages, thereby acting as a more conservative surrogate. From the perspective of implementing a monitoring program, counting somatic coliphages in plate assays is more practical because the plaques are larger and more readily countable than for F+ coliphages. However, F+ coliphages include ribonucleic acid coliphages that behave more similarly to RNA enteric pathogenic viruses of concern, such as norovirus and Hepatitis A, than somatic coliphages. Somatic coliphages more closely resemble DNA enteric viruses, such as parvovirus and rotavirus [for a list of RNA and DNA pathogenic viruses, see the work by the EPA (2015)].

Fig. 1. Summary of epidemiological studies reporting on the linkage between illness and various indicator types. [Reprinted from Korajkic et al. (2018), under Attribution 4.0 International (CC BY 4.0) license (https://creativecommons.org/licenses/by/4.0/).]

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is an RNA virus of great public interest, but little is known about its viability in environmental waters (Lodder and de Roda Husman 2020). SARS-CoV-2 has been detected in untreated sewage from countries across the globe, including the US (Wu et al. 2020a), India (Kumar et al. 2020), the Netherlands (Medema et al. 2020), and Australia (Ahmed et al. 2020). In the US, billions of dollars have been devoted to research related to SARS-CoV-2, including the development of methods for efficiently and effectively monitoring for SARS-CoV-2 at WWTPs (e.g., NIH 2020). The known primary routes of transmission for SARS-CoV-2 are the inhalation of aerosols and droplets, direct contact with infected individuals, or through contact with contaminated surfaces (Sommerstein et al. 2020). Fecal-oral transmission is an additional hypothetical route of transmission from sewage-contaminated surface waters (Lodder and de Roda Husman 2020), particularly in developing countries with limited infrastructure for collecting and treating wastewater (Usman et al. 2020). More research is needed to determine if SARS-CoV-2 in wastewater-contaminated surface waters poses a human health risk and, if so, to determine if F+ coliphages are suitable indicators for the virus.

CrAssphage

A newly described gut-associated bacteriophage, crAssphage (in reference to the cross-assembly method that was used to identify it), was recently discovered in human fecal microbiomes (Dutilh et al. 2014). The crAssphage has been found to be highly abundant in the human gut (Dutilh et al. 2014; Stachler and Bibby 2014) and in sewage (Ballesté et al. 2019; Tandukar et al. 2020). Bacteroides are the theorized host of crAssphage (Dutilh et al. 2014; Shkoporov et al. 2018), which suggests crAssphage could be useful for source tracking (Stachler and Bibby 2014). Multiple crAssphage qPCR markers have been developed, with some indicating greater human specificity than others (Ahmed et al. 2018c; García-Aljaro et al. 2017; Stachler et al. 2017).

The high concentration of crAssphage in untreated wastewater, wastewater-contaminated streams, and stormwater makes it a sensitive indicator of fecal contamination (Ahmed et al. 2018b; García-Aljaro et al. 2017; Stachler et al. 2018). In a US study, two crAssphage qPCR markers exhibited similar concentrations to the MST marker HF183 [as described by Green et al. (2014a)] in sewer influent samples collected from across the country and in environmental samples from Ohio (Stachler et al. 2017). In a study in Catalonia, Spain, crAassphage concentrations in sewage effluent and river samples were similar to those of human bacterial MST markers [HF183 as described by Haugland et al. (2010) and HMBif] and were more sensitive, but still correlated, with E. coli concentrations (Ballesté et al. 2019). In terms of geographical variation in crAssphage concentrations, Farkas et al. (2019) found lower, but still measurable, concentrations of crAssphage in WWTP influent in the United Kingdom than were found in Tampa, Florida (Ahmed et al. 2018b), and Spain (García-Aljaro et al. 2017). Geographic differences in crAssphage concentrations in untreated sewage had been noted in earlier studies (Stachler and Bibby 2014), potentially due to geographic differences in viral distributions (Stachler and Bibby 2014), levels of industrialization, or diet (Honap et al. 2020).

CrAssphage has shown promise as an indicator of enteric virus persistence in the environment and mixed results as an indicator of enteric virus removal during wastewater treatments. Wu et al. (2020b) found that crAssphage concentrations were strongly correlated with adenovirus and polyomavirus through the activated sludge wastewater treatment process, and bacterial indicators were removed more efficiently than crAssphage, which highlights the need for using a viral indicator to monitor the removal efficiency of viruses at WWTPs. CrAssphage was found to correlate well with concentrations of human enteric viruses in raw sewage (Tandukar et al. 2020). However, the same study also found crAssphage had a higher reduction ratio in both secondary-treated sewage and final effluent (tertiary-treated) than eight human enteric viruses, including human adenoviruses, and noroviruses, indicating it may overestimate the ratio of pathogenic viruses removed at WWTPs. In contrast, Farkas et al. (2019) did not find particularly strong relationships (

r < 0.5

) between crAssphage concentrations in raw sewage and human enteric viruses, with the exception of John Cunningham polyomavirus concentrations.

Other Viral Indicators

The pepper mild mottled virus (PMMOV) is an RNA virus that has been used as an indicator of enteric viral loads. PMMOV is a plant pathogen that affects a wide variety of peppers and was first proposed as a fecal indicator after it was found in high abundance in human feces (Breitbart et al. 2003). PMMOV has been reported in a high frequency of samples taken from surface waters and WWTPs, suggesting it could be an indicator of human fecal pollution and useful as a measure for effective wastewater treatment practices (Rosario et al. 2009; Shrestha et al. 2018; Symonds et al. 2018; Tandukar et al. 2020).

PMMOV showed greater stability than enteric viruses collected from environmental water samples from river water, marine water, and wetland waters (Kitajima et al. 2018), which makes it a conservative indicator of enteric viruses. PMMOV demonstrated 100% sensitivity in untreated wastewater samples (i.e., detected in all samples) across the US (Rosario et al. 2009) and in 19 of 20 surface water samples (95%) collected from four sites on the Chattahoochee River in metropolitan Atlanta, Georgia (Morgan 2016). However, PMMOV has also demonstrated cross-reactivity in chicken, geese, duck, cow, pig, and seagull feces, albeit at concentrations lower than is found in untreated wastewater (Rosario et al. 2009; Symonds et al. 2018).

The tobacco mosaic virus (TMV) is another plant pathogen fecal indicator that infects plants consumed by humans, including tobacco (primarily), cucumber, and tomato (Hu et al. 2011). Because of the widespread global use of tobacco, TMV has been suggested as a useful fecal indicator (Malla et al. 2019; Shrestha et al. 2018; Tandukar et al. 2018). However, few studies have indicated the utility of TMV as a fecal indicator in the US. Similar to PMMOV, TMV has exhibited high sensitivities in untreated wastewater but has also demonstrated cross-reactivity with nonhuman fecal-source samples, including pigs and cattle (Malla et al. 2019).

Recently, Bibby et al. (2019) reviewed several viruses, including PMMOV, that were discovered from a metagenomic approach (shotgun sequencing) and are under development as viral water quality indicators. Bibby et al. (2019) also outlined the pathway from viral discovery to a water quality monitoring tool, including the research needs of new indicators, including the geographic distribution, environmental fate, and viability of potential indicator viruses. A further investigation of human-associated viral diversity via metagenomic tools in combination with methodological advances for concentrating and culturing viruses in environmental samples presents an opportunity to develop improved viral water quality indicators.

Comparison among Indicators of Pathogens and Human Health Outcomes

Many factors, such as the number and diversity of pathogenic microbiota and viruses, preclude the existence of a single indicator that correlates perfectly with pathogen presence/abundance or human health outcomes (Wu et al. 2011). While highly abundant in human feces, fecal coliform, E. coli, and enterococci are known to propagate in environmental settings under particular conditions and occur across endothermic animals (Jang et al. 2017). Despite many recreational water-associated outbreaks attributed to zoonotic pathogens (Craun et al. 2005), research suggests that, with the exception of cattle, exposure to nonhuman fecal contamination generally presents a lower risk to human health than exposure to human fecal contamination (Schoen and Ashbolt 2010; Soller et al. 2010b, 2014). Human-specific, anaerobic (cannot survive outside the host’s gut system), and abundant enteric bacteria would therefore be expected to better correlate with levels of human fecal contamination than E. coli or enterococci and, thus, better predict human health risk from fecal-associated pathogens. Several studies have indicated positive correlations between human-specific (anaerobic and abundant enteric bacteria) markers and the presence of various pathogens [see the study by Harwood et al. (2014) for a review]; however, other studies have generally indicated that human Bacteroidales markers do not correlate well with human health outcomes (Harwood et al. 2014; Korajkic et al. 2018). One theory for this discrepancy is that the majority of diseases caused by fecal-contaminated waters are caused by human enteric viruses, which may behave differently than bacteria under various conditions (EPA 2015). For example, some WWTP processes are generally more effective at removing bacteria than viruses (McMinn et al. 2017), and therefore, bacterial indicators could underestimate the level of viral pathogens present in wastewater effluent. Bacteriophages that infect FIB can occur in high enough abundance to make them sensitive markers, and their persistence in WWTPs and in the environment may be more similar than FIB to human enteric viruses (Wu et al. 2011).

A metaanalysis by Wu et al. (2011) found that the most suitable microbial indicator of pathogen presence depended on the biotype of the pathogen (i.e., protozoal, bacterial, or viral). F+ coliphages were determined to be the best indicator of viral pathogens, and C. perfringens and fecal coliforms were likely useful indicators of all three biotypes of pathogens (Fig. 2). In contrast, E. coli and enterococci did not show any greater likelihood of correlating with pathogens than other indicators (Fig. 2). Wu et al. (2011) recommended that, at a minimum, the use of an indicator of recent fecal contamination (e.g., Bacteroidales) combined with an indicator of more long-term environmental persistence (e.g., human polyomaviruses) should be used to monitor and assess water quality for fecal contamination. Wu et al. (2011) also suggested that studies that failed to detect correlations between fecal indicators and pathogens had insufficient sample sizes to assess such relationships; studies that found correlations between indicators and pathogens generally had a higher percentage of samples positive for pathogens compared to studies that did not detect correlations. Therefore, site-specific monitoring efforts may need a large sample size to accurately assess human health risks and determine the most suitable indicator(s).

Fig. 2. Association (odds ratio, OR) of specific indicators with three classes of pathogens. BP = bacteriophages or coliphages; BPF = F-specific coliphages; BPS = somatic coliphages; CP = *C. perfringens*; EC = *E. coli*; ENT = enterococci; FC = fecal coliforms; FS = fecal streptococci; and TC = total coliforms. [Reproduced from Wu et al. (2011), with permission from the copyright holders, IWA Publishing.]

Harwood et al. (2014) reviewed four papers (Colford et al. 2007; Sinigalliano et al. 2010; Wade et al. 2006, 2008) that compared associations between fecal indicators and human health outcomes, which included GI and respiratory illness, and skin, eye, and ear infections. Human Bacteroidales markers (tested in four studies and included several different assays) and somatic coliphages (tested in one study) were not associated with human health outcomes. GI illness was associated with Enterococcus in two studies (tested in four studies) and with F+ coliphage in one study (tested in one study). Skin infections were associated with Enterococcus in one study (tested in three studies). Respiratory illnesses were associated with F+ coliphage in one study (tested in one study).

A more recent metaanalysis by Korajkic et al. (2018) found that general FIB tended to form statistically significant relationships more frequently with bacterial and protozoal pathogens than MST markers or bacteriophages. However, the differences in the frequency of significant relationships between indicators and pathogens may have been due to the variation in the frequency of detections and concentrations of indictors: FIB were detected in

> 90 %

of the samples, alternative indicators (C. perfringens and bacteriophages) were detected in

> 70 %

, and MST markers were detected in

< 10 %

of samples, with similarly low detections of pathogens. Similar to the conclusion by Wu et al. (2011), low frequencies of detections make it difficult to establish relationships between indicators and pathogens. Korajkic et al. (2018) also investigated the association between fecal microorganisms and human illness. No fecal microorganisms were linked with illness in more than 50% of the studies. F+ coliphage and enterococci were the indicators with the highest proportion of studies linked to illness (5/10 and 11/25 studies, respectively) compared to somatic coliphage, E. coli, total coliforms, fecal coliforms, general Bacteroidales spp. MST markers, and human MST markers from Bacteroidales and Bacteroidales-like organisms (Fig. 1; Korajkic et al. 2018).

With regard to indicators used for regulatory purposes in the US, the EPA outlined different attributes of the two current recommended indicators for recreation in freshwater (enterococci and E. coli) and coliphages (Table 2), for which RWQC (recreational water quality criteria) are under development (EPA 2017). Savichtcheva and Okabe (2006) also compared traditional FIB and alternative fecal indicators, including coliphages, in terms of their occurrence in feces and the environment, persistence and replication in the environment, associations with pathogens and human health outcomes, and the complexity of methods for assessing indicator presence and abundance. Results from epidemiological studies that investigated associations between coliphages and illness are reviewed in the work by the EPA (2015). Although there was not a clear relationship between coliphages and illness from exposure to contaminated recreational waters, five of the eight epidemiological studies included in the 2015 EPA review found associations between the presence of coliphages and swimming-related GI illness, suggesting an association.

Table 2. Attributes of fecal contamination indicators

Indicator attribute	Enterococci (e.g., EPA method 1600)	E. coli (e.g., EPA method 1603)	Coliphages (e.g., EPA method 1602)
Intestinal microflora of warm-blooded animals	Yes	Yes	Yes
Present when pathogens are present and absent in uncontaminated samples	Present when fecal pathogens are present but may also be present in ambient water without fecal contamination. Not indicative of viruses in WWTP effluent	Present when fecal pathogens are present but may also be present in ambient water without fecal contamination. Not indicative of viruses in WWTP effluent	Present when fecal pathogens are present but are likely absent in ambient water without fecal contamination. Better surrogate for viruses than enterococci or E. coli in WWTP effluent
Present in greater numbers than the pathogen (in this case, human viruses)	Depends on source^a	Depends on source^a	In most cases
Equally resistant as pathogens (in this case, viruses) to environmental factors	Not as resistant as viruses	Not as resistant as viruses	Under most conditions
Equally resistant as pathogens (in this case viruses) to disinfection in water and WWTPs	Not as resistant as viruses (except for ozone)	Not as resistant as viruses (except for ozone)	Under most conditions. However, adenovirus is more resistant than coliphages and other enteric viruses to UV inactivation
Should not multiply in the environment	Can multiply in the environment	Can multiply in the environment	Not likely enough to affect criteria levels
Detectable by means of easy, rapid, and inexpensive methods	Yes, but need EPA method 1611 for rapid enumeration. Other easy and rapid methods are available	Yes, but EPA method is not considered rapid (requires overnight incubation). Other easy and rapid methods are available	Yes, but EPA method 1601 needs validation for quantification. Other easy and rapid methods are available
Indicator organism should be nonpathogenic	Generally nonpathogenic^b	Generally nonpathogenic^c	Nonpathogenic
Demonstrated association with illness from epidemiological studies	Yes	Yes	Yes
Specific to a fecal source or identifiable as to the source of origin	Not EPA method 1600, but MST methods being developed	Not EPA method 1603, but MST methods being developed	Not EPA method 1602, but MST methods being developed

Source: Reprinted from EPA (2015).

a

In raw sewage FIB are present in greater numbers than pathogens. Viruses are less vulnerable to treatment processes than bacteria so that they could survive treatment in greater numbers than bacteria.

b

Enterococci can be pathogenic or antibiotic-resistant in some settings, like hospitals, but generally not in ambient water.

c

Enterohemorrhagic E. coli, specifically O157:H7, grows poorly at 44°C and is often negative for beta-glucuronidase, so it is not detected by EPA method 1603 (Degnan and Standridge 2006). Other pathogenic strains could be detected by EPA method 1603.

Compared to the aforementioned fecal indicators, few studies have investigated correlations between crAssphage and pathogen presence in environmental waters (Bivins et al. 2020). Ahmed et al. (2018d) investigated correlations between crAssphage and other sewage indicators [HF183 as described by Green et al. (2014a), E. coli, enterococci] with 11 bacterial pathogens in storm drain outfalls. They found no strong or moderate correlations between crAssphage and any of the potentially pathogenic bacterial markers, whereas culturable E. coli was moderately correlated with enteropathogenic E. coli. Further research is needed to assess associations between crAssphage and viral pathogens in recreational waters (e.g., Farkas et al. 2019) and its association with human health outcomes from exposure to fecal-contaminated recreational waters.

In addition to the strength of association between indicators and pathogen occurrence or human health outcomes, logistical constraints, such as cost, risks to analysts (i.e., potential pathogenicity of the indicator), technical expertise, time to results, and necessary equipment, are often taken into consideration when determining which indicator is most appropriate for monitoring at a given location. For example, a barrier to the implementation of qPCR-based assays could be the need for access to molecular laboratory facilities outfitted for DNA and RNA extraction and qPCR and personnel with specialized training in molecular techniques. Schang et al. (2016) compared the costs, time to results, and subjectivity of results for several analytical methods to measure E. coli and enterococci. Methods compared included enzyme-substrate cultures (IDEXX Colilert and Enteroalert), combined optical fluorescence sensing with bacterial cultures (TECTA system), qPCR (EPA method 1611 for enterococci), and HTS (enterococci only). Cost and measured results were compared relative to IDEXX analysis (approximately $10 for consumables per analysis). The authors concluded that the TECTA system was the least expensive analytical method, and the results were not subject to user bias because of the automated calculation of the results. HTS followed by qPCR were the most expensive methods, with HTS costs for consumables greater by a factor of 30.0 (30,000%) and qPCR costs for consumables greater by a factor of 3.3 (330%) relative to the costs of IDEXX analysis (Schang et al. 2016). However, multiple dilutions of a sample for IDEXX analysis may be necessary when bacteria levels are expected to be high (

> 2,000 MPN / 100 - mL

) to ensure the accuracy of results (Myers et al. 2014), increasing the cost of IDEXX consumables per sample by the cost of the number of dilutions analyzed. Culture-based FIB assays also often require 16–24 h for incubation, which can delay notifications of conditions that potentially pose a risk to human health. Payment et al. (2010) provide a general comparison of logistical constraints among various fecal indicators.

Direct Detection of Pathogens

There is currently no single method to detect all pathogenic microorganisms in a water sample. Although the direct measurement of pathogens presents numerous difficulties, it is possible, and it may be a practical approach for monitoring if a specific known pathogen is determined to be responsible for a majority of illnesses from recreational exposure to a waterbody (a list of common water-borne pathogens is found in Table 1). In the environment, pathogens that are present often occur at concentrations below the limit of detection for conventional methods. However, even at low concentrations, pathogens can cause illness (WHO 2017).

One of the methodological hurdles for detecting pathogens in environmental samples is the need to concentrate them from large volumes of water. The appropriate volume of water to sample is dependent on characteristics of the water (e.g., turbidity) (Francy et al. 2013), pathogen concentrations (Fout et al. 2014), and exposure routes (e.g., drinking water versus recreational exposure). In a comparison among filters for concentrating microbial indicators and viral, bacterial, and protozoal pathogens, Francy et al. (2013) determined that automatic ultrafiltration (UF) had higher recovery and lower variability in recovery rates when averaged across all microorganisms than four other filtration methods that were tested. Further, automatic UF could be scaled up for filtering 200 Ls or more of environmental water samples and was suitable for field deployment. However, Francy et al. (2013) also recommended the use of glass wool filtration if viruses are the target pathogens based on their better performance for the concentration of viruses than automatic UF.

Once the pathogens from a water sample have been concentrated, there are various methods for detection and enumeration. Additional information on pathogen detection, including methods not described subsequently, can be found in the studies by Bonadonna et al. (2019), Maurya et al. (2020), and Haramoto et al. (2018). Immunology-based methods, which use antibodies to identify pathogens, are the gold standard for detection of the protozoan waterborne pathogens Cryptosporidium and Giardia (EPA 2005). Culture-based methods have been considered the gold standard for the detection and identification of many bacterial and viral waterborne pathogens (Haramoto et al. 2018; Maurya et al. 2020), although PCR and qPCR methods are becoming increasingly more common. The purpose of culturing is to amplify the signal of low levels of the pathogen by growth either on selective solid (agar) or in liquid (broth) media. Cell cultures for viruses require an inoculation step in which host cells are infected with the virus. Estimates of viral abundance are then based on the observation of cytopathic effects (effects to host cells as opposed to counting the virus directly).

However, culture-based methods are time-consuming, and they are generally able to detect viable/infective pathogens (i.e., that are alive and able to replicate) as opposed to PCR-based methods, which cannot distinguish DNA or RNA from viable and nonviable pathogens. Therefore, culture-based methods will often underestimate, whereas PCR-methods may overestimate viable microorganisms (Girones et al. 2010). Culture-based methods may be more appropriate than PCR-based methods for monitoring pathogens in the circumstances, such as wastewater monitoring in which disinfection treatments have reduced the viability of pathogens, but they could still be detected by PCR-based methods resulting in an overestimate of the number of infectious microorganisms (Girones et al. 2010).

Culture-based methods of detection rely on the replication of viable organisms, which enables detection if present. PCR-based methods have a similar premise, except instead of the pathogenic organism replicating as with culture-based methods, the DNA or RNA-based template of the organism is replicated via temperature-controlled enzymatic reactions and detected. Often, cell-culturing is used in combination with PCR (i.e., integrated cell culture PCR) to improve further assay sensitivity (Dong et al. 2010), such as when working with environmental samples where pathogen abundance is likely to be low. Many PCR assays are measured with qPCR, which uses a fluorescent signal that increases as the number of DNA copies in the reaction increases to indicate the number of DNA copies in the reaction over time. One drawback of qPCR is the impact that naturally occurring PCR-inhibitory compounds, such as humic acid, can have on assay sensitivity and quantification. Inhibitory compounds make PCR reactions less efficient, which can result in underestimates of true abundance or an increased rate of false negatives. A more recently developed PCR-based method is droplet digital PCR (ddPCR), which is less impacted by inhibitory compounds than qPCR. In ddPCR, the reactions are spread into thousands of droplets in which the outcome (presence or absence) of PCR in each droplet is determined, and the ratio of droplets with the presence/absence is used to quantify the target pathogen. Loop-mediated isothermal amplification (LAMP) is an alternative method of DNA amplification that uses four to six primers, making it highly specific, occurs under isothermal conditions, and, therefore, does not require thermocycling equipment necessitated for PCR. LAMP has primarily been used as a clinical diagnostic tool in developing counties (Mori and Notomi 2009) but has shown promise for detecting pathogens from environmental waters as well (Gallas-Lindemann et al. 2016).

The advent of HTS technologies has enabled DNA and RNA sequencing to become an increasingly common method for detecting pathogens and surveying pathogen diversity. The type of pathogen of interest will determine necessary procedures prior to sequencing from an environmental sample. For protozoal or bacterial pathogens, a PCR step is generally necessary to increase the DNA ratio of the target pathogen type to DNA from nontarget organisms (Cui et al. 2017; Moreno et al. 2018). In contrast to bacterial and protozoal pathogens, a universal marker for all human (enteric) viruses does not exist. However, DNA sequencing markers have been developed for groups of viruses, such as adenovirus, norovirus, Hepatitis A virus, and enterovirus (Bonadonna et al. 2019), which have been used to study viral diversity in wastewater and river water (Ogorzaly et al. 2015).

Biosensors are devices that use bioreceptors (also referred to as biorecognition elements), which recognize the presence of a target biomolecule (e.g., pathogen or fecal indicator) of interest, and a transducer that gives an electrical signal when the bioreceptor has detected the target biomolecule. Advances in biosensor technology show promise for near-real-time monitoring of waterborne pathogens (Guo et al. 2012) and have added potential benefits of improved sensitivity, specificity, and lower cost compared to traditional detection methods. Kumar et al. (2018) summarize recent developments in biosensors and transducer technologies for the rapid detection of pathogens. Nanotechnology is a general term for the use of nanomaterials (at least one dimension in the nanoscale, 1–100 nm) and processes performed at the nanoscale for various applications and products (Bridle et al. 2015). The high surface area to mass and surface area to volume ratios can give nanomaterials and nanosurfaces unique properties (Auffan et al. 2009), such as excess energy at the surface and thermodynamic instability (Jolivet et al. 2004) that can enable different functionality from larger particles (Auffan et al. 2009). Nanotechnologies can be used in combination with biosensors, as components of biosensors, or coupled with other types of sensors (Maurya et al. 2020). Bridle et al. (2015) provide a comprehensive review of how nanotechnologies have been and could be used to monitor for waterborne pathogens.

Quantitative Microbial Risk Assessments

Quantitative microbial risk assessments apply principles from risk assessment (NRC 1983) to estimate the probability of illness from exposure to specific (i.e., reference) pathogens. Information needed to estimate the probability of illness includes the dose (calculated from the concentration of pathogens in the environmental sample and the volume of water ingested), the probability of infection given the dose (

P_{\inf}

), and the probability of illness given an infection (

P_{\inf | ill}

). Values or models for estimated volume ingested,

P_{ill}

and

P_{\inf | ill}

, can be obtained from the literature (e.g., Crank et al. 2019). However, without processing large volumes of water, pathogens can occur in environmental waters at concentrations difficult to detect or quantify with accuracy (Gerba et al. 2018). To address the issue of calculating the dose of a reference pathogen for which accurate concentrations in environmental waters are difficult to quantify, some studies have used a qPCR-based MST (qMST) marker as a surrogate. For qMST-QMRAs, the ratio between the qMST marker and reference pathogen concentrations in the fecal source(s) (e.g., host feces and wastewater) must be determined (e.g., Ahmed et al. 2018a). Ideally, information is also known about how that ratio changes due to environmental factors and during transport between the fecal sources and locations of potential human exposure (Zhang et al. 2019). Once that relationship is determined, the reference pathogen concentration can be extrapolated from qMST data from environmental samples and used to calculate the probability of illness (Zhang et al. 2019).

QMRA analysis has been used for a variety of other applications. For example, Eregno et al. (2016) combined QMRA analysis with hydrodynamic modeling to assess how human health risks change based on hydrologic conditions. E. coli was the surrogate indicator for QMRA development for the reference pathogens norovirus, Campylobacter, Salmonella, Giardia, and Cryptosporidium. Ratios between E. coli and the reference pathogens were determined from sewage. QMRA analysis was combined with hydrodynamic modeling and three different decay rates to estimate the health risk from the reference pathogens during and after rain events at several beaches in a Norwegian fjord. QMRA analysis combined with hydrodynamic modeling could be particularly useful for locations known to be affected by increased bacterial contamination after storm events. QMRA analysis has also been used to predict how different wastewater treatments would affect the likelihood of infection for recreational users receiving recreational waters. Purnell et al. (2020) investigated the changes in the probability of infection from several pathogens (adenovirus, Salmonella, and Cryptosporidium) during various water-based recreational activities under different water treatment and flow augmentation scenarios for a river in southwest England. The authors used estimates of pathogen concentrations in river water and wastewater effluent, predicted pathogen reduction ratios from additional reclaimed water treatment, and predicted the percent of increased river flow from augmentation with reclaimed water to predict the likelihood of illness under the different scenarios. They concluded that additional treatments to reclaimed waters used for flow augmentation were necessary for reducing the risk of illness from adenovirus due to recreational exposure while canoeing, fishing, kayaking, and rowing. Additional applications for QMRAs have included studies such as the determination of likely etiologic agents responsible for illnesses from recreational exposure to fecal-contaminated waters in the Great Lakes (Soller et al. 2010a) and the prioritization of areas for wastewater and drinking water pipe rehabilitation by combining QMRAs with spatial distribution system modeling in Hyderabad, Pakistan (Jamal et al. 2020). QMRAs can be a powerful tool when accounting for human health risks from waterborne pathogens across a wide range of issues.

Research Gaps

One of the major issues when developing models of pathogen concentration or human health risk as a function of fecal indicators is differences in the fate, transport, and survival of different pathogens and indicators throughout the stages of wastewater conveyance and treatment and under various environmental conditions. Uncertainties in recovery efficiencies among methods for concentrating, detecting, and quantifying pathogens and indicators add noise to data that would help clarify some of these differences in pathogen and indicator behavior under various conditions. Difficulties detecting and quantifying pathogens in combination with a lack of highly specific and sensitive MST markers for nonhuman sources also inhibit the ability to assess health risks from exposure to nonhuman sources of contamination. Areas for continued research to assess human health risks more accurately from pathogens in recreational waters include but are not limited to the following:

•

Determining recovery efficiencies for pathogens (particularly viral pathogens) from different concentration and extraction methods. This information is critical for comparing results across studies and more accurately estimating concentrations from environmental samples (Haramoto et al. 2018).

•

Determining fate, transport, and responses to environmental factors (including effects of differing WWTP treatment methods) for different fecal indicators, MST markers, and pathogens. This information is critical for accurately modeling the relationships between indicators and pathogens across different sample matrices (Zhang et al. 2019). Boehm (2019) provides an example of how simulation studies may help predict behavior in environmental waters of pathogens and indicators of different types from different sources.

•

Developing nonhuman MST markers with high sensitivity and specificity (Boehm et al. 2013) and host-specific markers that serve as better surrogates for pathogens (Harwood et al. 2014).

•

Determining the health risks from exposure to fecal contamination from different sources (EPA 2012).

•

Developing low-cost and highly sensitive technologies for detecting and quantifying pathogens. Innovation with biosensors and nanotechnology presents opportunities for growth in this area (Bridle et al. 2015).

Conclusion

Recreational waters in the US have been monitored in a regulatory capacity for indicators of fecal contamination for decades (EPA 1976). However, the regulatory emphasis on FIB, which is general to a wide range of hosts and, in some instances, can survive and reproduce in the environment, may result in the allocation of limited resources in improving water quality conditions of waters for which FIB may overestimate the risk to human health. To address this potential discrepancy, a variety of alternate bacterial, host-specific, and viral indicators have been developed in an attempt better to capture the health risk from exposure to fecal-contaminated waters. However, the diversity of possible pathogens and differences in survival, fate, and transport characteristics of pathogen types prevent the existence of a single ideal indicator for all pathogens. Further, pathogen presence and abundance can vary spatially and temporally. Therefore, optimal indicators for predicting human illness are likely to differ geographically and may change based on the season or over time for a given location. If a specific pathogen is known to be of concern or issue, technologies do exist for the direct monitoring of pathogens. However, if the goal is estimating the human health risk from exposure to specific pathogens, statistical methods (e.g., QMRAs) can help circumvent the logistical difficulty of direct pathogen monitoring. Technological advances in DNA and RNA sequencing, biosensors, and nanotechnology may help in the future for characterizing and monitoring pathogens in aquatic systems. However, until then, the optimal approach and method for assessing the human health risk in recreational waters impaired by fecal contamination may differ by location and entity based on the pathogens of concern and resources available for monitoring.

Data Availability Statement

No data, models, or code were generated or used during the study.

Acknowledgments

Funding to write this article was provided by a cooperative agreement between the US Geological Survey and Gwinnett County, Georgia, with guidance from Barbara Seal of the Gwinnett County Department of Water Resources. We thank Haley T. Olds at the USGS and anonymous journal reviewers for their reviews of the manuscript. Joseph Duris (USGS) and Marirosa Molina (EPA) provided helpful input on the logistical constraints of fecal indicators. The work by Marcella Cruz was done while serving as a Pathway Student with the US Geological Survey. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the US Government.

References

Ahmed, W., et al. 2020. “First confirmed detection of SARS-CoV-2 in untreated wastewater in Australia: A proof of concept for the wastewater surveillance of COVID-19 in the community.” Sci. Total Environ. 728 (Aug): 138764. https://doi.org/10.1016/j.scitotenv.2020.138764.

Abstract

Introduction

Methods

Background

Fecal Indicator Bacteria

Microbial Source Tracking

Metabarcoding and Community-Based Microbial Source Tracking

Management Applications of Microbial Source Tracking

Viral Source Tracking

Viral Indicators

Bacteriophages

Coliphages

CrAssphage

Other Viral Indicators

Comparison among Indicators of Pathogens and Human Health Outcomes

Direct Detection of Pathogens

Quantitative Microbial Risk Assessments

Research Gaps

Conclusion

Data Availability Statement

Acknowledgments

References

Information

Published In

Copyright

History

ASCE Technical Topics:

Authors

Affiliations

Metrics

Citations

Download citation

Cited by

Figures

Other

Share

Copy the content Link

Share with email

Share

Request Username

Create a new account

Change Password

Password Changed Successfully

Verify Phone

Congrats!