Wastewater-based surveillance as a tool for public health action: SARS-CoV-2 and beyond

SUMMARY

Wastewater-based surveillance (WBS) has undergone dramatic advancement in the context of the coronavirus disease 2019 (COVID-19) pandemic. The power and potential of this platform technology were rapidly realized when it became evident that not only did WBS-measured SARS-CoV-2 RNA correlate strongly with COVID-19 clinical disease within monitored populations but also, in fact, it functioned as a leading indicator. Teams from across the globe rapidly innovated novel approaches by which wastewater could be collected from diverse sewersheds ranging from wastewater treatment plants (enabling community-level surveillance) to more granular locations including individual neighborhoods and high-risk buildings such as long-term care facilities (LTCF). Efficient processes enabled SARS-CoV-2 RNA extraction and concentration from the highly dilute wastewater matrix. Molecular and genomic tools to identify, quantify, and characterize SARS-CoV-2 and its various variants were adapted from clinical programs and applied to these mixed environmental systems. Novel data-sharing tools allowed this information to be mobilized and made immediately available to public health and government decision-makers and even the public, enabling evidence-informed decision-making based on local disease dynamics. WBS has since been recognized as a tool of transformative potential, providing near-real-time cost-effective, objective, comprehensive, and inclusive data on the changing prevalence of measured analytes across space and time in populations. However, as a consequence of rapid innovation from hundreds of teams simultaneously, tremendous heterogeneity currently exists in the SARS-CoV-2 WBS literature. This manuscript provides a state-of-the-art review of WBS as established with SARS-CoV-2 and details the current work underway expanding its scope to other infectious disease targets.

INTRODUCTION

Wastewater-based surveillance (WBS; also known as wastewater-based epidemiology) is an emerging science that systematically analyzes residues in wastewater to understand the populations from which they derive. Intrinsic to WBS is its unique ability to monitor an entire population (irrespective of symptoms or the ability to access clinical testing) objectively, passively, and comprehensively, thus being inclusive and representative (1). Analogous to performing phlebotomy to assess for hematologic and biochemical markers of disease in a patient, wastewater can be collected dynamically within the sewershed and specific analytes measured to provide information on the health of the entire monitored population. Prior to the coronavirus disease 2019 (COVID-19) pandemic, WBS was predominately used to assess for chemicals in communities, including substance of abuse, tobacco, and alcohol (determinants of population health) and their distribution across space and time (2–5). Often, these data were correlated with static cross-sectional surveys (e.g., drug usage, alcohol and pharmaceutical sales, etc.) and then generalized across the population over time (6–8). The COVID-19 pandemic provided a unique opportunity to demonstrate the power and resolution of WBS. Widespread and rapid rollout of clinical testing combined with real-time data transparency on the burden of COVID-19 in communities provided a robust evidence base to compare with wastewater-measured SARS-CoV-2 RNA.

Early SARS-CoV-2 WBS pioneers adapted existing workflows used for monitoring chemical agents and/or viral causes of gastroenteritis (see Moving beyond SARS-CoV-2 to other pathogens) and developed assays that identified and quantified SARS-CoV-2 RNA in wastewater (9, 10). Shortly thereafter, these groups demonstrated strong spatial and temporal correlations between wastewater-measured SARS-CoV-2 RNA and clinical disease burden, prompting rapid adoption across the globe. Owing to the range of technical skills required for WBS, transdisciplinary teams with expertise involving engineering, water science, environmental and medical microbiology, infectious disease and public health, and bioinformatics were formed, spurring further innovation and rapid advancement of the field with more than 1,500 PubMed indexed articles on the topic (wastewater and SARS-CoV-2) within 3 years of the pandemic being declared. In this time, WBS has evolved into one of the hottest areas of science and medicine and captured the attention of researchers, clinicians, policy makers, and the public.

In this manuscript, we provide a state-of-the-art review of the field of WBS using SARS-CoV-2 as the test case. We detail the various aspects of the wastewater workflows including the different approaches associated with wastewater collection, concentration, cleanup, quality control and assessment, nucleic acid extraction, target quantification, and data analysis (Fig. 1). We discuss the clinical and ethical implications of this work, and how these lessons apply to additional infectious analytes, and the early data supporting this work.

Fig 1.

Wastewater-based surveillance workflow. (A) Wastewater can be collected at multiple points in the sewershed, providing objective data on the monitored population. Careful handling and expedient transport are required to ensure sample integrity. (B) Concentration and extraction of wastewater are required in order to increase the relative abundance of SARS-CoV-2 RNA (or for other infectious targets, DNA) and to remove and mitigate inhibitory substances. (C) A range of molecular assays including (RT) qPCR, ddPCR, and LAMP can be used to quantify SARS-CoV-2 RNA, and variant detection can be performed using allele-targeted assays or agnostic metagenomic assessments. (D) Wastewater-measured SARS-CoV-2 can be adjusted based on a number of extrinsic factors (i.e., flow rates) and intrinsic factors (i.e., normalization for fecal biomarkers) and then compared with clinical data to develop appropriate models of community infection.

SARS-CoV-2 AS A MODEL CANDIDATE FOR WASTEWATER SURVEILLANCE

The COVID-19 pandemic

The first cases of COVID-19 were identified in the City of Wuhan in the Hubei province of China at the end of 2019. Rapidly thereafter, clusters of community-acquired pneumonia, especially those centered around the Hunan Seafood Wholesale Market were reported (11). By 11 February 2020, the World Health Organization (WHO) had designated the virus SARS-CoV-2 as responsible for the clinical disease COVID-19 (12). On 11 March 2020, a global pandemic was declared (13). As of March 2023, more than 750 million cases were reported globally, with almost 7 million related deaths. This number underestimates the true burden of the disease, given the limited availability of testing in some areas and the limited role of clinical testing later in the pandemic, particularly since the onset of the Omicron BA.1 wave, which spread so rapidly that it exceeded clinical testing capacity (14). Indeed, seroprevalence studies demonstrate a globally diverse range of seropositivity, all considerably higher than diagnosed COVID-19 cases in those areas (15).

SARS-CoV-2: biology

Coronaviruses are enveloped, positive single-stranded RNA viruses (16). The family of Coronaviridae is divided into the subfamily Orthocoronavirinae, comprising four genera, alphacoronavirus, betacoronavirus (both infecting only mammals), gammacoronavirus, and deltacoronavirus (infecting a wider range of animals, including birds). Before COVID-19, four endemic human coronaviruses were known, HCoV-229E, HCoV-OC43, HCoV-NL63, and HCoV-HKU1. Infections by these endemic coronaviruses generally manifest as mild respiratory symptoms and are considered “common colds.” Other pandemic-capable viruses belonging to the betacoronavirus genus include SARS-CoV-1 (17) (originally referred to as SARS) and MERS-CoV (18, 19). In fact, SARS-CoV-2 shares 79% sequence identity with SARS-CoV-1 and 50% with MERS-CoV and clusters with coronaviruses found in bats, within the subgenus Sarbecovirus of the genus (20, 21).

The genetic sequence of SARS-CoV-2 was determined very quickly. The first reported genome was sequenced and shared via the GISAID database by 12 January 2020 (22). The genome is ~30,000 bp in size. At the upstream region, it contains a 5′ cap and a 3′ poly (A) tail at the downstream region (23). Ten open reading frames (ORFs) (1a, 1b, 3a, 3b, 6, 7a, 7b, 8a, 8b, and 9b) with four structural protein-coding ORFs have been identified: Spike (S), Envelope (E), Membrane (M), Nucleocapsid (N), and accessory factors (24). The single-stranded RNA genome is wrapped by N, whereas M and E proteins ensure the N with positive-sense RNA is incorporated in the viral particle during assembly. The S protein protrudes from the host-derived viral envelope and provides specificity for cellular entry receptors via the angiotensin-converting enzyme 2. Because of receptor homology in other mammals, SARS-CoV-2 has been recovered from a wide range of species.

Through mutation and selection, new variants of SARS-CoV-2 emerged and were disseminated in successive waves of infections across the globe (14, 25–29). Variants of concern (VOCs) by definition have any of: (i) increased replication advantage, (ii) increased immune evasion, thereby increasing the risk of reinfection and breakthrough infections despite prior vaccination, and (iii) increased potential for causing severe disease. Highly successful VOCs included Alpha (B.1.1.7 lineage), Beta (B.1.351 lineage), Gamma (P.1 lineage), Delta (B.1.617.2 lineage), and Omicron (B.1.1.529 and XBB.1.5 lineages and there many associated sub-variants). Omicron was first reported in Botswana, and shortly thereafter in South Africa. The United States Centers for Disease Control and Prevention track US clinical data in an interactive web display, https://covid.cdc.gov/covid-data-tracker/#variant-proportions (29).

COVID-19 clinical disease

SARS-CoV-2 is transmitted from person to person (30). Transmission generally occurs through close contact either directly or through respiratory droplets and aerosols (31). The risk of transmission exists even before symptom onset and is highest in the acute phase of the disease. In immunocompetent populations, transmission risk after 7 days is very uncommon (32) although viral RNA may be shed for much longer periods (see below) (33). The risk of infection depends on the type of exposure, with the greatest risk associated with prolonged close contact (34, 35). While many infections are not propagated, certain individuals are responsible for the wide dissemination of COVID-19. These individuals, commonly called “super spreaders,” are identified when large clusters of secondary infections can be traced back to a single index case (36, 37)

The disease spectrum of COVID-19 ranges from asymptomatic to mild disease to life-threatening illness. Asymptomatic infections vary by VOC and are generally more common in children, of whom as many as 30%–40% do not develop symptoms (38, 39). In those who develop symptoms, illness onset begins a median of 4–5 days after infection (range 1–14 days) (40, 41). The most common symptoms are fever, non-productive cough, fatigue (42–44), and occasionally the loss of sense of taste and smell (45, 46). In one of the first published cohorts involving 138 patients from Wuhan who were hospitalized for pneumonia due to SARS-CoV-2, dyspnea occurred after a median of 5 days after the first symptom onset, and hospital admission occurred a median of 7 days after symptom onset (40). Early in the pandemic, severe illness manifesting with dyspnea and hypoxia was seen in 14% of patients, and critical illness with respiratory failure and/or multi-organ dysfunction in 5%. Overall attributable fatalities in the early days of the pandemic were approximately 2.3% (47, 48). Severe illness can occur at any age but disproportionally occurs in older people or those with specific co-morbidities, including hypertension, diabetes, and immunosuppression (49–51). The Omicron variant has a shorter incubation time (3.8 days, 95% CI: 2.7–5.5) and a higher proportion of pre-symptomatic transmission than Delta variants; 62% vs 48%, respectively (52).

SARS-CoV-2 viral RNA shedding

While COVID-19 is most commonly recognized as a respiratory illness, SARS-CoV-2 infects multiple organ systems and is responsible for a wide array of clinical presentations (42, 43, 49, 51). The detection of SARS-CoV-2 RNA in wastewater represents contributions coming from multiple body fluids including feces, swallowed saliva and sputum, and, to a much lesser extent, the urine of infected individuals (53–55). In particular, the gastrointestinal tract is recognized as a site of SARS-CoV-2 infection, with common reports of abdominal pain, anorexia, nausea, and diarrhea, with or without concomitant respiratory involvement. In some individuals, gastrointestinal disease may be the sole manifestation of COVID-19 (56). Detection of SARS-CoV-2 in stool samples has been reported in 40%–50% of cases with diarrhea and 14% without diarrhea. In fact, diarrhea is reported to be the first manifestation of COVID-19 in 23.3% of cases (57–60). Individual viral shedding dynamics vary markedly. A meta-analysis by Cevik et al. (53) suggested that shedding occurred for a mean duration of 17 days (±0.8, CI 95%) in feces and begins even prior to symptom onset (32, 61). The study also reported no significant difference in SARS-CoV-2 viral loads in stool samples from persons with symptomatic COVID-19 infection relative to those who were asymptomatic, but a shorter duration of viral shedding was observed in asymptomatic infections (53). However, detailed data to establish fecal shedding dynamics and when peak shedding occurs have yet to be established (53).

Correlating COVID-19 occurrence and wastewater-measured SARS-CoV-2 RNA

WBS teams around the world have reported the correlation of SARS-CoV-2 RNA in municipal wastewaters measured at the level of wastewater treatment plants with COVID-19 clinical disease diagnosed in the monitored population (9, 62–68). Similar studies have established the critical role WBS plays in the geospatial modeling of disease across communities, providing critical information to decision makers and policy experts, and enabling real-time public health decision-making (69, 70). Relative to clinical testing, wastewater measurements have many great strengths. WBS does not suffer from limitations such as selective population sampling bias (i.e., the potential exclusion of members from some disadvantaged classes), does not suffer from testing fatigue and availability, and is not impacted by lags in clinical reporting (68). However, for WBS to function, providing real-time monitoring from a well-designed collection, processing, analysis, and reporting framework is required to generate a leading indicator relative to clinical case data (71).

Wastewater data correlates with clinical disease in the community and predicts health resource utilization. Several studies have demonstrated that hospitalization rates correlate strongly with SARS-CoV-2 RNA measured in associated sewersheds, providing a variable lead time (71–74). Furthermore, this correlation held true even for the most precious of hospital resources during the pandemic: ICU beds (72, 74). At many points in the pandemic, critical illness requiring ICU admission for life-saving supportive care far exceeded limited bed availability (75, 76). For example, Schenk et al. (74) monitored SARS-CoV-2 RNA across 123 wastewater treatment plants throughout Austria and established that wastewater measurements provided a lead time relative to general hospital admissions (8.6–11.6 days) and ICU (14.8–17.7 days) (although the lead time was shorter during the Omicron wave relative to the preceding Delta wave). COVID-19-related deaths have also been correlated with wastewater-measured SARS-CoV-2 (77, 78).

WBS has proven insightful in other areas of society beyond healthcare. For example, Acosta et al. correlated workforce absenteeism over a 2-year period with wastewater SARS-CoV-2 RNA abundance measured in Calgary, Canada demonstrating a 1-week lead time (79). Accordingly, WBS could be used by employers to optimize workforce distribution to ensure efficiency and cost-effectiveness while ensuring their mandates can be fulfilled.

The COVID-19 pandemic created the ideal circumstances to develop, implement, and hone WBS. Below, we detail the steps and supporting evidence behind WBS procedures and workflows.

SAMPLE COLLECTION

General considerations

For WBS to be an effective and reliable public health tool, representative wastewater samples need to be obtained, processed, and analyzed in a reproducible manner, which yields consistently accurate data and reduces the overall variability throughout the sample acquisition, handling, and analysis workflow (80, 81). Inaccurate data are worse than no data.

One of the greatest challenges associated with WBS is obtaining truly representative wastewater samples. Of all the sources of variability in the WBS workflow (i.e., collecting samples, processing, testing, analyzing, and reporting), sample collection is likely the largest source of variability since environmental samples, especially wastewater, are inherently variable (82–86). This challenge is further heightened by the often-intended quantitative use of the data. Unlike patient-specific clinical data that are often conveyed as a binary “state variable” (i.e., the patient’s infection status is either “positive” or “negative”), the most useful form of WBS data is quantitative (e.g., number of gene copies/mL), theoretically allowing for trends to be analyzed quantitatively or even action thresholds to be developed. This type of data use requires fairly stringent data-quality objectives, which begin with obtaining the most representative samples possible (87, 88). The biggest opportunity for obtaining those representative samples and increasing the monitoring accuracy is the proper design of the collection program.

Sample sources and locations

Achieving samples that are representative of the population under study requires an understanding of what factors contribute to the wastewater being collected. Wastewater collection systems are dendritic networks, running from the originating nodes (within a building) through increasing higher-order trunk lines (sewer pipes) until the terminal node (wastewater treatment plant or discharge to the environment). Sampling can be done at many locations throughout this network (88, 89).

Theoretically, the “most representative” wastewater sample (i.e., least contaminated with non-fecal inputs) that can be obtained for WBS is at the original source (i.e., the toilet). Every downstream point thereafter typically introduces other sources of wastewater (e.g., laundry and kitchen gray water, commercial, or industrial inputs) and dilution water (e.g., clean water discharges, rain and snow meltwater, and groundwater leaking into sewer pipes). Transit time through the system also changes the wastewater’s chemistry and biology (80). Therefore, different locations within the collection system will have different source contributions and degradation profiles resulting in different, location-specific wastewater characteristics, all of which affect analytical results (90). Because the wastewater of each municipality has a unique blend of these sources, it is not at this time advisable to directly compare WBS data between municipalities or even between locations within a single municipality’s collection system.

Given the structure of wastewater collection systems, a “nested scales” sampling regime can be used for WBS (66) (Fig. 2). Samples can be taken at a wastewater treatment plant or lift station that captures wastewater from the whole municipality. On a more “zoomed-in” scale, select areas (i.e., neighborhoods) can be monitored by sampling at maintenance holes or other access points (66, 91–93). Individual buildings or sections of a building (e.g., a hospital wing or ward) can be monitored by sampling at specific access points within that building’s plumbing system (94).

Fig 2.

Wastewater-based surveillance can be conducted using a “nested approach.” (A) A schematic representation of the wastewater-based surveillance program in the province of Alberta, Canada that enables monitoring of 83% of the population through 43 municipalities. (B & C) Key nodes within the sewershed can be separately assessed in large municipalities to provide granular data on the populations that comprise these sub-catchments as profound regional variation exists between neighborhoods in population makeup (i.e., differences in social, economic, and ethnodemographics). (D) Targeted facilities can serve a specific sentinel role for monitoring specifically at-risk populations (i.e., hospitals and long-term care) and how infections propagate.

Whereas wastewater treatment plants are typically equipped with sampling devices for routine collection of environmental compliance and quality control samples, sampling in neighborhoods and individual buildings brings unique challenges. These include the need to perform ongoing maintenance of the sampling equipment and provide adequate support and training to the field team. Safety is another key consideration, due to hazards such as potential exposure to sewer gases (i.e., hydrogen sulfide) (95), working above open pits, and traffic control or associated lane closures when the sample location is within a roadway. From a program design perspective, defining the boundaries of the sewershed being sampled and hence the population being monitored can also be challenging, especially if detailed records of the sewer pipe networks are not available. Similarly, within a building, difficulties can arise when attempting to identify suitable within-building plumbing access locations that provide coverage of the targeted area of the building (whether that be the entire building or a specific wing) while avoiding problematic sample locations (e.g., in a busy hallway) (96, 97). Sewershed usage patterns also need to be taken into consideration—for example, “buildings of visitation” (restaurants, sports stadiums, and schools) will have different toileting usage (i.e., the relative proportion of people defecating at those sites) than “buildings of residence” (houses and apartments, dormitories, etc.), which can dramatically affect the data (66).

Sampling at different scales within the sewershed can yield data suitable for different purposes and for the population of interest of a given WBS program. Generally, the more aggregate the wastewater source (e.g., wastewater treatment plant), the less variable the wastewater, the greater the monitored population (which reduces costs/capita), and the more the data are suited for general trend analysis and high-level community monitoring. Conversely, the more zoomed-in the sampling location is within the sewershed, the more temporal variability exists in wastewater strength and quality (because discrete events like toilet flushing or other discharge events have a greater proportional effect). The fewer people captured in the aggregate monitored population, the more actionable the data can be (e.g., if building-specific WBS detects a rise in SARS-CoV-2 in a homeless shelter or university, action can be taken to respond to that localized situation) (98–100). The appropriate lower bound of zooming in is likely based on ethical considerations (i.e., keeping a minimum size of the monitored population to avoid being able to identify individuals) and per capita cost rather than technical limitations (87, 88, 101, 102).

Sampling devices and approaches

The simplest sampling method for wastewater is to take a “grab sample” by collecting a sample at a single point in time from a location. However, this is also the least representative method (80, 103) due to the variability caused by the spatial considerations discussed above and the significant temporal variability inherent in wastewater. Even at its most aggregated location (the wastewater treatment plant), a distinct diurnal pattern is often observed caused by the population’s usage patterns of the facilities that contribute to wastewater (86). The farther upstream in the sewershed, the greater the temporal variability for the reasons discussed above. One method to address this temporal variability is to collect a composite sample over a given period, typically 24 hours, which smooths out the temporal variability (80, 104). This is usually done using an autosampler: essentially a programable pump plus the associated tubing and collection vessel(s), typically contained in an outer case. Autosamplers are available commercially from several sources. Custom-built autosamplers are also available and are well-suited for certain applications (e.g., in-building sampling) (105, 106). Numerous considerations need to be addressed when determining where and how to deploy autosamplers (87). Safety during deployment and retrieval is a major consideration. If not deployed at a secured location, strategies to prevent the public from interfering with the sampler are critical from a safety (i.e., unintended access to wastewater) and data quality perspective (e.g., interference with the samples or programmed sampling regime). Temperature control inside the sampler is also key in warm (preventing heat-related degradation of samples) and cold weather (preventing freezing). Indeed, several studies have documented nucleic acid degradation of SARS-CoV-2 and other viruses associated with both freezing and warming of wastewater (107, 108). This is even more important for live WBS targets such as antimicrobial-resistant organisms, where increased temperature deviations in the sewershed can manifest in changes in population structure owing to replication and amplification occurring outside of the human host (109). Sample collection can be impeded by a number of factors including tubing blockages, rags, and other wastewater debris blocking the inlet, changes in water levels, and the like. Therefore, sampler maintenance by trained personnel is essential.

Autosamplers can be programmed for time-weighted or flow-weighted sampling (110). Time-weighted sampling takes aliquots of equal volume at regularly spaced intervals, whereas flow-weighted sampling takes aliquots proportional to the total flow passing the sample location at each sampling interval. Flow-weighted sampling yields more representative samples but requires a flow meter to be linked to the autosampler and additional software to control the aliquot sampling dynamically. This makes flow-weighted composite sampling more complicated and thus less common. It also precludes certain sample locations as it is often impossible or cost-prohibitive to install a flow meter properly in some pipes (e.g., in a building’s sewer pipes or inside the smaller, inaccessible sewer lines beneath a neighborhood). Due to its importance in data interpretation, details of the sampling techniques should be included when reporting results.

Wastewater can be separated into solid and liquid fractions. Designing a sampling program for target analytes (e.g., pathogens and chemical targets) in WBS requires knowing: (i) which wastewater fraction the targets are primarily associated with, (ii) how long the targets stay associated with that fraction once they are in a mixed solution, and (iii) how these wastewater fractions behave at different points in the wastewater system. Within the sewer pipe network, the solid and liquid fractions are typically well-mixed. Once the wastewater arrives at a wastewater treatment plant, various treatment processes intentionally separate solids from liquids. Therefore, compared to the “raw” wastewater in the sewer system, different sample locations within a wastewater treatment plant will either have more solids (e.g., primary or secondary sludge and digestors) or less (e.g., clarifier effluent and supernatant). This will affect the abundance of target analytes(s). For example, enveloped viruses contain a lipid bilayer (like that of their hosts) surrounding the viral protein capsid that sits inside this membrane. This hydrophobic exterior can cause enveloped viral particles to be more associated with the solid fraction than the liquid fraction of wastewater (111–113). Since the design of a sampling program will significantly affect results, details of the sampling locations should be included when reporting results, especially when sampling within a wastewater treatment plant.

Composite samples and grab samples, under certain conditions, can produce quantitative data (i.e., number of gene copies/mL) (89, 114). These are “active” sampling methods. “Passive” sampling methods also exist. These methods suspend some form of absorbent media in the wastewater flow that will preferentially absorb/adsorb the target analyte(s) from the passing wastewater (86, 115). The media selected must also be capable of freely releasing these same targets when the media are subjected to the subsequent elution process. It is advisable to validate the absorption/adsorption and recovery of new analyte targets in the lab before initiating a field sampling protocol.

For WBS of SARS-CoV-2, the adsorption target is the suspended solids in the wastewater, as this virus is primarily shed in the feces of infected individuals (116–118), and therefore primarily associated with the solid fraction of the wastewater (see above) (61, 62, 119) at least until enough mixing and transit time occur to elute some of the virus particles into the liquid fraction (120, 121). Passive sampling for SARS-CoV-2 RNA, which generally employs simple adsorption media (e.g., a classic Moore swab, using cotton, gauze, or tampons), is often less expensive and simpler to use for the actual sampler deployment component of the workflow (115). However, most wastewater treatment plants already have autosamplers permanently installed as part of their routine monitoring for regulatory compliance. Also, deploying samplers of any kind for WBS within the sewershed will typically require opening maintenance holes (with the associated safety barricades and possible road closures). Thus, while passive samplers avoid the maintenance requirements of composite samplers, they still face similar deployment challenges as active samplers, which may outweigh sampler maintenance efforts.

The main limitation of passive samplers is that they typically cannot deliver data that can be converted into a representative concentration (122). This is because they separate (concentrate) the analyte from the bulk wastewater, and the original volume from which it was concentrated is unknown (i.e., the denominator in the concentration expression is missing). Some passive samplers, especially those designed to monitor environmental chemistry (e.g., Polar Organic Chemical Integrative Sampler [POCIS]), overcome this limitation by measuring the total flow that passes the sampler (123). Of course, this negates the original advantages of simplicity and lower cost. Passive samplers may best be suited for providing qualitative (e.g., presence/absence) data (as may be deployed in high-risk facilities such as long-term care). Since they also concentrate the target analyte, they can effectively achieve a lower detection limit and better sensitivity than a grab sample (124), which can be appropriate for sentinel applications where the switch from non-detectable to detectable is the key monitoring objective (87). This is a typical situation in environmental chemistry, where passive samplers are more commonly used. Studies have reported discrepant results as to whether passive samplers have improved sensitivity compared to composite samplers (125, 126).

Collection and use of associated data

In addition to collecting data on target pathogens (e.g., SARS-CoV-2), collection and use of wastewater-associated “metadata” are often useful during quality assurance and data interpretation (80). Typical metadata include sample location, collection date and time, total volume collected, type of sample (grab, time-weighted composite, etc.), flow data (if available), weather (especially recent rain events), ambient temperature during collection and transport, shipping conditions [e.g., both freezing and elevated temperatures can dramatically decrease recovery of SARS-CoV-2 RNA (127, 128)], and any field issues with sampler or associated observations (e.g., clogging or ragging of the inlet line and higher or lower flows than typical).

Collecting flow data, which is often easier at the wastewater treatment plant than farther upstream in the sewershed, allows the concentration of the target analyte(s) to be converted into a numeric flux for pathogens (e.g., gene copies/day) or a mass flux for chemicals (e.g., mg/day). This is sometimes referred to as “flow normalizing” the data but is actually just a data conversion process. Unlike concentration data, flux data from adjacent sewersheds (e.g., in a city or region with multiple wastewater treatment plants) can be summed together to give a total value for the overall region (66). Flux data can also be used in other analyses, including assessments per capita of general wastewater loading and associated trends. Using the population-adjusted level of pathogen detected in wastewater is another approach to aggregate WBS of different sites (67).

A significant challenge that WBS remains to overcome is standardization. A particular challenge within this context is the ability to understand how differences in a measured target signal can be interpreted between vastly different populations (i.e., large cities vs towns or neighborhoods). One tool being explored in this regard relates to secondary signals that, in theory, enable an objective and reliable correction or “normalization factor” between populations (Fig. 3). Ultimately, being able to accurately determine how many people contributed to the wastewater from which a signal is measured at a given point in time may allow for more accurate analysis and use of WBS data (87, 129). Approaches to normalize the measured pathogen target are discussed in more detail below.

Fig 3.

Wastewater-measured analytes may be normalized against both population and biologically relevant factors to increase the validity and reliability of site-to-site comparisons. (A) Communities differ in their population size, demographics as well as the nature of water usage—and therefore the makeup of their wastewater output. Understanding dynamic concentrations of target analytes is therefore dependent on factors relating to wastewater flow, contributions, and population size. (B) Behavioral factors relating to toileting patterns likewise may influence the accurate interpretation of wastewater-measured analytes. If an analyte is shed in the feces, measuring the analyte relative to a fecally shed biological marker may improve its correlation with case occurrence in the defined population. This may matter more on a granular scale such as individual facilities (i.e., a hospital/long-term care facility vs a bar or public venue) where defecation is expected to occur among patients/residents but not patrons transiently visiting an establishment.

FACTORS TO UNDERSTAND THE POPULATION FROM WHICH SAMPLES DERIVE AND TO ALLOW FOR NORMALIZATION

The ability to measure analytes and microbial targets in wastewater depends on many factors beyond the release/shedding of that analyte/target in the population being studied. For example, in combined sewer systems, stormwater from precipitation/snowmelt enters the wastewater collection system and can increase the flow of wastewater substantially, sometimes as high as 3–10-fold at the wastewater treatment plant (130). Compensating/correcting for the receiving volume can be insufficient. Also, sampling sites that are proximal to the wastewater treatment plant (i.e., in neighborhoods and individual facilities) generally cannot measure wastewater outflow (131). Additionally, all wastewater systems can be impacted by variable groundwater seepage into the wastewater collection system. As many wastewater treatment plants receive a highly variable amount of non-feces-containing wastewater from non-resident sources (i.e., commercial and industrial discharges), changes in volume received at the plant may not reflect the true burden of human feces passing through. Furthermore, there are significant variations in terms of the contribution of fecal matter into the wastewater system on an individual level by time scale, i.e., the population being monitored.

Normalizing fecal content to adjust the measured quantity of the pathogen target in WBS has long been advocated for SARS-CoV-2 WBS, with the concept that feces is the main contributor of SARS-CoV-2 in wastewater. Whereas pathogen targets measured longitudinally in wastewater across a relatively stable, large population center generally reflect accurate trends over time, comparing values between different wastewater treatment plants or in dynamic populations is challenging. When normalization is performed, the target of interest (e.g., SARS-CoV-2 RNA) is benchmarked as a ratio against a second analyte in those samples. Numerous biomarkers have been proposed as normalization benchmarks (Table 1) and have been hypothesized to improve correlations between clinical cases and wastewater-measured target analyte at sampling sites where significant population flux might exist or between sites to account for differences in population sizes and water inputs. In this review, we mention only those most promising/commonly used in the context of WBS for COVID-19.

TABLE 1.

Common biomarkers explored in wastewater-based surveillance studies

Classification	Agent	Characteristics/theory	Primary means of sewage entry	Advantages	Disadvantages
Chemical analytes	Creatinine (132)	Breakdown product of energy-producing processes in muscle	Urine	Widely used in clinical medicine	Unstable and breaks down quickly
	Ammonium (133)	Breakdown product of metabolic activity	Stool	Widely performed as part of routine monitoring panels at wastewater treatment plants	Non-specific to human activity
	5-Hydroxy-indoleacetic acid (134, 135)	Breakdown product of serotonin	Urine	Stable in the environment	Very low levels, variations based on the health of the individual/population
	Caffeine (134, 136, 137)	Stimulant predominately used in food and beverages	Urine and beverage disposal	Stable in the environment, widely used	Variation in individual/population usage
	Paraxanthine (134, 136)	Breakdown product of caffeine	Urine	Stable in the environment, widely used	Variation in individual/population usage
	Coprostanol (138, 139)	Bacterial metabolite of human cholesterol	Stool	Highly specific for human fecal matter	Minimal exploration to date
Viral nucleic acid analytes	Pepper mild mottle virus (131, 140–143)	ssRNA virus (non-enveloped) analogous to many WBS targets	Stool	Abundant in common foods derived from peppers and tomatoes	Variations over time and in individual/ population usage
	crAssphage (144, 145)	Widely distributed bacteriophage of Bacteroidetes spp. present in the human gastrointestinal tract	Stool	Globally distributed	Variation in individual/population levels
Bacterial DNA analytes	Bacteroides HF 183 (146, 147)	Bacterial species with specific carriage in the human gastrointestinal tract	Stool	Globally distributed	Variation in individual/population levels
	Lachnospiraceae (146, 147) (Lachno3)	Bacterial family in high numbers in the human gastrointestinal tract	Stool
Human DNA analytes	18S rDNA (148)	Targeting specific alleles in 18S rRNA can allow for human contributions to be assessed	Stool	Specific to human populations in sewersheds	Considerably lower abundance relative to fecal bacterial/viral biomarkers
	Mitochondrial NADH dehydrogenase (149, 150)	Highly specific to human populations	Stool

Characteristics of a good biomarker include:

1. Easy to accurately measure. Candidate biomarkers should be easy to measure using similar protocols as target analytes or easily adapted using existing workflows. Ideally, there should not be factors that interfere with the target. In general, the more abundant a signal, the easier it is to measure. For example, nucleic-acid-derived biomarkers for normalization generally demonstrate the pattern that viral biomarkers (i.e., crAssphage is more abundant than pepper mild mottle virus, PMMoV) are more abundant than those from bacteria (HF183) and more abundant than direct human biomarkers (i.e., mt DNA or 18s RNA) (151–153).

2. Derives exclusively from human populations residing in the sewershed. Before the COVID-19 pandemic, significant work was already underway to model populations through analytes measured in wastewater. Several chemical biomarkers were identified, which increase in wastewater in proportion to the number of residents in the sewershed catchment (137, 154). Some of these markers are routinely recorded at wastewater treatment plants as water quality indicators [e.g., biological oxygen demand (130) and ammonium (133)], whereas others are investigational tools being explored specifically for WBS [e.g., mt DNA (149, 150)]. For example, caffeine measured in community wastewater correlates the best with population size, followed by other biomarkers including paraxanthine (PARA), PMMoV, and hydroxy-indoleacetic acid (HIAA). In contrast, creatinine has minimal correlation (134, 155). These normalization biomarkers are expected to control for dynamic changes in the population being studied (as might occur in tourist destinations or communities subject to an influx of weekday commuters or temporary workers) (114, 156).

In addition, there should be an understanding of how the biomarker enters the sewershed (i.e., mechanism of excretion, and how this correlates with the analyte of interest) and how it may vary diurnally with circadian rhythms. The biomarker should have the same dilutional characteristics and parallel linearity as the target pathogen, e.g., graywater and industrial input into the wastewater treatment plant or monitored site.

1. Stable in the environment, such that the signal is not rapidly degraded during transit. Normalization markers should not decay at a significantly different rate than the target analyte. For example, first-order decay has been described for SARS-CoV-2 RNA entering the sewershed, with a half-life of 0.99 day, similar to many biological normalization factors (157). In contrast, creatinine is unstable and highly variable in excretion from person to person and across populations (134, 155).

2. Low variance in daily excretion per capita. Ideal candidates are released into wastewater at a relatively stable rate across a population such that little day-to-day variance occurs. Whereas PMMoV is unaffected by seasonal variation (131, 140–143), considerable variations in extraction rates have been documented, and the amount detected is heavily influenced by the concentration protocol (see below) (158). Several biomarkers that depend on dietary patterns are subject to considerable variation and differences between populations [i.e., caffeine, PARA and PMMoV (159)].

3. Should not vary with seasons, weather, or geography. Markers should be relatively stable in the contributing population and have stable levels year-round (130). crAssphage and PMMoV are good examples (131). Among chemical biomarkers, PARA had the highest precision, lowest variation, and least temporal variations in populations, followed by CAF, PMMoV, and 5-HIAA (134).

4. Enters the wastewater in the same manner as the pathogen target. SARS-CoV-2 and several WBS targets (e.g., antibiotic-resistant organisms and viral causes of gastroenteritis) enter wastewater predominately through fecal shedding. It may be advantageous to have normalization biomarkers that are similarly associated with fecal excretion, and analytes that partition into solid factions include PMMoV, crAssphage, and HF183 (119, 141, 143, 160–162). In contrast, caffeine and PARA are highly soluble in water and exhibit low hydrophobicity. As a result, they adhere less well to solids, making their assessment in soluble fractions more relevant (134, 136).

Most studies comparing markers of fecal bioburden have been performed on wastewater samples collected from larger catchments, such as municipal wastewater treatment plants. Studies comparing the relative performance of different fecal biomarkers and their impact on the measured correlation between WBS data and clinical data in smaller catchments, such as neighborhoods and individual buildings, have not been pursued to the same degree (66). At present, no candidate meets all of the desired characteristics of an ideal biomarker, and consequently, a gold standard does not currently exist.

Most importantly, in general populations, SARS-CoV-2 RNA measured in wastewater correlates exceptionally well with clinical case data even in the absence of normalization (131). In fact, normalization with these candidate biomarkers does not significantly improve correlations, and in many instances, it actually reduces correlations with clinical case occurrence when assessed at large wastewater treatment plants (131, 158, 163–165). There is no unity even in studies that identify a specific normalization target as performing superior to others with respect to clinical case occurrence. Depending on the study, mtDNA (130), creatinine (130), crAssphage (103), PMMov (62), coprostanol (151), and HF183 (151) have all been described as producing the strongest correlations between WBS and clinically diagnosed COVID-19 cases when comparing different normalization markers. Furthermore, in studies involving multiple sites where these comparisons are assessed, considerable variations in the effects of normalization have been observed between sites (131, 158, 163). Biomarkers that improved correlations at some sites did not at others, suggesting that other factors (e.g., chemical or biological characteristics of the wastewater among different sampling sites) affect some biomarkers differently than the pathogen targets. This is still a knowledge gap with no consensus as to the optimal strategy for normalization.

SAMPLE CONCENTRATION AND NUCLEIC ACID EXTRACTION

Principles

Influence of “target” biology on assay concentration performance

Concentrating viruses in wastewater samples is a key first step to improve detection and quantitation. Viruses come in a variety of shapes, sizes, and electrochemical charges, all of which impact the methods used to obtain their nucleic acids for molecular and genomics applications. Additionally, the differences between enveloped and non-enveloped viruses and the different functional groups on the surface of these viruses impact their partitioning in the wastewater (i.e., solid or liquid fraction) (112), which ultimately directs which fraction to target for nucleic acid extraction to maximize recovery. SARS-CoV-2 is an enveloped RNA virus but other targets of WBS may be different. For example, the PMMoV was initially embraced as a potentially ideal normalization target owing to its prevalence in sewage due to the widespread occurrence of pepper products in human diets (142). While SARS-CoV-2 and PMMoV are RNA viruses excreted in feces, the latter is a non-enveloped virus, and the two are structurally different. Therefore, different strategies to obtain viral nucleic acids can generate variable efficiency with these targets in the same sample. Accordingly, normalization with PMMoV has been met with mixed success (134, 158, 166) and is not applied universally.

Non-enveloped viruses lack an exterior lipid bilayer and tend to be more resistant to harsh treatment conditions (e.g., heat, pH, desiccation, disinfection, and detergents) that are typically used in nucleic acid extraction protocols. Accordingly, strategies for non-enveloped viruses in wastewater samples may be optimized differently than those for enveloped viruses. Examples of non-enveloped viruses detected in environmental samples include norovirus in river water (167) and wastewater (168), and rhinovirus (169), adenovirus (170), and enterovirus in wastewater (171). The hydrophobic exterior and functional groups of enveloped viruses give rise to their higher association with wastewater solids rather than suspended in the liquid fraction (111–113). Structural differences between viruses also influence which nucleic acid extraction protocols are effective for WBS programs. Respiratory viruses SARS-CoV-2, influenza A and B, and respiratory syncytial virus (RSV) are all enveloped RNA viruses. Accordingly, WBS teams have incorporated reverse transcription-quantitative PCR (RT-qPCR) assays for influenza A and B (172) and RSV (173) into existing COVID-19 surveillance programs by using the same protocols to extract viral genomic RNA from wastewater (174). Practicalities aside, how influenza A and B and RSV signals from wastewater correlate with clinical disease burden remains to be fully elucidated (see below). Differences in viral structure do not necessarily render extraction protocols unusable for non-intended target viruses. It is possible and even likely that a common nucleic acid extraction protocol may work for a range of enveloped and non-enveloped viruses in the same sample. Clearly, this would be attractive for methodological efficiency and cost savings. Investigators should be cautious in this regard.

Wastewater concentration and nucleic acid extraction

Developing and optimizing methods to extract enveloped SARS-CoV-2 in wastewater were important objectives early in the pandemic when WBS protocols and programs were being established (113, 175–179). Virus recovery methods often apply a concentration step before extracting the nucleic acid. Concentration reduces the volume of the sample while retaining viral particles and genomic fragments. Viruses can be concentrated using methods that include ultrafiltration, flocculation, precipitation, centrifugation, or magnetic beads (180–185). Some key strategies deployed in WBS during the COVID-19 pandemic are described below (Fig. 1B).

Ultrafiltration

Ultrafiltration has gained popularity as a method of concentrating specific targets in wastewater samples. Ultrafiltration relies on the principle of size exclusion as water samples are passed through capillaries, hollow fibers, or flat sheets using tangential flow. Filters have nominal molecular weight cut-offs of 30–100 kDa and pore sizes in the range of 0.5–2.0 nm. Because of the pore size, water and low molecular weight substances can pass through the fibers into the filtrate, whereas larger substances, such as viruses and other microorganisms, are retained. Ultrafiltration can be dead-end (the sample passes through a single time) or tangential flow (multiple circulations of the sample through the filter). Typically, filters are then backwashed to remove any microorganisms that are retained, and this wash is combined with the retentate of the final sample for analysis (186, 187).

Anisotropic membranes are used in various ultrafiltration products, such as the Centricon and Amicon ultracentrifugal filter units (Millipore). These are available with various nominal molecular weight limits (NMWL) depending on the molecular weight of the particle or the molecule of interest. It is generally recommended to select a membrane with an NMWL three times smaller than that of the viral target of interest (187). Retention depends on the molecular weight as well as the three-dimensional shape of the viral particles. Devices with an NMWL of 30 kDa have been used to concentrate non-enveloped enteric viruses from wastewater (181). Enveloped viruses are generally larger than non-enveloped viruses, such that devices with NMWLs of 30 kDa can isolate both types of viruses (188). Once the samples have been concentrated, the nucleic acid can be extracted using a preferred method. The major disadvantages to the use of ultrafilters are slow filtration rates, difficulty in field deployment, and tendency to clog under conditions when particulates exceed 1.6 g L⁻¹ (186).

PEG precipitation

Polyethylene glycol (PEG) precipitation was widely and successfully used at the onset of the COVID-19 pandemic to concentrate SARS-CoV-2 from wastewater (189–191). This method was known to be effective as it had been tested with other viruses (112, 192, 193). However, this method requires long incubation/centrifugation times (>2 h) making it unattractive for high-volume WBS programs with fast turnaround time requirements.

Electronegative membrane filtration

Most viral particles are extremely small (0.03–0.10 µm in diameter) and have an isoelectric charge between 3.5 and 7, thus they are negatively charged in pH > 7 solutions (113, 194). Accordingly, viruses and genetic fragments can be concentrated using positively or negatively charged membranes to which viral particles become adsorbed and eluted in a concentrated manner by adjusting the electrical charge of either the wastewater samples or the filters via changes to pH or using chemical attractants or repellants (195, 196). For electronegative filtration, virions in wastewater adsorb directly to negatively charged membranes when the pH of the sample is adjusted below the isoelectric point of the viral particles. Alternatively, a salt solution (e.g., MgCl₂) can be used to donate cations that bridge the negative surface of the membrane and the negative charges on the virions. For electropositive membrane filtration (also known as direct filtration), the virions adsorb directly to the charged membrane (197). For either technique, viral particles must then be eluted from the surface of the membrane for the nucleic acid to be extracted. This can be done using protein-based solutions with a high pH (i.e., glycine buffer pH 9.5) to facilitate the elution of virions from the filter. Eluents can then be further concentrated for virions or viral RNA fragments by additional methods like ultrafiltration or polyethylene glycol precipitation (197). Nucleic acid can then be extracted from the eluent, the filter directly, or various concentrates (from ultrafiltration or polyethylene glycol) (183, 197). Charged membranes can also be used in passive sampling devices to improve viral recovery (124).

Flocculation

Flocculation methods such as skim milk and aluminum-driven flocculation are relatively inexpensive and straightforward methods for virus concentration and recovery. Skim milk flocculation was developed to recover viruses from seawater (198) and has been used to determine viral titers in various food products (199). A skimmed milk solution is pre-flocculated by acidification and added to an aqueous sample. The resulting mixture is incubated for an extended time to allow viral particles to adhere to the flocs, which are then collected via centrifugation and used to extract nucleic acid. This method can generate consistent and reproducible results for recovering viruses (198, 200, 201). In aluminum-driven flocculation, the sample is acidified and an AlCl₃ solution is added to form a precipitate, which is collected by centrifugation for further processing and ultimately nucleic acid extraction (191, 202–204).

Direct extraction

In a complex matrix such as wastewater, using a concentration method before extracting nucleic acid assumes viral particles are intact and does not effectively capture free-floating nucleic acids. At the onset of the COVID-19 pandemic, researchers at UC Berkeley developed a direct extraction method they termed 4S (i.e., sewage, salt, silica, and SARS-CoV-2) (205). This method recovers nucleic acids without any size, mass, or charge bias and co-captures nucleic acids from intact and lysed viral particles in wastewater. This method was used successfully for WBS during the COVID-19 pandemic (66, 94, 105, 165, 205–208). 4S can be scaled up to mid-throughput, is cost-effective, and minimizes workflow steps by eliminating the need for a separate downstream nucleic acid extraction step with associated reagents and/or commercial kits. The wastewater sample is first lysed with a high concentration of sodium chloride and stabilized with Tris EDTA (TE) buffer (205, 209). By lysing before removing solids, viral particles associated with the solids are also lysed thus increasing the total recovery of nucleic acid. Heat-based pasteurization inactivates any infectious particles in the wastewater and helps lyse other non-enveloped viruses, such as PMMoV, thus increasing their recovery. After a coarse filtration step and ethanol precipitation, nucleic acids are captured using commercial silica-based columns or a silicon dioxide slurry.

Magnetic bead-based methods

More recently, magnetic bead-based approaches have been developed and applied to concentrate viruses in wastewater. This has been prompted by more conventional virus concentration methods often being expensive, resource-intensive, and time-consuming. Commercial magnetic bead particles such as Dynabeads virus enrichment beads (Invitrogen), which are strong anion-exchange magnetic beads that bind negatively charged molecules (i.e., viruses) or magnetic hydrogel particles (Ceres Nanosciences Inc), among others, have been used to concentrate SARS-CoV-2 from wastewater (184, 185, 210). Research teams continue to develop other magnetic-bead-based approaches to target viruses more specifically, such as magnetic beads coated with porcine gastric mucin; these glycans can act as virus receptors, possibly promoting higher binding affinity to the magnetic particles (211).

Internal quality control checks and balances

Nucleic acid amplification technology (NAAT) has been widely and successfully applied to detect and quantify SARS-CoV-2 RNA levels in sewage samples. However, validating methods in different settings for WBS of SARS-CoV-2 requires investigators to understand and control the interfering factors that can impact assay performance. Potential interfering factors in the analytical procedure can be categorized into two types: intractable and submissive. Intractable factors in the WBS of SARS-CoV-2 are the effects of inhibitory, diverse, and pervasive substances in wastewater samples, and the ways they impact (e.g., inhibit) the recovery of targeted analytes (e.g., RNA from SARS-CoV-2) and their analysis (e.g., RT-PCR assays). Submissive factors, on the other hand, are manageable through good laboratory practice, including the variables around sample selection, collection, storage, and treatment (as described above). This section focuses on the quality control of intractable factors in the WBS of SARS-CoV-2.

Certain chemicals may inhibit or reduce the sensitivity of the RT-PCR reaction by promoting inefficiency during amplification (e.g., the erroneous activity of polymerase enzymes during reactions), resulting in false-negative results and impacting reproducibility (212–214). Wastewater samples contain large amounts of those chemicals and substances, including humic and phytic acids, bile salts, polyphenols, detergents, metal ions, bacterial cells, and other biological debris in relatively high concentrations (213, 215–218). To assess the impact of these factors on reducing PCR efficiency, an internal control should be assessed in each sample. Commonly used surrogates are generally viruses that structurally and morphologically resemble SARS-CoV-2, such as human coronaviruses 229E and OC43 (219, 220), bovine coronavirus (66, 91, 96, 119, 158, 205, 208, 221–230), murine hepatitis virus (183, (221, 231), or bacteriophages (231). No single internal process control performs better than others (222). As consensus is lacking, selecting a positive process control relies on available resources, laboratory experience, technical expertise, and, in some cases, supply chains and availability of the selected control (232–234). Exogenous internal controls can then be used to understand the overall performance of the RT-PCR assay by calculating the recovery rate of targeted analytes in the presence of a large quantity of inhibitory substances in the wastewater sample matrix (235). Introducing a constant amount of exogenous control material to samples can allow the quantitative and qualitative qPCR analysis to be normalized, regardless of the initial sample type. Inhibitory substances in the sample can suppress PCR amplification partially or completely, resulting in incorrect quantification or false-negative results (236). Several spiked-in exogenous quality controls have been described for identifying PCR inhibitors in SARS-CoV-2 wastewater studies, such as the Sketa22 (9, 237), salmon DNA (219), transmissible gastroenteritis coronavirus genome (238), coliphage MS2 (ATCC 15597-B1) (190, 239), hepatitis G armored RNA (221), VetMAX Xeno internal positive control (228); and haloalkaliphilic archaeon’s gyr a gene (225).

Calculating the thermal cycle threshold (Ct) of a real-time PCR amplification profile is a simple approach to identify the presence of PCR inhibitors in wastewater samples. One common definition for inhibited PCR reaction is a different amplification profile with 2 Cts greater than the expected Ct values of amplification (240). Liu and Saint explored the mathematical relationship between the initial amount of a target (RNA or DNA), the amplification efficiency of the real-time PCR reaction, and the Ct value and illustrated that the Ct values of samples with a known amount of target (indicator) can be used to determine the PCR efficiency of a reaction, i.e., detect the presence of inhibition (241). Kontanis and Reed demonstrated that analyzing the real-time PCR profile is a sensitive way to find contamination of tannic acid, which decreases PCR efficiency remarkably (242). The presence of 0.4 ng of tannic acid in a real-time PCR reaction resulted in an increase of 4 Cts compared to the PCR amplification profile without tannic acid controls. Based on a logistic regression analysis of their study, Liu and Saint also recommended to use changes in the Ct value to estimate the degree of inhibition on PCR efficiency (241).

Reducing or removing PCR inhibitors can be accomplished through:

• optimizing sample concentration and nucleic acid extraction approaches (see Factors to understand the population from which samples derive and to allow for normalization) that minimize the co-concentration or removal of PCR inhibitors (243, 244),

• using inhibitor-tolerant reverse transcriptase and DNA polymerase that could relieve or suppress the effect of PCR inhibitors when performing PCR-related approaches (214, 245), and

• using heat and chemical treatment before the PCR (217).

Future work should focus on optimizing the best process control material (artificial or natural) and developing national or international standard protocols or best practice guidelines pertaining to analytical procedures for the WBS of SARS-CoV-2 to promote process harmonization and data comparability.

MOLECULAR TARGET DETECTION

SARS-CoV-2 genomic targets

Recovery of replication-competent viruses from wastewater is particularly challenging owing to rapid degradation over time, inhibitory chemicals in the wastewater matrix preventing cell culture work, and, in particular, for SARS-CoV-2, the requirement for the work to be accomplished in a BSL-3 laboratory (246). In fact, viable SARS-CoV-2 virus has yet to be recovered from wastewater. Accordingly, WBS for viruses and, in particular, SARS-CoV-2 has been accomplished using molecular and genomic tools.

SARS-CoV-2 has a large RNA genome (~30 kilobases, kb) that encodes about 29 proteins (247, 248). Several genomics regions involved in different functions during the assembly of progeny virions have been used as targets to detect SARS-CoV-2. Those include sub-genomic RNAs that encode for conserved structural proteins [i.e., envelope protein (E) and nucleocapsid protein (N)], genomic regions that encode for non-structural proteins from two ORFs (i.e., ORF1a and ORF1b), and genomics regions involved in replication and transcription [i.e., RNA-dependent RNA polymerase (RdRP)] (249).

Multiple nucleic acid amplification tests have been developed and used to detect SARS-CoV-2 RNA. The US Centers for Disease Control and Prevention designed three assays targeting the N gene (i.e., N1, N2, and N3 assays) for the SARS-CoV-2 Real-Time RT-PCR Diagnostic Panel (250, 251), of which the first two are the most widely used. The N3 assay is no longer recommended as part of the diagnostic panel as it was designed for the universal detection of SARS-like coronaviruses (252). Corman and collaborators (253) designed additional assays to identify SARS-CoV-2 RNA targeting the RdRp gene, E gene, and N gene. They found that the RdRp and E assays were more sensitive than the one targeting the N gene (253), although investigators have reported varying sensitivities of each target. See Table 2 for a detailed description of the most commonly employed assays used to detect and quantify SARS-CoV-2 RNA.

TABLE 2.

Assays used to detect SARS-CoV-2 RNA

Assay	Gene target	Primer/probe name	Oligonucleotide sequence (5′–3′)	Reference
US-CDC-N1	N gene	2019-nCoV_N1-F	GACCCCAAAATCAGCGAAAT	(250, 251)
		2019-nCoV_N1-P	ACCCCGCATTACGTTTGGTGGACC
		2019-nCoV_N1-R	TCTGGTTACTGCCAGTTGAATCTG
US-CDC-N2		2019-nCoV_N2-F	TTACAAACATTGGCCGCAAA
		2019-nCoV_N2-P	ACAATTTGCCCCCAGCGCTTCAG
		2019-nCoV_N2-R	GCGCGACATTCCGAAGAA
US-CDC-N3		2019-nCoV_N3-F	GGGAGCCTTGAATACACCAAAA
		2019-nCoV_N3-P	AYCACATTGGCACCCGCAATCCTG
		2019-nCoV_N3-R	TGTAGCACGATTGCAGCATTG
Charité-E	E gene	E_Sarbeco_F1 Forward Primer	ACAGGTACGTTAATAGTTAATAGCGT	(253)
		E_Sarbeco_P1 Probe	ACACTAGCCATCCTTACTGCGCTTCG
		E_Sarbeco_R2 Reverse Primer	ATATTGCAGCAGTACGCACACA
Charité-N	N gene	N_Sarbeco_F	CACATTGGCACCCGCAATC
		N_Sarbeco_P	ACTTCCTCAAGGAACAACATTGCCA
		N_Sarbeco_R	GAGGAACGAGAAGAGGCTTG
Charité- RdRp	RdRp gene	RdRp_SARSr-F	GTGARATGGTCATGTGTGGCGG
		RdRp_SARSr-P2	CAGGTGGAACCTCATCAGGAGATGC
		RdRp_SARSr-R	CARATGTTAAASACACTATTAGCATA
China-CDC-N	N gene	Forward primer (F)	GGGGAACTTCTCCTGCTAGAAT	(254)
		Fluorescent probe (P)	TTGCTGCTGCTTGACAGATT-TAMRA
		Reverse primer (R)	CAGACATTTTGCTCTCAAGCTG
China-CDC-ORF1ab	ORF1a and ORF1b	Forward primer (F)	CCCTGTGGGTTTTACACTTAA
		Fluorescent probe (P)	CCGTCTGCGGTATGTGGAAAGGTTATGG
		Reverse primer (R)	ACGATTGTGCATCAGCTGA
US-CDC-SC2	N gene	SC2 For	CTGCAGATTTGGATGATTTCTCC	(255)
		SC2 Probe	ATTGCAACAATCCATGAGCAGTGCTGACTC
		SC2 Rev	CCTTGTGTGGTCTGCATGAGTTTAG

NAAT drives the venue of WBS of SARS-CoV-2

Detecting traces of viral pathogens in wastewater and using the results as evidence of infection transmission among humans were pursued long before PCR was discovered (see below for polio example) (256, 257). Unlike polioviruses that primarily shed in the feces and contain a high number of viral particles, SARS-CoV-2 sheds at a relatively low level in feces and is detectable mainly as non-infectious RNA fragments in wastewater (59, 258–260). Thus, adequate sampling and sensitive, specific, and quantitative detection are critical when using NAAT in WBS to monitor for COVID-19. NAAT-based methods, including the qualitative and quantitative detection of target pathogens, are considered superior to other conventional technologies in terms of rapid testing, fast turnaround and reporting times, ability to confirm clinical results and trends, and providing meaningful epidemiological data at the individual and community level. Commonly employed NAAT for COVID-19 monitoring includes RT-qPCR, RT-digital PCR (RT-dPCR), and RT-loop-mediated isothermal amplification (RT-LAMP). This section describes the application of these methods in the WBS of SARS-CoV-2 during the COVID-19 pandemic globally, including their advantages and limitations (Fig. 4). There is no “perfect” technology; each laboratory should adapt a practical platform to meet expectations according to its infrastructure, budget, and expertise. As experience accumulates and knowledge increases, improving NAAT and its application on WBS of SARS-CoV-2 and other emerging pathogens will be warranted.

Fig 4.

Molecular strategies to identify and quantify target analytes in wastewater. (A) RT-qPCR, (B) RT-dPCR, and (C) RT-LAMP.

Reverse transcription-quantitative PCR

PCR was a revolutionary advancement in molecular biology (261). Conventional PCR, also known as endpoint PCR, requires post-PCR confirmation of the amplified target (i.e., gel electrophoresis and additional labeling steps) that are labor- and time-intensive and prone to cross-contamination when handling amplified products (262). In comparison, fluorescence-labeled real-time quantitative PCR (qPCR) and/or real-time RT-qPCR assays have become the principal assays used in diagnostic and research laboratories due to their higher sensitivity, enhanced specificity, lower risk of cross-contamination as closed-tube reaction, real-time integrated detection and quantification, versatile reaction platforms, ability to detect multiple targets, and rapid turn-around-time using advanced instruments.

Since the first RT-qPCR used to detect SARS-CoV-2 in untreated wastewater in Australia provided a proof-of-concept for the WBS of COVID-19 in communities (9), an enormous number of papers on detecting SARS-CoV-2 in wastewater by RT-qPCR-based assays have been published. The RT-qPCR assay has been widely used to detect and quantify SARS-CoV-2 RNA levels in sewage samples from settings such as municipal wastewater treatment plants, neighborhoods, university dormitories, hospitals, correctional facilities, shelters, and long-term care facilities. In all these settings, the assays indicated the burden of COVID-19 infection and provided an early warning system for COVID-19 outbreaks, enabling interventions to prevent transmission. Most laboratories in Canada’s WBS network for SARS-CoV-2 implemented RT-qPCR assays to monitor COVID-19 trends. However, different RT-qPCR assays using various platforms, reaction chemistries, standard materials for calibration, reference materials, and QA/QC controls have been developed leading to a large variation in quantitative outcomes (263). Given the heterogeneity of samples and the lack of traceability to a standardized reference system, it is nearly impossible to reliably compare outcomes of RT-qPCR assays on SARS-CoV-2 levels between laboratories (263). Reference standards for calibrating SARS-CoV-2 RT-qPCR assays are urgently needed, which is a critical step in the quality control and improvement of quantitative RT-qPCR assays for the WBS of SARS-CoV-2. Standardizing the assay will allow investigators to perform inter-laboratory comparisons nationally and internationally, facilitating insights into possible factors that interfere with RT-qPCR reactions. For example, the Ontario Clean Water Agency has led an inter-laboratory trial with the Ontario Ministry of the Environment/Conservation and Parks and the Public Health Agency of Canada to compare methods on this aspect since 2021 (263). The international water community has also made great inputs toward developing and standardizing methods for measuring SARS-CoV-2 in wastewater.

Digital PCR

Digital PCR (dPCR) can be used to detect and differentiate single nucleotide polymorphisms in template DNA (264), providing a way to enumerate target loci. This technology distributes the target locus across a large number of physical partitions on a reaction chip by limiting dilution so that a single DNA molecule is present in some but not all the partitions (265, 266). Therefore, the absolute quantification of targets using dPCR platforms depends on limiting dilution and the most probable number of targets premised on Poisson statistics (267, 268). Due to its endpoint measurement for each partition, an external calibration curve is not required, improving this assay’s reproducibility and accuracy to quantify target molecules, potentially a significant advantage of dPCR over qPCR.

Early incarnations of dPCR used limiting dilution so that each PCR reaction contained a single DNA molecule. Recent dPCR platforms use nanofluidic chambers (arrays) or nanodroplet emulsions to partition single molecules more effectively and precisely (269). Nanofluidic arrays divide the target DNA into numerous reaction chambers with nanoliter volumes where PCR takes place. The number of chambers with positive reactions (i.e., end-point PCR results) is counted by fluorescence imaging (270–273). Another means of partitioning involves emulsifying target DNA and PCR master mixes into thousands of nanodroplets that serve as PCR chambers and are subsequently enumerated for the number of droplets containing positive PCR reactions (268, 274, 275). Both approaches allow target DNA in a sample to be quantified accurately.

The first study comparing the sensitivity of SARS-CoV-2 detection in wastewater between RT-qPCR and RT-dPCR was by Boogaerts et al. (276). Following this, other studies reported comparable or slightly better sensitivity and detection by RT-dPCR compared to RT-qPCR for the WBS of SARS-CoV-2 (62, 221, 276–282), superior reproducibility of digital PCR to qPCR (283), and less inhibition using RT droplet digital PCR (Bio-Rad) resulting in fewer false-negative results compared to RT-qPCR (279, 284). However, digital PCR remains a relatively new technology in the context of WBS, such that potential sources of bias need to be identified and mitigated. It has been observed that the size of fragmented RNA and its derived DNA can significantly impact digital PCR results of SARS-CoV-2 in wastewater (272, 283). Similarly, detection can be affected by sustained template exposure to high temperatures and variations in partition volumes. When the sources of variation in a dPCR platform are better understood, standard guidelines and best practices can be developed for its application in WBS. This will be valuable as this field expands to detect other viruses and pathogens.

Loop-mediated isothermal amplification

Although RT-qPCR and RT-dPCR are the main technology platforms used to detect and quantitate SARS-CoV-2 in wastewater, RT-LAMP is a related application offering the advantages of speedy reaction, simple procedure in a single tube, without requiring multi-thermal cycle amplification, and easy-to-read end-point results.

In general, LAMP has a high tolerance for biological material, such that extensive extraction of nucleic acid is often not necessary. No large or exotic instrument is required since amplification occurs in a single and unchanged temperature condition (normally 60°C–65°C), making LAMP assays attractive for point-of-care diagnostics and for laboratories with limited resources or experience using NAAT. LAMP outcome measurements can be tailored in many ways, such as changes in fluorescent signals using intercalating dyes, DNA probes with gold nanoparticles, turbidity changes caused by magnesium pyrophosphate precipitation (285), or gel electrophoresis followed by UV light detection (286, 287). The most common end-point reading is a visible color change in a colorimetric master mix containing a pH indicator for rapid detection. Alternatively, the precipitates in the reaction mix can be formed by adding magnesium-containing agents, and turbidity can easily be analyzed using a small reader. Fluorescent dye can also improve the measurement of the end results of an RT-LAMP reaction. Mautner and colleagues reported a RT-LAMP assay that was 10 times cheaper and 122 times faster than RT-qPCR for the clinical diagnostic of SARS-CoV-2 when working with respiratory swab samples from patients (288).

RT-LAMP detection for COVID-19 has mainly been used in clinical diagnosis, with only a few reports of detecting SARS-CoV-2 in wastewater. Rapid detection, within 3 hours from wastewater sample collection to result reporting, is an attractive feature of RT-LAMP applications for WBS with very good target specificity (281, 282, 289, 290). However, the sensitivity of the RT-LAMP assay is lower than that of RT-qPCR (290, 291). Ongerth and Danielson reported that RT-LAMP could detect SARS-CoV-2 in wastewater directly without the need for a dedicated nucleic acid extraction step, implying that the RT-LAMP reaction was not unduly inhibited by inhibitory substances in the wastewater matrix (292). In contrast, Donia and colleagues showed that nucleic acid extraction was a necessary preliminary step for successful RT-LAMP detection by comparing results with and without extraction for seven wastewater samples (290). Since RT-LAMP is still at an early stage of development in the context of WBS, further evidence needs to be collected and evaluated to assess its pros and cons. A major concern is that the RT-LAMP assay cannot quantify the levels of SARS-CoV-2 in wastewater precisely, which is very important for dynamic monitoring in some situations (i.e., municipal monitoring) but less so in others (i.e., where presence/absence is more important, such as individual high-risk facilities). The advantages mentioned above still make RT-LAMP an attractive option for adding value to the WBS of SARS-CoV-2 in some situations, including resource-poor settings.

Tracking variants of SARS-CoV-2 in wastewater

WBS has been effective in monitoring the diversity of SARS-CoV-2 variants, including VOCs, within a population. The ability of PCR-based and genomic sequencing methods to distinguish diagnostic mutations within RNA isolated from SARS-CoV-2 in wastewater enables these applications.

Allele-specific PCR

In principle, PCR-based assays for VOCs are identical to the technology used to monitor quantitatively for SARS-CoV-2 in wastewater and clinical samples. RT-qPCR assays consist of two primers used to amplify the target region of interest as well as an additional hybridization probe targeting a region in the middle of the amplified fragment. Provided that the primers and probes can hybridize to regions of the genome that have not experienced mutations, in principle, all SARS-CoV-2 genomes should be amenable to these assays. WBS programs routinely mitigate this risk by simultaneously tracking total levels of SARS-CoV-2 by implementing multiple universal assays in parallel (i.e., simultaneous monitoring for N1 and N2). These same principles allow new primers and/or probes to be designed that specifically hybridize with sequences that detect VOCs by targeting regions with mutations. This approach has been useful for monitoring the progression of successive waves of COVID-19 by deploying assays for specific VOCs (293–298). In some instances, high-resolution tracking with different VOC assays revealed the displacement of one VOC by another. This was illustrated in a province-wide study in Alberta, Canada, whereby VOC assays determined an increasing proportion of Omicron (BA.1) compared to Delta in large and small communities over time as a function of distance to airports with international arrivals (206). In this example, like in other parts of the world, the increase of Omicron coincided with a dramatic increase in the total level of SARS-CoV-2 quantified in the same wastewater samples using “universal” RT-qPCR assays.

The advantages of allele-specific PCR for monitoring VOCs relative to alternatives (e.g., genomics sequencing of wastewater samples; see below) mainly relate to turnaround time, ease of use, and affordability. Similar to real-time monitoring programs that use universal PCR assays to track overall levels of SARS-CoV-2, the same RT-qPCR (or dPCR) technology and infrastructure can enable VOC tracking, provided VOC-specific primers and probes are available. Developing specific and differentiating assays can be the bottleneck that delays WBS for VOCs, as it can take weeks to develop, validate, and optimize a PCR assay. On the other hand, the rapid posting of genomics information about SARS-CoV-2 mutations and variants provides the necessary information for assay development and testing. Furthermore, once an assay is validated by one group and posted publicly, it can be used by other groups around the world, potentially before a new variant emerges in the location that is onboarding the assay. However, this also underscores the inherent limitation of allele-specific PCR (and indeed any PCR)-based approaches, i.e., that they are based on prior knowledge of the target sequence. These assays may have been more useful at earlier stages of the pandemic when circulating variants were far fewer.

Amplicon tiling genomic sequencing

PCR strategies using two primers and one probe that are each approximately 20 nucleotides long translate into diagnostic specificity that relies on 60 of the 30,000 bases in the SARS-CoV-2 genome (0.2%). Continually measuring SARS-CoV-2 in wastewater using these assays therefore depends on mutations in the genome not occurring in these target regions. While the chances of this happening are modest (~0.2%), the high mutation rate in RNA viruses (299, 300) and the well-documented evolution of SARS-CoV-2 (301, 302) make it inevitable that emerging variants will evade detection using established quantification PCR workflows (301). Genomics sequencing of SARS-CoV-2 genomes in wastewater is a complementary strategy that theoretically requires no a priori knowledge of the genomic diversity of the viruses in the sample. Thus, whereas allele-specific PCR represents a “reactionary” posture and strategy that requires assay development and onboarding, genomics sequencing is inherently “proactive” in that all variants present can be included in the resulting libraries. In principle, rather than relying on the genome sequences of new variants reported in other jurisdictions, sequencing enables prospectively monitoring for the emergence of variants locally.

A key challenge that sequencing strategies must overcome is the enormous heterogeneity of biological material in wastewater. Unlike samples from COVID-19 patients that are relatively homogenous (containing only a single or a few SARS-CoV-2 variants together with host-derived human nucleic acids), wastewater samples contain a plethora of nucleic acids from humans, other eukaryotes, bacteria, and viruses. As such, in community wastewater, genomic RNA from SARS-CoV-2-infected individuals co-exists with large amounts of interfering signals. This means that the diversity of SARS-CoV-2 lineages present represents only a tiny fraction of the total nucleic acids in wastewater samples. For this reason, shotgun metatranscriptomic sequencing of RNA from wastewater without any additional enrichment has yielded very low proportions of reads mapping to SARS-CoV-2 (303, 304).

Targeting only sequences of interest for amplification has mainly been achieved using amplicon tiling or probe-based enrichment strategies (64, 303, 304). The most widely-used strategy for SARS-CoV-2 genomics sequencing from wastewater, combining efficiency and costs, is an amplicon tiling scheme (305), whereby the 30-kb SARS-CoV-2 genome is divided into smaller sections that are independently amplified to generate overlapping fragments of each section (not to quantify but to generate enough copies for routine sequencing) (305). The resulting amplicons are sequenced and mapped to the SARS-CoV-2 reference genome to measure the diversity of SARS-CoV-2. Targeted amplification circumvents other genomic material in wastewater, affording the advantage that all variants present in the wastewater sample should be represented. While this is not without biases owing to the requirement for primer-based amplification, more robust ratios of variants including known and emerging VOCs can be calculated and reported.

Detecting variants and their relative abundance depends on software packages, such as ASPICov (306), Kallisto (307), Vaquero (308), COJAC (309), or Freyja (184). Whether or not these pipelines can identify new variants accurately relies on the underlying databases being continually updated. Similarly, the primer pairs used to generate the amplicons (or “tiles”) across the SARS-CoV-2 genome must be checked to ensure emerging mutations do not impact performance.

Different groups have used amplicon tiling for the WBS of SARS-CoV-2 in wastewater in a variety of settings, including municipal wastewater (310, 311) and on university campuses (184). While the resulting genome sequences offer an excellent opportunity to compare with SARS-CoV-2 genomes derived from clinical samples, investigators also noted that community wastewater samples can reveal hundreds of variants, including some “cryptic” lineages. This led to suggestions that some variants may be derived from individuals with long-term infections or from infected animals providing an alternative incubator for the virus. Regardless of the source of cryptic lineages, these types of findings further underscore that wastewater sampling is the ultimate inclusive strategy for capturing large numbers of individuals in populations as well as the diversity of variants that infect them.

Semi-agnostic probe-based genomic sequencing of viruses in wastewater

A potential disadvantage of the semi-targeted tiling approach described above is that it only targets one virus. As described in the section Moving beyond SARS-CoV-2 to other pathogens, WBS holds great potential for tracking other viruses, including through the use of genomics sequencing applications.

To obtain useful data for a greater number of viruses in the heterogeneous wastewater mixture, hybridization probe-based technologies have been developed whereby different combinations (or “panels”) of targets can be combined and enriched in a semi-selective sequencing strategy. Indeed, attempts to sequence a low-abundance target will be overwhelmed by abundant nucleic acid sequences unless mitigating steps are taken (303, 304). Other ways to tip the scales in favor of successfully measuring low-abundance targets include looking for similar viruses, such as respiratory RNA viruses SARS-CoV-2, influenza, and RSV at the same time. In this example, nucleic acid extracts can be processed to concentrate relevant signals, e.g., using DNases (so that only RNA genomes remain) and ribosomal RNA depletion (to remove the plethora of bacterial and human rRNAs) thereby enriching viral genomic RNA. Following this, specific hybridization probes are applied for pathogens of interest before sequencing. The resulting sequencing does not rely on amplicon tiling (i.e., no associated primer bias) and delivers metagenomics information on the remaining nucleic acids in the sample similar to other environmental metagenomics applications. Applying this strategy in a longitudinal sampling of wastewater from a large urban center using Qiagen’s xHYB panel (132 viruses) revealed trends for influenza, RSV, and coronaviruses that confirmed clinical results over the same time frame (312). These respiratory viruses were only detectable using this enrichment step but not in shotgun metagenomics libraries of the same samples, which instead highlighted human astroviruses and enteroviruses, i.e., the most abundant signals in the samples. This underscores the utility of panel-based enrichment, provided that panels for pathogens of interest are commercially available or are constructed.

DATA SHARING, REPORTING, AND TRANSPARENCY

Government vs public facing

The application of WBS for SARS-CoV-2 has evolved to be remarkably diverse, making guidance on data sharing, reporting, and transparency inevitably context-specific (63, 87). Most commonly, wastewater has been monitored at centralized wastewater treatment plants, providing an integrated SARS-CoV-2 signal for the entire community served by that system.

A major hurdle in sharing WBS data has been the absence of universal standardized methods for processing samples, analyzing RNA fragments, and reporting data. Ahmed et al. reviewed the many elements of WBS that can lead to differing quantitative results (103). These technical variations reflect the rapid adoption of WBS by many investigators, many of whom were not previously engaged in public health surveillance, seeking to respond to the urgent demands on the public health system caused by the COVID-19 pandemic. The University of California Merced (https://arcg.is/1aummW) reports the adoption of the WBS approach in 70 countries, at 3,892 sites involving 284 universities (as of January 2023) (313, 314). Interlaboratory studies evaluating quantitative WBS results for SARS-CoV-2 (121, 263) demonstrated large quantitative differences (orders of magnitude) among laboratories analyzing common wastewater samples. These differences preclude direct comparison of quantitative results from different locations unless the samples were prepared and analyzed by a single laboratory, an uncommon situation outside of regional programs.

WBS data are generally reported to public health authorities for their internal use. Other than individual retrospective research publications about specific monitoring applications, real-time reporting of WBS data to the public has been limited to dashboards of WBS data from wastewater treatment plants or neighborhoods. In September 2020, the City of Ottawa, University of Ottawa, Children’s Hospital of Eastern Ontario Research Institute, and Ottawa Public Health established one of the first public-facing dashboards providing near real-time (data posted within days of sample collection) WBS data (https://613covid.ca/wastewater/). These data were progressively used to triangulate with clinical test data and evaluate public health interventions (315). Over 2021, other jurisdictions added similar dashboards. For example, Alberta’s dashboard monitored 43 communities, covering more than 83% of the province’s total population of 4.5 million [https://covid-tracker.chi-csm.ca (206)]. As of January 2023, 25 public-facing dashboards were active across Canada (https://health-infobase.canada.ca/covid-19/wastewater/#links), and 152 worldwide (https://arcg.is/1aummW). As more transmissible VOCs, such as Omicron, became prevalent in late 2021 and early 2022, WBS data emerged as the only comprehensive source of information on the prevalence of COVID-19 in communities because centralized clinical testing could no longer keep up with demand, and many cases were identified by home or workplace testing that was not centrally recorded.

Public-facing dashboards of near real-time data allow individuals to apply their own risk management measures such as vaccination, wearing masks, and avoidance of individuals beyond their inner circle. Although no focused research has been done on the public use or uptake of the dashboards, the Alberta site recorded up to 8,000 hits/day, suggesting substantial public interest in locally relevant WBS data. The US Centers for Disease Control and Prevention established a National Wastewater Surveillance System (https://www.cdc.gov/nwss/index.html) to coordinate reporting of WBS among the initial 40 US sampling sites in 2020 (https://covid.cdc.gov/covid-data-tracker/#wastewater-surveillance), up to 1,450 as of January 2023. When CDC first launched a national public-facing dashboard for WBS data in February 2022 after extensive internal consultation about how to present the data from multiple disparate sources, it experienced rapid and overwhelming critical feedback (https://www.cdc.gov/surveillance/blogs-stories/wastewater-surveillance.html) that led to a more successful re-launch in April 2022 with a clear and primary aim at public use vs satisfying the needs of researchers. Dr. Amy Kirby who led this exercise concluded that effective sharing of WBS data with the public requires serious consultation with communications and visualization experts before going live, an openness to new approaches, and a commitment to being highly selective to ensure that the displays are appropriate for public use.

Ethics of WBS

Health or bioethics guidance was originally developed primarily around connections between patients and healthcare providers where a clear relationship requires informed consent. Ethics guidance for public health surveillance, which clearly includes WBS, has been much less prevalent (316).

In 2017, prior to COVID-19 and the widespread application of real-time WBS, the World Health Organization commissioned a panel to develop ethics guidance for public health surveillance (317). When evaluated for WBS, 14 of the 17 resulting guidelines were directly applicable and adaptable to WBS. Perhaps most important is “Guideline 3. Surveillance data should be collected only for a legitimate public health purpose.” Also particularly relevant is “Guideline 13. Results of surveillance must be effectively communicated to relevant target audiences.” (317) Notably, the preamble to the guidelines emphasizes an imperative to engage communities transparently.

Nonetheless, criticism is still raised about potential ethical issues associated with WBS. Joh argued that WBS could be adopted for police surveillance to track illicit drug use (318). Van der Sloot (319) argued, without much discussion of the common applications of WBS for tracking pathogens, that potential ethical abuses may be possible with WBS and that it may be “flushing privacy down the drain” (319).

In spite of the WHO (2017) guidelines, McClary-Gutierrez et al. found that the “ethics of wastewater surveillance data sharing and use [was] not yet established.” (320, 321) Coffman et al. (322) noted that previous applications of WBS for monitoring infectious diseases including polio, hepatitis, and salmonellosis as well as narcotics usage and alcohol consumption supported the development of WBS ethics guidance, noting that many WBS investigators are natural scientists or engineers who may be unfamiliar with the aspects of health ethics guidance (322). Hrudey et al. (88) recognized the ethical issues involved with WBS for investigators and participants in the Canadian Water Network COVID-19 Wastewater Consortium (https://cwn-rce.ca/covid-19-wastewater-coalition/) (88).

Clearly, any public health surveillance activities, including WBS, need to attract mainstream public support. Given the politicization and polarization that public health measures attracted during the COVID-19 pandemic, applying sound ethical guidance to WBS to achieve mainstream support is more important than ever.

The WBS of SARS-CoV-2 at the community level via sampling at wastewater treatment plants provides valuable epidemiological information on COVID-19 in entire populations and contributes to evidence-informed decision-making for public health measures. In order to derive more granular data on COVID-19 kinetics in the community and to increase the actionability of this data, multiple efforts have been made to perform WBS on targeted sub-populations, including neighborhood- and university-campus-level monitoring as well as in individual facilities.

Targeted monitoring

WBS for SARS-CoV-2 across individual facilities has taken place in long-term care or correctional facilities, schools, university campuses, and hospitals. Evidence suggests that the WBS of these facilities may be a cost-effective tool for monitoring COVID-19 status (a single sample for a whole facility if the design of the sewage system allows) (323). There is no risk, and the routine operation of the facility is not interrupted (silent monitoring). Neighborhood and institutional-level WBS of SARS-CoV-2 have different objectives than at the community level. For example, 43% of COVID-19 deaths in Canada were in LTCF (324). Therefore, WBS of SARS-CoV-2 in long-term care aims to protect vulnerable elderly populations and prevent mortality. Conversely, in university dormitories, correctional facilities, and hospitals with congregate living, the aim is to prevent and control large-scale outbreaks to maintain the normal functioning of these critical infrastructures (325). Asymptomatic or pre-symptomatic individuals pose a high-risk potential for COVID-19 transmission.

Neighborhood-level monitoring

A modest number of investigative programs have been undertaken across communities at levels more granular than wastewater treatment plants (66, 91–93, 326, 327). Indeed, across an entire municipality, considerable variation exists across its geography with respect to cultural, social, and economic factors of its various residents. Community-based nested monitoring strategies that collect wastewater at different spatial scales within a larger city (e.g., from municipal zones, down to individual neighborhoods) in theory have the potential to provide a much more nuanced understanding of socioeconomic factors associated with infections (66, 93, 326). However, neighborhood-level WBS introduces additional challenges related to sample collection relative programs administered through wastewater treatment plants where infrastructure exists for sample collection. Accordingly, innovation in performing dynamic sample collection and the capacity to troubleshoot device malfunction in the field are required. While a more detailed understanding of analyte dynamics through communities is possible through granular neighborhood-level monitoring, relating measurements back to clinical cases is more challenging as confirmed cases are typically linked to the addresses of affected individuals. This is because individuals are likely to move between neighborhoods/sewershed catchments through the course of a day and their excretions reflect those movements, yet all diagnoses are related back to their original address (66).

Long-term care facilities

In contrast to the enormous amount of information on WBS of SARS-CoV-2 at the community level, few published reports exist on WBS of SARS-CoV-2 in LTCF. Davó et al. reported a pilot study of WBS of SARS-CoV-2 with clinical outbreak investigations in five nursing homes in northeast Valencia, Spain (328). Grab samples of wastewater were taken five times/week during variable study periods over 2.5 months. SARS-CoV-2 RNA levels in wastewater samples increased exponentially over the course of confirmed COVID-19 outbreaks in two out of five nursing homes. Lee et al. conducted WBS of SARS-CoV-2 at 12 LTCFs in Edmonton, Canada, since January 2021, aiming to develop, validate, and implement an early warning-rapid public health action system to protect the vulnerable (329). In this study, positive detection of SARS-CoV-2 in wastewater of these LTCF was highly concordant with clinical COVID-19 cases. Furthermore, more than half of SARS-CoV-2 positive detections led to confirmed COVID-19 outbreaks in the corresponding LTCF. In a few instances, known COVID-19 cases were in the facility but SARS-CoV-2 RNA was not detected in wastewater. These discrepancies were postulated to be caused by non-daily collection, COVID-19-positive staff not working in the facility, and COVID-19-positive residents using diapers and other continence aids. The key outcome, however, was that real-time reporting (within 24 hours) to public health authorities triggered an immediate outbreak investigation of those facilities with no confirmed cases, alerting them to strengthen infection prevention and control measures across the institution (330, 67).

Pico-Tomàs et al. (125) reported that active autosamplers were more sensitive than passive sampling devices using the modified Moore swabs (331); the former was able to detect SARS-CoV-2 in wastewater when COVID-19 prevalence was at 0.4% compared to the latter at 2.2% in their study (125). Using an active autosampler to collect 24-hour composite samples at least twice/week from the participating facilities, preliminary data analysis indicates that WBS of SARS-CoV-2 in LTCFs can be effective assuming adequate sampling/schedule and timely responses from the public health authority, despite some challenges remaining as mentioned above (330).

University campuses

Institutions of higher learning received disproportionate attention for SARS-CoV-2 WBS and COVID-19 modeling. While practical reasons for this high level of interest exist (i.e., teams innovating this technology have very frequently been university-based, making sample access and collection relatively easy), several features made universities important biological models:

• dormitory living where risk of COVID-19 rapid transmission is commonplace (229, 332–334),

• a vast array of building complexes with varying roles ensuring an appropriately complex model system that can be used to understand larger, more diverse municipal systems (335–337),

• high levels of social connectivity,

• more restrictive policies to mitigate the risk of COVID-19 transmission, and therefore the potential to assess the burden of disease relative to the larger external community without the same levels of intrinsic controls (97), and

• the presence of occupational health teams that enabled the assessment of disease transmission. Challenges relate to the lower proportion of symptomatic infections among younger people.

Campus monitoring confirmed that WBS-measured SARS-CoV-2 correlates with clinical cases, identifies asymptomatic infections that would not otherwise be observed, and identifies outbreaks and hot spots of COVID-19 activity (106, 229, 332–334). In monitoring residences, WBS demonstrated a 75%–85% positive predictive value and an 85%–90% negative predictive value for identifying locations with clinically confirmed infections (98, 100, 334). Many of these studies were designed as real-time learning and response models. For example, in a study at the University of California that applied WBS to the ~7,600 resident population, real-time alerts of positive wastewater signals were issued if one of the study’s 68 monitoring stations yielded a positive sample, encouraging residents to seek asymptomatic tests. The study was reported to have identified ~85% of cases (98).

Correctional facilities

Correctional facilities are unique: they represent a crowded congregate living scenario with guarded enclosures and limited ability for social distancing, making them different from settings such as university campuses and LTCFs. They can be hotspots for prolific SARS-CoV-2 transmission due to social and biological risk factors contributing to higher rates of COVID-19 morbidity and mortality (338). SARS-CoV-2 can spread rapidly in these conditions, especially with the arrival of newly infected incarcerated individuals (339). In a social study, Riback et al. indicated that integrating WBS of SARS-CoV-2 into surveillance strategies at correctional facilities may minimize the impact of future COVID-19 outbreaks (340).

Hospitals

Increasingly, WBS of SARS-CoV-2 in hospitals has been performed, including studies from Canada, USA, Brazil, Slovenia, and Thailand (94, 105, 341–345). The rationale for hospital WBS relates to the following:

• Hospitals are ideal systems to model virus shedding in wastewater. As a pandemic response system, hospitals contain detailed accounts of patients, and staff ensure accurate denominators exist for the WBS of SARS-CoV-2 RNA (94, 105). Unlike the community at large, hospitals are well-resourced for real-time investigations and active case finding, potentially capable of capturing all suspected cases regardless of symptoms. Furthermore, hospitals are well equipped with expert Infection Prevention and Control staff trained to review the timeline of COVID-19 hospitalized cases and classify them as community- vs hospital-acquired in the context of investigating and managing outbreaks. Furthermore, many use electronic health record systems that contain easily accessible and accurate data including co-morbidities, exposures, course of infection, treatments, and outcomes.

• No system benefits more from real-time WBS. Hospitals are where the most vulnerable and high-risk populations for severe COVID-19 infection reside. Furthermore, the health services offered in hospitals have very limited capacity for disruption and no backup system. Any service disruption related to COVID-19 outbreaks, e.g., closing outbreak units for admission to decrease exposure, has a tremendous impact and carries possible harm to all those in the hospital as well as those in the community needing high-level medical/surgical care. Strategies that might identify in-hospital transmission, allowing for early intervention and outbreak abrogation, would be profoundly impactful and serve to safeguard these critical resources.

Studies have generally identified increasing amounts of measurable SARS-CoV-2 RNA in hospital wastewater as a function of the total number of patients for a range of SARS-CoV-2 gene targets including N, E, and S genes (94, 105, 342). Furthermore, targeted nested qPCR (105, 341) and untargeted metagenomic sequencing (345) documented a range of mutations and variants. Impactfully, even among a large number of individuals hospitalized and recovering from community-acquired infections, spikes in WBS-measured SARS-CoV-2 N1 and N2 were found to strongly correlate with nosocomial infections and unit-specific outbreaks (94, 224). These data further highlight the importance of understanding viral shedding dynamics from infected populations where, in hospitals, it was first apparent that SARS-CoV-2 fecal shedding is highest in the first days of illness, and thereafter rapidly dropping to low levels that persist for an extended time. These studies demonstrate that SARS-CoV-2 WBS is particularly sensitive to acutely occurring cases, exactly what is required in outbreak situations.

MOVING BEYOND SARS-COV-2 TO OTHER PATHOGENS

WBS existed before COVID-19. However, it was principally an academic pursuit, only rarely considered or applied for real-time surveillance. However, the COVID-19 pandemic demonstrated the transformative potential of this platform technology in managing and responding to an infectious disease health crisis. WBS has since been embraced as an objective, inclusive, cost-effective, real-time tool to identify and measure SARS-CoV-2 in wastewater systems, enabling COVID-19 to be modeled across the world. Based on these successes, WBS is rapidly being applied to new areas of infectious disease or operationalized for real-time surveillance of established targets in others. However, for WBS to detect and track an infectious pathogen, a detailed understanding of the pathobiology of each agent is required (Fig. 5). This will include:

Fig 5.

Natural history of analyte. The performance of wastewater-based surveillance of infectious targets is influenced by the natural history of infections they cause and how the measured analyte enters the sewershed. (A) For an agent to be tracked by WBS, its nucleic acid must ultimately end up in the wastewater. Nucleic acid may come directly or indirectly from stool, urine, vomit, and that exfoliated from the skin during bathing/showering. (B) The wastewater-measured signal will provide different population-level information depending on the natural history of the pathogen targeted, incidence (i.e., acute infection with rapid resolution of analyte shedding such as COVID-19), overall prevalence (i.e., chronic colonization or disease associated with protracted analyte shedding), or a combination of both (i.e., chronic viral diseases with acute seroconversion illnesses, i.e., HIV).

• determining how viral dynamics in wastewater correlate with the natural history of human infection (i.e., acute shedding only at symptom onset vs persistent shedding in the context of chronic infection) to understand its distribution across space and time in communities.

• establishing how the characteristics of the agent of study (i.e., enveloped vs non-enveloped viruses), hydrophobicity, etc., influence its distribution within the wastewater as a complex matrix.

• understanding how each agent enters the wastewater system (i.e., through secretions such as feces and urine or through hygienic activities such as washing or teeth brushing). As with SARS-CoV-2, understanding in which fraction of wastewater the analyte of interest separates is key, as is identifying relevant normalization agents.

SARS-CoV-1

WBS was not used as a real-time surveillance tool during the first SARS pandemic in 2003. Only one study reported detecting SARS-CoV RNA in sewage from two hospitals in China, with no infectious virus isolated (346). Conversely, sewage management was made famous due to a suspected link between sanitary plumbing and a common source outbreak, with 99 cases diagnosed in one building of a large private apartment complex in Hong Kong (347, 348). The index case had diarrhea with high fecal viral shedding, and factors besides person-to-person contact seemed to lead to many secondary cases also experiencing diarrhea in the building. Environmental investigations found that the U-traps of many bathroom floor drains were dry (thus non-functional), and a crack was identified in a sewer vent pipe that theoretically could allow virus-laden aerosols to disperse throughout the lightwell of the building. However, no conclusive evidence of transmission was found. In terms of transmission pathway, compared to SARS-CoV-2, SARS-CoV-1 peaks later in illness (days 6–10) in respiratory samples, and the viral load in fecal samples is much higher than in respiratory samples in some patients (349). On the other hand, one study observed that stool samples from five patients who had cultivable SARS-CoV-1 from autopsied intestinal tissue did not carry any infectious virus. This is consistent with the short duration of viral infectivity in stool samples (350, 351). The short time span of the SARS-CoV-1 pandemic, with the first case on 16 November 2002, in China and the last reported case on 13 July 2003, in the US, was a relief globally, and thus further efforts to hone SARS-CoV-1-related WBS were abandoned (352).

Respiratory viruses beyond SARS-CoV-2

When/if SARS-CoV-2 becomes an endemic respiratory virus, it will still be possible for antigenically distinct variants that can evade vaccine- or natural-acquired immunity to emerge, as we learned with Omicron (353). Thus, WBS still has a role in disease surveillance in a post-pandemic world, especially with the dramatic reduction in clinical COVID-19 testing capacity. However, in a post-pandemic world, COVID-19 will only represent a fraction of the total respiratory viral disease burden in the community, given the diverse agents responsible for colds and influenza-like illnesses (354, 355). Given the profound impact of seasonal respiratory viral infections, efforts to assess respiratory disease activity comprehensively through WBS have great promise.

Influenza

Minor changes in surface-expressed hemagglutinin and neuraminidase genes manifest in antigenic drift for both influenza A and influenza B, resulting in annual seasonal epidemics in temperate climates (i.e., November–March in the Northern hemisphere and April–September in the South) (356). In the US alone, 9–45 million infections/year result in 140–810,000 hospitalizations and 12–61,000 deaths on average (357). In 2023, strains in circulation include influenza A: H1N1 and H3N2, and influenza B: B/Yamagata and B/Victoria. Co-circulating in animal reservoirs in addition to humans, influenza A exists in a great number of subtypes (H1–18 and N1–11) further susceptible to antigenic shift where major changes in these genes due to recombination manifest in a novel virus that is antigenically distinct, and for which no cross-protection exists. The result may be a pandemic with variable potential for morbidity and mortality (356, 358). The most recent human pandemics include 1889–1890 (H3N8 vs H2N2), 1918–1919 (H1N1, avian origin), 1957–1958 (H2N2, avian origin), 1968–1969 (H3N2, avian origin), 1977–1978 (H1N1), and 2009–2010 (H1N1, avian and porcine origins). Sporadic cases of H5N1 (avian origin) continue to cause great concern for its pandemic potential (359).

Accordingly, influenza has long been recognized as a high-value target for WBS. Early efforts were generally insensitive at detecting influenza (360). However, with the rapid expansion of WBS, teams across the world have started to report on wastewater-measured influenza A and B RNA gene targets using extraction protocols developed to assess for SARS-CoV-2 RNA. As of 2023, efforts to establish how well these measurements correlate to clinical cases are only just being determined using municipal (172, 361) and campus-based model systems (362). Mercier et al. after establishing that influenza A RNA partitioned into the primary sludge fraction, compared wastewater collected over 4 months with clinical-test positivity rates in Ottawa, Canada. They demonstrated a moderate correlation with the RT-ddPCR-measured influenza A membrane protein gene normalized for fecal matter using PMMoV, potentially proving a leading indicator of case occurrences (172).

Ultimately, influenza-based WBS programs could inform administrative and patient-care decisions. Prospective WBS relying on quantitative measurements can identify the changing prevalence of infection and redistribute health resources and labor to anticipate the increased care requirements of afflicted individuals/populations (in hospitals, LTCF, etc.). WBS could also inform the appropriateness of empiric anti-viral therapies for those presenting with influenza-like illness (363–365). Monitoring gene targets for allelic variation could inform on the strains in circulation, assessing in real time for vaccine match/mismatch (366, 367) [even potentially informing adaptive platforms to account for this (368)] and monitoring for resistance to first-line antiviral therapies (369, 370). Adapting these technologies to monitor the natural environment and industrial aviaries for the changing prevalence and patterns of influenza A would allow early detection of changing patterns and new strains of pandemic potential (371, 372).

Respiratory syncytial virus

Human RSV, is an enveloped non-segmented negative-sense RNA virus with two major antigenic subtypes (A and B) and follows a similar seasonal pattern of disease (373, 374). While it can infect anybody, it causes disproportionate morbidity and mortality among neonates and young infants and, to a lesser degree, the elderly and immunocompromised (375). A systematic analysis of the global burden of RSV in 2019 estimated 33 million RSV-associated respiratory infections, 3.6 million associated hospital admissions, and 26,300 in-hospital deaths, and >100,000 overall RSV-attributable deaths in children aged 0–60 months (376). As of 2023, no effective antivirals against RSV exist (except for the potential of ribavirin as a rarely used supportive therapy among patients with severe immune dysfunction and life-threatening infections (377, 378). Until recently, the main strategy to decrease RSV-related hospitalization and morbidity in countries with resources and access is passive immunization with monoclonal antibodies in high-risk neonates and infants (379). With the high cost of this treatment and the annual variation of the start and the end of the RSV season, it makes perfect sense that passive prophylaxis programs should be initiated according to the local RSV circulation pattern (373, 375, 380). Recently, vaccines protecting against the RSV F protein have been developed and have demonstrated efficacy in older adults, and its implementation is expected to reduce incident infections in at-risk populations (381, 382).

The progressive lifting of COVID-19 restrictive measures coincided with a surge in RSV cases, causing considerable pressure on the acute care system (383–385). Some work has begun to forecast the changing burden of RSV infection, leveraging SARS-CoV-2 surveillance networks, and taking advantage of the similar biology of viruses (173, 361). Using a custom-designed RT-ddPCR-assay (detecting both RSV A and B) for the nucleocapsid protein gene in settled solids from wastewater treatment plants in the Santa Clara and San Mateo counties of California, Hughes et al. compared values with state-wide RSV-testing positivity rates, finding strong correlations when assessed raw and normalized for PMMoV (173). Such assays could be used to inform the timing of RSV preventative strategies for high-risk infants (386, 387). From the viewpoint of managing individual pediatric patients presenting with acute viral respiratory infections, there have been calls, e.g., Choosing Wisely, to decrease pathogen testing among pediatric patients with bronchiolitis due to similar supportive treatment irrespective of the viral pathogen (388, 389). With the decrease in clinical testing for RSV, an urgent knowledge gap needs to be filled about whether the WBS of RSV, while shown to correlate with overall clinical prevalence, is a good surrogate for defining the start and end of RSV among neonates and infants who—interestingly—are in diapers and therefore not contributing to wastewater (173, 361, 390).

Gastroenteritis viruses: norovirus, rotavirus, and others

The five viruses classically included as common causes of acute gastroenteritis are norovirus and sapovirus (both belonging to the family of calicivirus), rotavirus, astrovirus, and enteric adenovirus (391, 392). With the advancement of molecular diagnostics and after rotavirus vaccine programs were implemented in many countries, norovirus became well known as the most common cause of sporadic gastroenteritis and institutional foodborne and waterborne outbreaks causing significant disease and economic burden (393–400).

Several studies have been published on the epidemiology of gastroenteritis virus in wastewater that showed a correlation between viral RNA level in wastewater with expected seasonality and human disease (401–403). WBS can be useful in population-based surveillance of the disease burden of the gastroenteritis virus especially where the capacity for clinical diagnosis is limited, as was the case with Omicron in the COVID-19 pandemic. Another important application of WBS for the gastroenteritis virus is characterizing strains in wastewater samples to identify the full spectrum of circulating and emerging strains to develop vaccines, and for post-vaccination surveillance for, e.g., rotavirus (404–409). Surveillance for strains responsible for atypical syndromes including neurologic disease (i.e., EV71) (410) and non-classical hepatitis (411) would be of tremendous benefit.

Adenoviruses are a diverse family of more than 60 serotypes, causing a range of respiratory and gastrointestinal illnesses and rarer neurologic and acute hepatitis presentations (412, 413). Their impact is disproportionate in children and elderly immunosuppressed populations. Despite their prevalence in general populations, clinical testing is generally limited. Pellegrinelli et al. described early efforts to use WBS to model adenovirus disease in Milan, Italy. Between 2021 and 2022, the authors collected wastewater from associated treatment plants and used a custom-designed real-time PCR assay to quantify the adenovirus hexagonal gene DNA. These were compared to adenovirus tests performed at the largest associated local research hospital (all respiratory and fecal tests), demonstrating a correlation with the percentage of positive tests (414).

Enteric viral hepatitis

In much of the developing world, fecal-oral transmission of viral hepatitis continues to be common. Work prior to and since the COVID-19 WBS revolution has identified hepatitis A (403, 415, 416) and hepatitis E (415–418) virus RNA in the wastewater of municipal populations. More recently, investigators have even been able to identify the occurrence of specific strains associated with disease in particular at-risk communities, such as men who have sex with men (419). Correlating wastewater-measured virus and clinical disease has been challenging owing to the high frequency of asymptomatic infections (particularly in younger individuals) and much work remains.

Polio

While the number of publications related to WBS has increased dramatically since 2020 with the global use of WBS for the COVID-19 pandemic, the World Health Organization has used WBS as an environmental surveillance tool in monitoring its goal of polio eradication since 2003, first using viral culture to detect poliovirus (420). In particular, effective real-time programs have been developed and widely used by Israel (421–428). Modern-day WBS combined with genetic sequence and phylogenetic analysis was a useful tool in mapping geographic areas at risk for wild-type poliovirus in Pakistan (429). Unfortunately, besides wild-type poliovirus, the paradox of the oral polio vaccine (OPV) (430), i.e., vaccine-derived poliovirus (VDPV) becoming a concerning cause of polio, came true due to implementation challenges, including the shortage of inactivated trivalent polio vaccine (IPV) that was to be administered with a globally synchronized switch from trivalent OPV to bivalent OPV (serotypes 1 and 3) planned between April and May 2016 after the wild-type poliovirus serotype 2 was certified as eradicated in September 2015 (431–433). IPV protects 99%–100% against paralytic myelitis but does not 100% prevent infection and transmission of wild-type poliovirus or VDPV. Despite efforts to improve mucosal immunity and protection against the serotype 2 virus with supplemental immunization with trivalent OPV, the gap in protective immunity after the global switch resulted in a large number of VDPV serotype 2 (VDPV2) outbreaks, especially in Africa, with more than 500 cases in each of 2020 and 2021 (434). Reports of a VDPV serotype 3 (VDPV3) paralytic poliomyelitis case in an unvaccinated young child in Israel, a VDPV2 paralytic poliomyelitis case in an unvaccinated adult in the US, and the identification of VDPV3 from airport sewage-testing in Poland illustrated that the circulation of VDPV strains from local or travel-related OPV recipients can present risks to susceptible individuals globally (435–437). Routine poliovirus WBS since 2014 in the UK identified that geographic transmission of VDPV2 provided early warning to initiate rapid public health responses, including enhanced IPV campaign, and demonstrates the utility of WBS (438). However, limitations of WBS, including the sensitivity of identifying a target depending on the sewershed and the population monitored, and the public health objectives and intended actions for expected results need to be determined before implementing WBS for poliovirus (439, 440).

More recently, WBS programs have identified imported VDPV2 in wastewater collected in large urban centers including New York (436, 441), London (442,443 ), and Montreal (444). Whereas clinical cases were never identified in London or Montreal, a case of acute flaccid myelitis attributed to VDPV2 in an unvaccinated resident of Rockland County was observed (the first case of polio in the United States since an imported case was documented in 2013) (440). How common VDPV2 is in wastewater in the developed world and how this relates to the travel/immigration of individuals receiving OPV is as yet unknown. However, even these preliminary observations have prompted early actions including expanded testing in the US in areas with low vaccination coverage (445), and in the UK, the administration of booster or catch-up IPV vaccines to children (446).

Zoonoses and emerging zoonotic infections capable of human-to-human transmission

While the vast majority of research in WBS focuses on agents that are transmissible from person to person, efforts in different jurisdictions are underway to assess several zoonotic infections. For example, arboviruses such as dengue virus are shed in feces and urine and are therefore potentially amenable to WBS as has been pursued by some groups (447). The potential for informing public health response is high, especially given the seasonality of dengue, the variability of symptoms, and the potential for outbreaks of dengue hemorrhagic fever with the importation of new serotypes into an area with high levels of prior infections (448, 449).

Mpox (formerly known as monkeypox) is an enveloped, DS-DNA member of the Orthopoxvirus genus and Poxviridae family. Mpox is a zoonosis and has been identified as causing disease in a range of mammals. In humans, infection typically begins with systemic symptoms of fevers and rigors, followed 1–3 days later by characteristic vesiculo-pustular skin lesions first on the face and spreading caudally (450). Endemic in Central and West Africa, Mpox cases began to increase after the discontinuation of smallpox vaccination after its successful eradication (451). Clusters of cases have been reported in non-endemic areas, including an outbreak in the mid-West US in 2003 related to contact with prairie dogs exposed to rodent populations from Ghana (452). In early 2022, clusters of disease unrelated to travel and with an atypical epidemiology and clinical manifestations were first identified in Europe. Cases were occurring predominately in men who have sex with men (453). The 2022 Mpox outbreak is hypothesized to be transmitted primarily through sexual contact (although household, non-sexual secondary cases have infrequently been identified) (450). As of 15 March 2023, about 86,500 cases worldwide have been identified. Given that Mpox represented the first global “new infection” in the post-COVID-19 age, it was not surprising to see its study by WBS teams. Investigators hypothesized that Mpox DNA may enter the sewer system through multiple routes (viral DNA is known to be shed in feces, urine, and skin lesions potentially shed during bathing) and have been found in wastewater systems (454–457). Early work in California demonstrated positive correlations between wastewater-measured DNA in nine wastewater treatment plants and incident Mpox clinical cases within their catchment areas (458). Efforts to pre-emptively vaccinate at-risk populations have proved effective; incident cases have fallen to <5 per day since January 2023 (450, 459). Efforts to onboard programs in other areas have slowed due to the very low incidence of new cases and the rarity of asymptomatic cases (460).

Parasites

It is well known that parasitic exposure and infections vary vastly between geographic areas, but a review observed that the geographic prevalence of many human parasitic worms still remains unknown (461, 462). It may be possible for WBS to estimate the burden of parasites whose lifecycle involves the human gastrointestinal and urinary tracts. However, most of the work involving wastewater and parasites relates to waterborne parasites, with most of the published work on Cryptosporidium spp. showing variable prevalence in different communities (463, 464).

Cryptosporidiosis can cause infections in humans ranging from asymptomatic to severe and prolonged gastrointestinal symptoms in immunocompromised hosts. There are at present 44 Cryptosporidium spp. with variability host range (human and animal susceptibility) depending on the species. Treatments for Cryptosporidium spp. infections have variable effectiveness (465).

The disproportionate effort placed on studying Cryptosporidium spp. as a WBS target relative to other parasites likely relates to its predominant role in outbreaks (466). Indeed, there is an obvious reporting bias, with over 80% of all waterborne parasite outbreaks worldwide between 2017 and 2020 being reported by the US (56%), the UK (20%), and New Zealand (10%) despite their greatest impact occurring in developing nations. This is another knowledge gap that WBS can fill if the topic is of public health importance and resources are made available to low-resource countries (467). WBS can also provide useful information on whether climate change may be modifying the epidemiology of parasites (468). The reclassification of Cryptosporidium to a new subclass, Cryptogregaria, and its ability to reproduce without a host could represent a paradigm shift in terms of the risk its distribution might cause in the water industry (469).

Antimicrobial resistance

The emergence and spread of antimicrobial resistance (AMR) are a profound global health threat (470, 471). Individuals infected with antimicrobial-resistant organisms (ARO) experience higher morbidity and mortality and disproportionally consume limited health resources (472–477). AMR will continue to proliferate, driven by the use, misuse, and overuse of antibiotics in humans (471), animals (478–481), and agriculture (482–484). By 2050, ARO-related deaths are expected to rise to 10 million/year globally (485–487). Beyond just health outcomes, AMR is expected to have broad social impacts including reduced quality of life, social connectivity and trust, and discrimination (487–489). The World Bank estimates that AMR will result in >$100 trillion/year in lost productivity across the globe by 2050. Within the world of AMR, the One-Health approach to surveillance and control of AMR has long been advocated (490, 491). Using a One-Health approach, strategies to monitor and mitigate AMR must simultaneously be implemented across human, animal, and natural environment reservoirs since bacteria and ARO genes can and will be exchanged between compartments (492). Water serves as the unifying link. Water transports nutrients, pollutants, and other population analytes and is a major bacterial habitat, and therefore a potent reservoir that aids in AMR dissemination (493–495). Accordingly, the WBS of AMR is increasingly seen as a tool that must evolve to counter this impending crisis.

WBS of AMR is likely to be particularly effective for pathogens that naturally colonize or pass through the gastrointestinal tract and enter the wastewater system through feces. Metagenomics studies have increasingly identified ARO genes in wastewater, including metallo-beta-lactamases (MBL), such as New Delhi MBL (ndm), vim, imp, and mobile colistin resistance (mcr) (496–498). Few population-based studies of ARO genes in wastewater have been performed (499–508). One notable study compared the AMR gene burden measured from wastewater treatment plants in seven countries, confirming that both raw and 16S rRNA-normalized gene levels were disproportionally high in countries with greater antibacterial consumption and lower socioeconomic status based on static WHO measurements (509, 510).

Both culture-based and molecular/genomic strategies have been applied for the WBS of AMR. For example, a number of culture-based studies have focused on Enterobacterales (507, 511). For instance, Blaak et al. used this approach to demonstrate the wide dispersion of carbapenemase-producing Enterobacterales (CPE) across wastewater treatment plants in the Netherlands, suggesting a prevalence of 1 in 5,000 to 1 in 20,000 individuals (507). In this study, CPE were identified in all wastewater samples, regardless of whether hospitals were involved. The authors estimated that CPE were present at a level 250 times lower than extended-spectrum beta-lactamases (ESBLs). However, in contrast, others found that CPE were only present in wastewater treatment plants that received hospital-based wastewater (508). Several hospital-based wastewater studies have likewise been performed. Cai et al. performed weekly WBS in a single ~250-bed hospital in Shantou, China for ARG using metagenomics over several months. They compared their data with pathogens from 104 individuals (16% of submitted clinical specimens) and used network analysis to demonstrate strong correlations with many ARG types and recovered clinical pathogens (512). Carlsen et al. performed culture-based surveillance over 4 days to identify CPEs related to prior outbreaks (513).

Many complications will impact the future operationalization of AMR WBS. Many of the most important AMR gene targets (i.e., ESBLs and carbapenemases) are challenging to track through One-Health compartments. Their low initial inoculum and potential for overgrowth in environmental reservoirs by faster-growing organisms require novel strategies (493, 514, 515). Furthermore, an intrinsic limitation to short-read-based metagenomics analysis is that it is not possible to determine if the gene of interest is from a “pathogen” or a commensal (516), nor can it be attributed to a living or dead organism (517). In contrast, culture-based methods can identify organisms carrying ARO genes of interest, but such strategies are resource-intensive and cannot provide a comprehensive and agnostic characterization of the resistome (508, 516, 518). Also, bacteria measured in wastewater may come from organisms released from biofilm and biofouling in pipes and therefore may not represent direct contributions of colonized/infected individuals, unlike their viral counterparts (519). Relative to virus-targeted WBS, ARO poses unique challenges. While the processes required for sample collection, concentration, DNA extraction, sequencing, and analysis are analogous, the sheer volume and diversity of ARO gene targets are much greater. Furthermore, resistance to individual drugs is frequently mediated by multiple gene types (related and unrelated), and the selection pressure for a resistance gene may be independent of the mechanism of action of that enzyme owing to the co-location of ARO genes on mobile elements.

Ultimately, WBS may serve to provide both an early warning of local and regional changes in ARO gene abundance (i.e., the introduction of a new target) and comparisons across sewersheds that provide insight into healthcare, socioeconomic, and environmental factors contributing to those patterns (509, 520). The WBS of AMR in hospitals could be transformative for infection control and antimicrobial stewardship, providing a leading signal relative to conventional clinical surveillance, which requires those colonized to develop symptoms of an infection and specimens to be collected that eventually result in an pathogen being identified, and could therefore identify outbreaks before clinical cases are identified.

Tuberculosis

Tuberculosis (TB) continues to cause considerable morbidity and mortality, particularly in the underdeveloped and developing countries. The majority of disease is caused by Mycobacterium tuberculosis, although other members of the Mycobacterium tuberculosis complex can cause human infections with similar presentations (i.e., Mycobacterium africanum and Mycobacterium bovis). The spectrum of diseases caused by TB is vast. After primary infection, ~90% of immunocompetent individuals control bacillary replication-forming granulomas, and the bacteria enter a latent phase. Only a portion of those with latent disease will reactivate. Risks include increasing age and immunosuppression. An estimated 22% of the world’s population has been infected with TB, with an estimated 10 million new infections per year (521). TB is responsible for ~1.5 million deaths per year (522). TB’s morbidity and mortality are further accentuated in those co-infected with HIV, disproportionally occurring in sub-Saharan Africa (523, 524). Accordingly, there is great interest in population-level surveillance for TB, given the challenges associated with its clinical diagnosis. Preliminary studies have assessed the viability of M. tuberculosis as a WBS analyte in Durban, South Africa (525, 526). Investigators collected wastewater from three municipal wastewater treatment plants. qPCR probes targeting total Mycobacteria 16S rRNA DNA, MTB complex (Rv0577), M. tuberculosis (RD9), M. bovis (RD1), and Mycobacterium caprae (RD4) were used on the extracted DNA from pelleted wastewater, where they identified each target. Furthermore, the authors observed significant differences between sites and at individual sites over time. Values were elevated at plants that received sewage from TB programs (525). Studies to correlate WBS gene content with clinical burden in sewershed catchments have yet to be reported. As a wastewater-measured analyte, TB is more complex than many traditional markers based on its complex pathobiology. It would be unlikely and less desirable for those with latent TB to be detected by WBS. More likely, those with active pulmonary disease are expected to be identified as a result of swallowed respiratory secretions (527) or those ~20% with extrapulmonary TB excreted in stool (528). In addition to monitoring for changes in disease activity across a region, WBS of TB could play a considerable role in empiric treatment. Given the strong correlation of in vitro anti-mycobacterial susceptibility testing with specific genomic markers (529–531), it is conceivable that WBS for multi-drug resistant and extreme drug-resistant TB could occur, and regimens could be tailored based on the local wastewater-directed antibiogram.

WBS for fungal and AMR fungal pathogens has not been explored to any significant effect as yet. Notably, however, a strain of Candida auris, linked to a clinical disease across multiple healthcare facilities including hospitals was cultured in community wastewater in Southern Nevada, demonstrating that this may indeed be possible (532).

WASTEWATER-BASED SURVEILLANCE IN RESOURCE-LIMITED SETTINGS

Many low- and middle-income nations (LMIC) experience a high burden of incident and prevalent infectious diseases that are either current (i.e., norovirus, Mpox, hepatitis A, and hepatitis E) or potential future targets of WBS (i.e., tuberculosis, malaria, dengue, Ebola, and zika). Despite the early success of academic WBS programs for SARS-CoV-2 in LMIC (313), most countries are poorly equipped to perform WBS on a large, nationwide scale (533). For example, many LMIC have a dearth of formal wastewater treatment plants and programs. More commonly, gray and black water may be discharged directly into nearby water systems without treatment. For example, India is estimated to treat <40% of wastewater, and in some LMIC rates may be <10% (533–535). Accordingly, WBS programs developed in these settings lack traditional centralized infrastructure and operational personnel to collect composite samples favored by high-income countries (HIC). To compensate, many LMIC WBS programs more frequently operate using “grab” samples from open drains and pit latrines (536, 537). Other differences too exist (314). Keshaviah et al. conducted an online survey regarding wastewater-based surveillance programs, which included representatives from 43 countries (536). Major differences between LMIC and HIC programs generally related to financial support, which limited the frequency with which sampling could be performed and precluded more advanced investigations such as those to detect and monitor emerging variants.

A significant proportion, and in some cases, the majority of homes in many of the larger urban centers of several LMIC, may not be connected to centralized sewer systems (538). Accordingly, with the limited centralized infrastructure, WBS programs in LMIC may not be able to claim to be “comprehensive and inclusive”—as is the case for programs in HIC. However, strategically implemented WBS sites in sentinel sewersheds would still have profound impacts in these resource-limited settings. In particular, WBS offers a cost-effective, real-time measure enabling an understanding of the local prevalence of individual infectious agents within these resource-constrained settings (where individual diagnostic testing may not be feasible for acutely ill patients, and clinicians therefore generally rely on empiric therapies). Ultimately, innovation in WBS allowing for point-of-use testing (i.e., analogous to point-of-care testing for individuals) is required in order to achieve a more rapid turnaround—which has been a greater challenge in LMIC (536). For this to occur, the incorporation of novel technologies, such as paper-based microfluidic devices, will likely be required (539), which will reduce the reliance of testing on those with significant technical expertise, which to this point has necessitated centralization of services.

SAFETY MEASURES FOR THOSE WORKING WITH WASTEWATER

This section focuses on risk assessment and the prevention of infections when working with wastewater, i.e., it does not include occupational hazards occurring as a result of fieldwork or from exposure to chemicals, e.g., hydrogen sulfide, dust, and endotoxin previously studied (540–544). In terms of risk assessment and prevention of infectious disease for WBS, there are several key principles:

• understanding pathogen transmission pathways and bioburden,

• suitable personal protective equipment (PPE) (including the appropriate donning and doffing protocols to reduce the risk of transfer) with decontamination steps as needed to prevent self-contamination,

• vigilance, with rigid hand hygiene coupled with equipment and environmental decontamination as needed to prevent residual exposure,

• proactive preventative measures including active immunizations for vaccine-preventable illness, and immunization, active or passive with/without prophylactic treatment in the event of exposure.

These principles are applicable for wastewater treatment plant workers and research laboratory staff who handle and process wastewater samples.

With the advancement of WBS and the use of molecular technology including metagenomics, the breadth of potential pathogens that will be monitored in wastewater increases (545). While most of the studies described herein involve the detection of nucleic acids of pathogens (which is not equivalent to replication-competent pathogens that can in theory be acquired from wastewater), important steps to ensure the safety of wastewater treatment plant workers have been highlighted by Brisolara et al. (546) and a PPE matrix proposed by LeChevallier et al. (547). In short, the authors stated the importance of being mindful and performing ongoing risk assessment when working with wastewater, taking precautions to minimize the generation of droplets or aerosols, wearing appropriate PPE, washing hands often, cleaning and disinfecting the working environment, avoiding eating and drinking in laboratories and environment at risk of contamination, handling and washing possibly contaminated work clothes separately, and in research laboratories, specimens with infectious risks should always be processed in a biosafety level-2 cabinets. During a pandemic with emerging pathogens, continually reviewing guidelines (548–553) and reassessing potential risks are essential for ensuring ongoing safety. Communicating safety procedures and providing training and adequate PPE for staff, affirming that appropriate processes protect, are all important components of maintaining occupational health and safety. Receiving routine childhood vaccines including tetanus, polio, and hepatitis B are important steps to provide immunity and protection (548). In addition, the US CDC recommends hepatitis A and typhoid vaccines as well for those at occupational risk.

THE FUTURE (AND POTENTIAL) OF WBS

WBS is a platform technology with transformative potential for studying infectious diseases and public health. Serial monitoring of populations can be performed at a cost of mere pennies per person per year, with minimal intrusion through surveillance at municipal wastewater treatment plants. The information generated is in near-real time, providing objective, comprehensive, and inclusive data on the prevalence and distribution of the target pathogen across space and time. However, for WBS to be meaningful, significant preparatory work will be required for each individual infectious agent including optimizing workflows for the efficient extraction, concentration, identification, and quantification within wastewater, and a complete and thorough understanding of how, when, and under which circumstances the pathogen enters the sewershed. Correlating the presence and abundance of target analytes in wastewater with the burden of clinical disease in a catchment area will likely prove the most challenging, as it is unlikely that the clinical resources for active population monitoring (to establish their prevalence) of conventional pathogens will ever rival the resources dedicated to COVID-19 monitoring. Accordingly, novel approaches will be needed as WBS continues to evolve. In addition to infectious diseases, a vast range of population health determinants can be assessed through wastewater using similar strategies and workflows as developed for infectious diseases, providing insight into great many other factors including chemicals and toxins, dietetics, substances of abuse/misuse, etc. Given the significant costs associated with a logistical network required for the efficient and appropriate collection and processing of wastewater, the greater the collected samples are leveraged across a wide range of surveilled analytes the greater the impact, and the lower the relative cost of each additional analyte.

Biographies

Michael D. Parkins completed his training in Internal Medicine and Microbiology at the University of Calgary. He is now a Professor in the Department of Medicine and Department of Microbiology, Immunology and Infectious Disease in the Cumming School of Medicine at the University of Calgary. Dr. Parkins serves as the Section Chief of Infectious Diseases and Medical Director of the Home Parenteral Therapy Programs across Calgary in Alberta Health Services and as the Clinic Director of the Southern Alberta Adult Cystic Fibrosis Clinic. His research interests focus on CF airway disease, infection transmission, and the epidemiology of antibiotic-resistant organisms. Dr. Parkins has most recently worked to advance wastewater-based surveillance as a platform technology capable of identifying and tracking targets across the sewershed as an actionable public health tool.

Bonita E. Lee is a professor at the Department of Pediatrics, University of Alberta, Edmonton, Canada. She received her MD in 1992 and completed a pediatric residency, pediatric Infectious Disease fellowship, and an MSc in Epidemiology there. She worked for almost 10 years as a medical virologist at the Provincial Laboratory for Public Health in Edmonton and Alberta and was program leader for gastroenteritis virus, respiratory virus, HIV, prenatal programs, and laboratory-based surveillance until 2010. She also worked as a pediatric infectious disease consultant until switching her clinical work to Infection Prevention and Control (IPC) in 2013. Currently, she is the IPC medical lead for the Stollery Children’s Hospital in Edmonton and was one of the recipients of the Pediatric Chairs of Canada 2020 COVID Leadership Award. Her research and publications focused on infectious disease epidemiology of gastroenteritis, respiratory, bloodborne, and perinatal infection pathogens.

Nicole Acosta received her PhD in Microbiology and Biotechnology at the University of Alberta, Edmonton, Canada. She is a research associate at the Cumming School of Medicine at the University of Calgary, Calgary, Canada. Her research focuses on understanding the epidemiology of infectious diseases by using molecular biology and bioinformatic approaches. Her publications focus on the molecular analysis of heterogeneous and polymicrobial medical and environmental samples to understand their constituents. She is the current Molecular Team Leader for the wastewater-based surveillance (WBS) project at the University of Calgary.

Maria Bautista is an environmental microbiologist and virologist specializing in molecular methods for studying diverse microbial communities. She focuses on assessing microbial and viral diversity, community changes following disturbances, and temporal/spatial comparisons. In her current role as a Senior Research Associate at the Microbial Markets and Geomicrobiology group at the University of Calgary, Maria Bautista serves as one of the technical leads for the Pan-Alberta wastewater monitoring program. Since 2020, she has been dedicated to testing, developing, and applying methods for extracting, quantifying, and sequencing SARS-CoV-2 and other pathogens in municipal wastewater.

Casey R. J. Hubert is a professor and research chair in the Department of Biological Sciences, where he leads the Geomicrobiology group. This research team focuses on molecular microbial ecology and metagenomics on samples from extreme environments such as deep-sea marine sediments and other deep biosphere habitats. The group specializes in extracting nucleic acids from these challenging environments, enabling microbiome assessments using different sequencing approaches. Bacterial endospores and their biogeography are of particular research interest. This approach has led to new insights at the intersection of geology and microbial ecology, as well as different practical applications in the energy sector. In 2020, the group began applying these methods and approaches to municipal wastewater for both quantification and genomic sequencing of SARS-CoV-2 and other pathogens, contributing to providing population-level monitoring as a valuable public health tool in Alberta, Canada.

Steve E. Hrudey is Professor Emeritus, Faculty of Medicine and Dentistry, University of Alberta. He chaired the 2022 Royal Society of Canada Expert Panel on Wastewater Surveillance for SARS-CoV-2 in Canada. Among 28 expert panels over his career, he also served from 2000 to 2002 on the Research Advisory Panel to Justice Dennis O’Connor’s Walkerton Inquiry into a fatal drinking water outbreak, on the Expert Panel for federal Minister Jim Prentice on Safe Drinking Water for First Nations in 2006, and he chaired the international expert panel on bladder cancer and chlorination DBPs in 2014–2015 for the Water Research Foundation, Denver. His research, involving over 200 refereed journal publications, 10 books, and 29 book chapters, has focused on public health, environmental risk, and ensuring safe drinking water. Dr. Hrudey was elected as a member of the Order of Canada in 2020 and the Alberta Order of Excellence in 2017.

Kevin Frankowski is the Executive Director of Advancing Canadian Water Assets (ACWA), a partnership between the City of Calgary and the University of Calgary that focuses on research and innovation, including technology de-risking and validation. Kevin is also a co-leader of the pan-Alberta wastewater monitoring program, which monitors 80% of Alberta’s population (and >94% of Alberta’s urban population) three times per week. Kevin Frankowski is also a business leader and technology development specialist who leverages innovation and strategic vision to unlock growth. He has led a large consultancy business unit, is a serial entrepreneur who has co-founded five startups, and has led a technology accelerator and a technology test bed. Kevin has a Bachelor of Science Honors degree in Ecology from the University of Calgary and a Masters of Applied Science in Environmental (Civil) Engineering, specializing in wastewater treatment, from the University of British Columbia.

Xiao-Li (Lilly) Pang obtained her Bachelor of Medicine (M.D.) from the University of Southeast, Nanjing, China, in 1982 and her Ph.D. in Medical Virology from the University of Tampere, Finland, in 2000. She had been a visiting scholar at the Centers for Disease Control (CDC) and the National Institutes of Health (NIH) in the USA. She is a professor at the Department of Laboratory Medicine and Pathology, University of Alberta, and a molecular virologist and program leader in the Alberta Precision Laboratories. Her research focuses on the study of viral etiology, clinical diagnostics, molecular epidemiology, waterborne viral diseases, and environmental virology. Her interests cover a broad range of important human and environmental viruses with high epidemic potential. She has made significant contributions to Wastewater-Based Epidemiology Surveillance (WBS) on SARS-CoV-2 during the COVID-19 pandemic, both provincially and nationally.