Quantitative microbial risk assessment of microbial source tracking markers in recreational water contaminated with fresh untreated and secondary treated sewage

Environment International 117 (2018) 243–249

Contents lists available at ScienceDirect

Environment International journal homepage: www.elsevier.com/locate/envint

Quantitative microbial risk assessment of microbial source tracking markers in recreational water contaminated with fresh untreated and secondary treated sewage

T

⁎

Warish Ahmeda, , Kerry A. Hamiltonb, Aldo Lobosc, Bridie Hughesa, Christopher Staleyd, Michael J. Sadowskyd,e, Valerie J. Harwoodc a

CSIRO Land and Water, Ecosciences Precinct, 41 Boggo Road, QLD 4102, Australia Drexel University, 3141 Chestnut Street, Philadelphia, PA 19104, USA Department of Integrative Biology, SCA 110, University of South Florida, 4202 East Fowler Ave, Tampa, FL 33620, USA d BioTechnology Institute, University of Minnesota, 1479 Gortner Ave, St. Paul, MN 55108, USA e Department of Soil, Water and Climate, 1991 Upper Buford Circle, Room 439, Saint Paul, MN 55108, USA b c

A R T I C LE I N FO

A B S T R A C T

Keywords: Sewage contamination MST QMRA Recreational water Health risk

Microbial source tracking (MST) methods have provided the means to identify sewage contamination in recreational waters, but the risk associated with elevated levels of MST targets such as sewage-associated Bacteroides HF183 and other markers is uncertain. Quantitative microbial risk assessment (QMRA) modeling allows interpretation of MST data in the context of the risk of gastrointestinal (GI) illness caused by exposure to known reference pathogens. In this study, five sewage-associated, quantitative PCR (qPCR) MST markers [Bacteroides HF183 (HF183), Methanobrevibacter smithii nifH (nifH), human adenovirus (HAdV), human polyomavirus (HPyV) and pepper mild mottle virus (PMMoV)] were evaluated to determine at what concentration these nucleic acid markers reflected a significant health risk from exposure to fresh untreated or secondary treated sewage in beach water. The QMRA models were evaluated for a target probability of illness of 36 GI illnesses/1000 swimming events (i.e., risk benchmark 0.036) for the reference pathogens norovirus (NoV) and human adenovirus 40/41 (HAdV 40/41). Sewage markers at several dilutions exceeded the risk benchmark for reference pathogens NoV and HAdV 40/41. HF183 concentrations 3.22 × 103 (for both NoV and HAdV 40/41) gene copies (GC)/100 mL of water contaminated with fresh untreated sewage represented risk > 0.036. Similarly, HF183 concentrations 3.66 × 103 (for NoV and HAdV 40/41) GC/100 mL of water contaminated with secondary treated sewage represented risk > 0.036. HAdV concentration as low as 4.11 × 101 GC/100 mL of water represented risk > 0.036 when water was contaminated with secondary treated sewage. Results of this study provide a valuable context for water quality managers to evaluate human health risks associated with contamination from fresh sewage. The approach described here may also be useful in the future for evaluating health risks from contamination with aged or treated sewage or feces from other animal sources as more data are made available.

1. Introduction Humans may be exposed to a variety of disease-causing microorganisms in recreational waters. A meta-analysis found that the risk of contracting diarrhea for non-swimmers was 35/1000, which rose to 59/ 1000 after swimming at beaches where the fecal indicator bacteria (FIB) Enterococcus spp. exceeded 35 CFU/100 mL of water (Arnold et al., 2016). Fecal pathogens can cause diarrhea, abdominal pain, cramping, nausea, and vomiting in healthy humans. Among sources of fecal contamination, human feces, including untreated or improperly ⁎

treated sewage and septage, generally represents a greater risk to human health than contamination from animal feces, largely due to the presence and greater abundance of human-specific enteric viruses (Soller et al., 2010a). Furthermore, enteric viruses have low median infectious doses, resulting in the potential for substantial health risks in healthy populations (Haas et al., 1993). However, fecal contamination from some animals, such as cattle, may also pose a significant human health risk (Soller et al., 2010a). Since sewage contamination poses greater risks to recreational water users than does animal fecal pollution, it is important to identify

Corresponding author at: Ecosciences Precinct, 41 Boggo Road, Dutton Park 4102, QLD, Australia. E-mail address: [email protected] (W. Ahmed).

https://doi.org/10.1016/j.envint.2018.05.012 Received 4 April 2018; Received in revised form 5 May 2018; Accepted 5 May 2018 0160-4120/ © 2018 Elsevier Ltd. All rights reserved.

Environment International 117 (2018) 243–249

W. Ahmed et al.

five sewage-associated MST markers targeting HF183, nifH, HAdV, HPyV and PMMoV in beach water samples seeded with fresh untreated and secondary treated sewage samples in Australia. Subsequently, concentrations of sewage markers that would pose a human health risk were estimated. The current study will enhance our understanding of the interpretation of sewage marker concentration data in the context of risk identification and mitigation.

the sources of such contaminants in recreational water so that appropriate mitigation strategies can be implemented, and risks can be accurately assessed (Soller et al., 2010a). However, based on concentrations of FIB, it is not possible to attribute the sources to human and specific animal hosts because FIB are excreted in the feces of all warmblooded animals. To overcome this limitation, researchers have developed a range of quantitative PCR (qPCR) assays to quantify host-associated molecular markers in environmental waters. These methods and others have been referred to as microbial source tracking (MST) tools. Sewage-associated molecular markers in current use for MST are Bacteroides HF183 (Green et al., 2014), BacHum-UCD (Kildare et al., 2007), HumM2 (Shanks et al., 2009), Methanobrevibacter smithii nifH (Johnston et al., 2010; Ahmed et al., 2012), human adenovirus A-F (HAdV; designated as a marker in this study) (Rusiñol et al., 2014), human polyomaviruses (HPyV) (McQuaig et al., 2009; Ahmed et al., 2010) and pepper mild mottle viruses (PMMoV) (Rosario et al., 2009). While the presence and concentrations of these markers in a water body provide information regarding potential sources, and perhaps the magnitude of the issue (i.e., high or low levels of contamination), it is difficult to interpret qPCR-generated marker concentration data in terms of human health risks (Boehm et al., 2015; Wang et al., 2013). Staley et al. (2012) linked the HF183 and HPyV markers to human health risk in recreational waters by determining the process limit of detection (PLOD) in dilutions of fresh untreated sewage and estimating the risk of gastroenteritis from exposure to NoV in the same diluted sewage. Both the HF183 and HPyV markers were detectable in water samples from several sites containing diluted sewage, and the probability of illness (Pill) exceeded a 10/1000 benchmark, based on a quantitative microbial risk assessment (QMRA) analysis. HF183 was also detectable in water samples containing diluted sewage where human health risk was below the 10/1000 benchmark, suggesting a low concentration of this marker may not pose a meaningful risk to recreational water users. The HPyV marker, which is less concentrated in fresh untreated sewage than is HF183, was not detected in diluted water samples seeded with sewage compared to the HF183, leading the authors to conclude that the HPyV marker may not provide adequate public health protection when used as the sole indicator of sewage contamination (Staley et al., 2012). A similar QMRA approach was undertaken in a recent study for the PMMoV marker in Florida (Symonds et al., 2016). The risk benchmark was set to 0.036 (36 GI illnesses/1000 people). Based on the PMMoV limit of detection, it was found that at a dilution of 10−2, the median GI illness risk was > 0.036. A recent study also used a QMRA approach to simulate the risk of GI illness associated with swimming in waters containing different concentrations of HF183 and HumM2 markers (Boehm et al., 2015). The volume/volume ratio of untreated sewage to ambient water was determined by comparing marker concentrations in recreational water to untreated sewage across the USA. The results indicated that median concentrations of 4.20 × 103 and 2.80 × 103 GC of HF183 and HumM2 markers/100 mL, respectively, in recreational water will surpass a risk benchmark of 30 GI illnesses/1000 swimmers/swimming event. The study by Boehm et al. (2015) is the first to establish a risk-based approach for interpreting concentrations of HF183 and HumM2 markers in ambient waters. This is particularly important because the concentrations of sewage-associated markers can be translated into a health risk. However, such data (i.e., marker concentration threshold) are not available for other sewage-associated markers such as nifH, HAdV, HPyV, and PMMoV. Furthermore, these QMRA estimates were done in the USA, and limited information is available from other countries where these markers are frequently used to detect sewage contamination of waterways. In addition, these studies used fresh untreated sewage as analytical material, thus the GI risk for a scenario where the water is contaminated with secondary treated sewage is not known. The main objective of this study was to undertake an exploratory QMRA analysis based on the process limit of quantification (PLOQ) of

2. Materials and methods 2.1. Determination of PLOQ for sewage-associated MST markers A previous study determined the PLOQ values of six sewage-associated markers targeting Escherichia coli (EC) H8, Bacteroides HF183, M. smithii nifH, HAdV, HPyV and PMMoV (Hughes et al., 2017). For this study, the Bacteroides HF183, M. smithii nifH, HAdV, HPyV and PMMoV markers were chosen, while EC H8 was omitted. The latter, EC H8 marker, and identical sequences can be found in genera like Yersinia and Klebsiella, thereby compromising host-specificity (Ahmed et al., 2015). Briefly, to determine the PLOQ of the sewage-associated nucleic acid markers (HF183, nifH, HAdV, HPyV and PMMoV), 3 mL of untreated sewage was seeded into 297 mL of filter-sterilized beach water samples (at a salinity of 34‰), in triplicate (Hughes et al., 2017). Similarly, 150 mL of secondary treated sewage was seeded into 150 mL of the filtered beach water samples, in triplicate. Tenfold serial dilutions (10−1 to 10−6) of all sewage-seeded samples were then made prior to analyses. The samples were filtered through negatively charged 47-mm, 0.45-μm-pore-size HA membranes (Merck Millipore, Tokyo, Japan). Nucleic acid samples were extracted directly from the membranes using the Mo Bio PowerWater DNA and RNA isolation kits (Mo Bio Laboratories, Carlsbad, CA, USA). The qPCR analyses including primers and probes, chemistry and cycling parameters have been described elsewhere in detail (Hughes et al., 2017). During qPCR analysis, all DNA and RNA samples were run in triplicate with three negative controls (sterile water) on 96-well plates using the CFX 96 thermocycler (BioRad Laboratories, CA, USA) (Hughes et al., 2017). 2.2. Process limit of quantification (PLOQ) Defined criteria were established to determine the PLOQ. The PLOQ was defined as the smallest volume of sewage that could be subjected to the complete sample preparation process, including dilution, filtration, and nucleic acid extraction and still be reliably quantified in 2/3 qPCR reactions (Staley et al., 2012; Symonds et al., 2016). 2.3. Quantitative microbial risk assessment NoV and HAdV 40/41 were selected as reference pathogens as these viruses are known to cause swimming-associated illnesses in recreational waters (Zlot et al., 2015; Kauppinen et al., 2017). Use of reference pathogens is an accepted practice in the field of QMRA (Soller et al., 2006; Soller et al., 2010a; Schoen et al., 2017). Furthermore, information such as concentrations of NoV and HAdV in sewage (Pina et al., 1998; Eftim et al., 2017; Hughes et al., 2017), and dose-response models exist in the literature (Crabtree et al., 1997; Teunis et al., 2010; Teunis et al., 2016) making it possible to assess human health risks. The concentrations of NoV and HAdV 40/41 in untreated sewage and secondary treated sewage were obtained from our previous study (Hughes et al., 2017). NoV concentrations data in untreated sewage produced a mean and standard deviation of 9.66 × 105 ± 2.85 × 105 gene copies (GC)/L. A lognormal distribution with parameters (μ = 13.7, σ = 0.28) was used calculate the dose (Table 1). The HAdV 40/41 concentration data in untreated sewage produced a mean and standard deviation of 9.66 × 105 ± 2.85 × 105 gene copies (GC)/L. A lognormal distribution with parameters (μ = 14.7, σ = 0.67) was used to calculate the dose. 244

Environment International 117 (2018) 243–249

W. Ahmed et al.

Table 1 Parameter values used in the QMRA analysis. Reference pathogens

Log normal concentrations/ L Untreated sewage

Secondary treated sewage

NoV

μ = 13.7; σ = 0.28

μ = 8.97; σ = 0.62

HAdV 40/ 41

μ = 14.7; σ = 0.67

μ = 8.60; σ = 0.67

Parameters

Units

Constant morbidity

α = 0.04 and β = 0.055 α = 5.24 and β = 2.95

GC

0.6

TCID50

0.5

μ = dose; GC = gene copies.

The NoV concentration in secondary treated sewage was 9.62 × 103 ± 6.71 × 103, GC/L, corresponding to a lognormal distribution with parameters (μ = 8.97, σ = 0.62). The HAdV 40/41 concentration in secondary treated sewage was 6.85 × 103 ± 5.16 × 103 GC/L, corresponding to a lognormal distribution with parameters (μ = 8.60, σ = 0.67). The volume (V) of water (in mL) ingested during swimming is based on a study of pool swimmers (Dufour et al., 2006). A lognormal distribution with parameters (μ = 2.92; σ = 1.43) was used to model the volume of water ingested. This distribution corresponded to a mean density of 18.6 mL and standard deviation 4.2 mL. The NoV dose (DNoV) for each sewage dilution was estimated using Eq. (1).

DNoV = (CNoV,untreated sewage or secondary treated sewage) × V × DF

(1) Fig. 1. Relationship of the probability of gastrointestinal (GI) illness caused by NoV in fresh untreated and secondary treated sewage and the process limit of quantification (PLOQ) in diluted sewage for (a) HF183 (b) nifH, (c) HAdV, (d) HPyV and (e) PMMoV. PLOQ data for HPyV (d) in secondary treated sewage samples were not available for this study. Minimum and maximum on the bars show the 25th and 75th percentile ranges for illness probability, the line in each bar represents the median probability of GI illness from NoV. The illness benchmark is marked as the level of risk where 36 cases result in GI illness out of 1000 exposures to NoV.

where (CNoV, untreated sewage or secondary treated sewage) is the untreated sewage or secondary treated sewage concentration/L, V is the ingestion volume in mL and DF is the dilution factor. The hypergeometric doseresponse model was used to estimate the probability of infection (Pinf). Parameters α = 0.04 and β = 0.055 for DNoV were measured using qPCR analysis for non-aggregated virus suspensions (Teunis et al., 2010). Lastly, the probability of illness (Pill) was estimated using Eq. (2).

Pill = Pinf × 0.6

(2)

3. Results

where Pinf is the probability of infection and 0.6 is the constant morbidity (fraction of infections resulting in illness) (Soller et al., 2010a). The HAdV 40/41 dose (DHAdV 40/41) for each sewage dilution was estimated using Eq. (3).

3.1. PLOQ of nucleic acid targets in sewage seeded, serially-diluted samples Supplementary Fig. S1 shows the PLOQ and corresponding concentration of each sewage marker in serially-diluted beach water samples seeded with (A) untreated (B) or secondary treated sewage (Hughes et al., 2017). The PLOQ is represented by the dilution at which sewage-associated markers were still quantifiable.

DHAdV 40/41 = (CHAdV 40/41,untreated sewage or secondary treated sewage) × V × DF (3) where (C HAdV 40/41, untreated sewage or secondary treated sewage) is the untreated or secondary treated sewage concentration/L, V is the ingestion volume in L and DF is the dilution factor. The exponential dose-response model with parameters α = 5.24 and β = 2.95 was used to obtain Pinf (Teunis et al., 2016). Finally, Pill was estimated using Eq. (4).

Pill = Pinf × 0.5

3.2. Probability of GI illness from NoV and HAdV 40/41 associated with sewage dilutions and MST marker presence The probability of GI illness due to NoV infection from recreational contact with waters contaminated with untreated and secondary treated sewage (except HPyV) was estimated for HF183, nifH, HAdV, HPyV and PMMoV for beach water samples (Fig. 1). The Pill is high when it is greater than the illness benchmark (i.e., 36 GI illnesses/1000 exposures or 0.036) and is low when it is less than < 0.036, as specified by US EPA (US EPA, 2012). The median Pill associated with the HF183 (denoted as “a” in Fig. 1) PLOQ at the dilution 10−5 for beach water sample seeded with untreated sewage was 0.065 which is approximately two times greater than the 0.036 US EPA GI illness benchmark. HPyV (denoted as “d”) and PMMoV (denoted as “e”) PLOQ at the dilution 10−3 for beach water samples seeded with untreated sewage was 0.286 which is approximately eight times greater than the 0.036 US

(4)

where Pinf is the probability of infection and 0.5 is the morbidity ratio (fraction of infections resulting in illness) (Haas et al., 1993). A Monte Carlo simulation was conducted to capture the variability of NoV and HAdV 40/41 concentrations, and volume of water ingested with 100,000 iterations in R version 3.3.3. The median, first, and third quartiles of the Pill for each qPCR marker concentration were calculated from the Monte Carlo iterations.

245

Environment International 117 (2018) 243–249

W. Ahmed et al.

benchmark. At dilution 10−1, the median Pill associated with the nifH and HAdV markers was 0.280, which is approximately eight-fold greater than the GI illness benchmark. The probability of GI illness due to HAdV 40/41 infection from exposure to waters seeded with fresh untreated and secondary treated sewage (except HPyV) was also estimated (Fig. 2). The median Pill associated with the HF183 (denoted as “a” in Fig. 2) at dilution 10−6 for beach water samples seeded with fresh untreated sewage was 0.032, which is slightly lower than the US EPA GI illness benchmark (i.e., 36 GI illnesses/1000 exposures). The median Pill associated with the HPyV (denoted as “d”) and PMMoV (denoted as “e”) at dilution 10−3 was 0.489, approximately 14-times greater than the GI illness benchmark. At the 10−2 dilution, the median Pill associated with nifH and HAdV markers exceeded the GI illness benchmark. The median Pill associated with the HF183 and PMMoV PLOQ at dilution 10−3 for beach water samples seeded with secondary treated sewage was 0.069, two times greater than the benchmark value. Below this dilution, HF183 and PMMoV were quantifiable, but the median Pill did not exceed the GI Illness benchmark. At dilution 10−1, the median Pill associated with the nifH and HAdV markers was 0.443, which is much greater than the prescribed GI illness benchmark. 3.3. Sewage marker concentration threshold corresponding to GI illness benchmark The water sampling protocol used in the study involved filtration of 300 mL of water sample to capture sewage-associated markers on the membrane, followed by nucleic acid extraction with Mo Bio Power Water DNA and RNA Isolation Kits (Hughes et al., 2017). This allowed us to obtain a final volume of 100 μL of nucleic acid from 300 mL of water sample seeded with either fresh untreated sewage or secondary treated sewage. A 3 μL template of 100 μL nucleic acid was subjected to qPCR analysis. Therefore, it was possible to estimate the concentration of the sewage markers/100 mL water sample by back-calculation corresponding to a median GI illness benchmark of 0.036. Among all the markers tested, HF183 concentration needed to be greater in the water sample than other markers to correspond to an exceedance of the illness benchmark (Table 2). When NoV and HAdV 40/41 were used as reference pathogens, a median concentration of 3.22 × 103 GC of the HF183 in 100 mL of water sample represented a risk above the GI illness benchmark value when beach water was contaminated with fresh untreated sewage. Similarly, 3.66 × 103 GC of HF183 in 100 mL of water sample represented a risk above the benchmark for both NoV and HAdV 40/41, when beach water was contaminated with secondary treated sewage. Concentrations of nifH and HAdV as low as 1.44 × 102 and 4.11 × 101 GC/100 mL, respectively exceeded the benchmark risk of GI illness when beach water was contaminated with either fresh untreated or secondary treated sewage. For both NoV and HAdV, a HPyV median concentration of 1.01 × 103 GC (water contaminated with untreated sewage) in 100 mL of water represented risk above the benchmark value. For fresh untreated sewage contaminated recreational water 5.44 × 102 GC in 100 mL of

Fig. 2. Relationship of the probability of gastrointestinal (GI) illness caused by HAdV 40/41 in fresh untreated and secondary treated sewage and the process limit of quantification (PLOQ) in diluted sewage for (a) HF183 (b) nifH, (c) HAdV, (d) HPyV and (e) PMMoV. PLOQ data for HPyV (d) in secondary treated sewage samples were not available for this study. Minimum and maximum on the bars show the 25th and 75th percentile ranges for illness probability, the line in each bar represents the median probability of GI illness from HAdV 40/ 41. The illness benchmark is marked as the level of risk where 36 cases result in GI illness out of 1000 exposures to HAdV 40/41.

EPA GI illness benchmark. Below this dilution, only HF183 was quantifiable up to 10−6, but the median Pill did not exceed the GI illness benchmark at the dilution 10−6. The median Pill associated with the nifH (denoted as “b”) and HAdV (denoted as “c”) at dilution 10−2 was 0.321 which is approximately nine times greater than the GI illness benchmark. The median Pill associated with the HF183 and PMMoV PLOQ at dilution 10−3 for beach water samples seeded with secondary treated sewage was 0.062 which is greater than the 0.036 US EPA GI illness benchmark. Below this dilution, HF183 and PMMoV (10−4) were quantifiable, but the median Pill did not exceed the GI illness

Table 2 Concentrations (GC) of sewage markers in 100 mL of beach water sample contaminated with either fresh untreated or secondary treated sewage that exceeded the median illness rate of 36/1000 people at a single event for reference pathogens NoV and HAdV 40/41. MST markers

Fresh untreated sewage

Fresh secondary treated sewage

NoV HF183/100 mL nifH/100 mL HAdV/100 mL HPyV/100 mL PMMoV/100 mL

HAdV 40/41 3

3.22 × 10 1.44 × 102 1.22 × 102 1.01 × 103 5.44 × 102

± ± ± ± ±

3

2.33 × 10 1.33 × 102 9.99 × 101 1.55 × 102 2.66 × 102

3

3.22 × 10 1.44 × 102 1.22 × 102 1.01 × 103 5.44 × 102

± ± ± ± ±

NoV 3

2.33 × 10 1.33 × 102 9.99 × 101 1.55 × 102 2.66 × 102

ND: Not determined. 246

HAdV 40/41 3

3.66 × 10 3.89 × 102 4.11 × 101 ND 2.00 × 103

3

± 3.89 × 10 ± 1.03 × 102 ± 3.33 × 101 ± 4.22 × 102

3.66 × 103 3.89 × 102 4.11 × 101 ND 1.78 × 102

± 3.89 × 103 ± 1.03 × 102 ± 3.33 × 101 ± 1.67 × 102

Environment International 117 (2018) 243–249

W. Ahmed et al.

greater compared to secondary treated sewage. This is because the concentrations of reference pathogens are generally 2–3 orders of magnitude greater in untreated sewage compared to secondary treated sewage. At lower dilutions, the probability of GI illness exceeded the 36/1000 benchmark for both reference pathogens, and all sewage markers were quantifiable. However, certain markers, such as HF183, HPyV, and PMMoV, were quantifiable at greater dilutions compared to nifH and HAdV, suggesting results should be interpreted with caution depending on the marker used. The PLOQ and corresponding GI risks obtained in this study also need to be interpreted with caution as the sample processing strategy and nucleic acid extraction protocols may vary from study to study. The QMRA approach used in this study assumed that sewage was the only source of pathogens in recreational waterways. However, pathogens of zoonotic origin can also enter the waterbody and could pose a risk to swimmers. Since qPCR assays were used to determine the marker concentrations in each dilution, we were able to calculate the concentration of markers in 100 mL of water that represented a risk above the GI illness benchmark (i.e., 36 cases of GI illness/1000 swimming events). This is particularly important for water quality regulators as they will be able to use the qPCR generated marker concentrations data to interpret the GI risk associated with fresh untreated sewage contamination in the event of a sewage spill or a significant stormwater runoff. Two previous studies did not provide the concentration benchmark values for the HF183, HPyV or PMMoV markers as they used PLOD (Staley et al., 2012; Symonds et al., 2016). However, in another study the authors reported that a median concentration of 4.2 × 103 GC of the HF183 marker in 100 mL of water sample represented a risk above the GI illness benchmark of 0.30 based on a QMRA simulation using NoV as a reference pathogen (Boehm et al., 2015). This value was generated from a linear relationship between log10-transformed HF183 median concentrations per 100 mL and simulated log10 GI risk. The value reported here (3.22 × 103 GC/100 mL) for the HF183 marker using NoV as a reference pathogen is slightly lower than the value reported by Boehm et al. (2015). Untreated sewage used for the PLOQ analysis in this study contained a mean concentration of 6.15 × 106 GC/mL of HF183 markers, which was > 1 order of magnitude greater than the concentration value (1.63 × 105 GC/mL obtained from Shanks et al., 2010) used in Boehm et al. (2015). While it is logical that the concentration of the sewage-associated markers in sewage will affect QMRA model outputs, more quantitative data on the spatial and temporal variability of sewage-associated markers and reference pathogens in fresh raw and secondary treated sewage samples will be required for improved risk assessment. In this study, several assumptions were made for the risk calculation. For example, the model did not consider any marker or pathogen decay in the environment. The decay rates of markers especially HF183 or nifH, can be faster than HAdV (Ahmed et al., 2014). If enteric viruses persist longer in the environment and remain infective, the GI risk estimates generated by HF183 and nifH markers need to be interpreted carefully. However, this study determined the risk associated with fresh sewage scenarios, and therefore, decay rates consideration into the models may not be applicable. One of the assumptions of the QMRA estimates was the ratios of sewage-associated markers, and reference pathogens were constant in diluted recreational water. These ratios may change over time under real-world conditions, as well as vary by treatment facility, when these markers and pathogens are transported from their source to the recreational sites. It was also assumed that all NoV and HAdV 40/41 quantified using qPCR are viable and infective. One notable limitation is that NoV cannot be cultured and available concentration data were generated using reverse-transcriptase (RT) qPCR. In this study, we did not consider MST marker recovery efficiencies from water samples to account for their effective or corrected concentrations in water samples. These need to be considered in the future studies as all these factors may affect

water represented risk above 0.36. 4. Discussion Sewage can enter recreational waters from defective sewer collection systems, treatment plant failures, poorly functioning on-site septic systems, extreme weather events, such as flooding, or by the direct release by treatment plants. It has been suggested that sewage contamination represents the dominant risk of GI illness in recreational water (Schoen and Ashbolt, 2010). Direct monitoring of pathogens in recreational water, such as enteric viruses is possible, but their uneven distribution in the community and low concentrations in sewage make the probability of false-negative results (failure to detect when health risk is elevated) high (Petterson et al., 2001). Alternatively, monitoring sewage-associated markers, which co-occur with pathogens in sewage and water can be an attractive option because their concentrations are 3 to 4 orders of magnitude greater than enteric viruses, providing a more sensitive level of detection of sewage in environmental waters (Hughes et al., 2017). Sewage-associated markers are useful for specifically detecting sewage contamination, but limited information is available on the concentrations of most commonly used sewage-associated markers in water and associated human health risk. The QMRA analysis undertaken here was intended to establish a link between concentrations of five sewage-associated markers and risk of GI illness from primary recreational water use (i.e. swimming). The QMRA approach adopted in this study was previously used for the HF183, HPyV and PMMoV markers in fresh untreated sewage in the USA based on PLOD analysis (Staley et al., 2012; Symonds et al., 2016). This study expanded on the previous analyses by modeling risk from exposure to water contaminated with fresh secondary treated sewage in addition to untreated sewage using PLOQ analysis. Fresh treated effluent (i.e., UV and/or chlorinated) was not included in this study as the concentrations of sewage-associated markers can be low in the final effluent making it difficult to detect (McQuaig et al., 2009). Furthermore, data on the concentrations of enteric viruses (i.e., NoV and HAdV 40/41) in treated effluent is limited. Previous studies (Staley et al., 2012; Symonds et al., 2016) used NoV as a reference pathogen to estimate the GI risk, however, in this study, HAdV 40/41 was also used as an additional reference pathogen to obtain risk estimates. In this study, we used hypergeometric dose-response models for both NoV and HAdV 40/41. This dose-response model is the most commonly used to estimate the risk of NoV in a variety of water matrices (Van Abel et al., 2017). A limitation of this model is that it does not use an aggregation factor of NoV (Van Abel et al., 2017). While NoV in fresh sewage may not aggregate, we lack data and information on the degree of aggregation that may be accumulated during wastewater treatment processes (e.g., activated sludge processes) (McBride, 2014). Like this present study, most NoV QMRA studies in the literature have ignored the aggregation factor due to the complexity and lack of knowledge of virus aggregation in water matrices (McBride et al., 2013; Schoen and Ashbolt, 2010; Soller et al., 2010b; Viau et al., 2011). The hypergeometric dose-response model used in this study for HAdV 40/41 accounts for variation in infectivity and/or pathogenicity across few HAdV types for an oral ingestion scenario (Teunis et al., 2016). This model combines dose-response information from five published studies that used HAdV 4, 7 and 16 for oral ingestion, intranasal, inhalation and intraocular droplet inoculation. Therefore, this model was more suitable than the exponential dose-response model (based on respiratory HAdV 4), which has previously used in assessing the risk of human health of HAdV in several studies (Couch et al., 1969; Crabtree et al., 1997; van Heerden et al., 2005; Chigor et al., 2014; Lim et al., 2015). Risk estimates indicated that more diluted sewage posed less GI risk compared to less diluted sewage for both reference pathogens. This trend was consistent for both untreated and secondary treated sewage scenarios. The GI risk associated with the fresh untreated sewage was 247

Environment International 117 (2018) 243–249

W. Ahmed et al.

the risk estimate. The scenarios examined in the current QMRA can be used as the basis for examining more complex scenarios in the future. For example, it would likely be possible to estimate risk from contact with septage, recently excreted feces, treated sewage, as well as aged sewage when more data are available. Marker/pathogen inactivation data and transport time can also be incorporated into the model, and perhaps multiple dose-response models can be used in the future to provide a range of Pill. Different exposure scenarios other than swimming such as canoeing, kayaking, boating, and fishing can also be explored (Sunger and Haas, 2015). The risk was estimated for a single event, but risks could be annualized to take into account exposure frequency and duration patterns.

Boehm, A.B., Soller, J.A., Shanks, O.C., 2015. Human-associated fecal quantitative polymerase chain reaction measurements and simulated risk of gastrointestinal illness in recreational waters contaminated with untreated sewage. Environ. Sci. Technol. 2 (10), 270–275. Chigor, V.N., Sibanda, T., Okoh, A.I., 2014. Assessment of the risks for human health of adenoviruses, hepatitis a virus, rotaviruses and enteroviruses in the Buffalo River and three source water dams in the Eastern Cape. Food Environ. Virol. 6 (2), 87–98. Couch, R.B., Knight, V., Douglas Jr., R.G., Black, S.H., Hamory, B.H., 1969. The minimal infectious dose of adenovirus type 4; the case for natural transmission by viral aerosol. Trans. Am. Clin. Climatol. Assoc. 80, 205–211. Crabtree, K.D., Gerba, C.P., Rose, J.B., Haas, C.N., 1997. Waterborne adenovirus: A risk assessment. Water Sci. Technol. 35 (11−12), 1–6. Dufour, A.P., Evans, O., Behymer, T.D., Cantu, R., 2006. Water ingestion during swimming activities in a pool: a pilot study. J. Water Health 4 (4), 425–430. Eftim, S.E., Hong, T., Soller, J., Boehm, A., Warren, I., Ichida, A., Nappier, S.P., 2017. Occurrence of norovirus in raw sewage – a systemtaic literature review and metaanalysis. Water Res. 111, 366–374. Green, H.C., Haugland, R.A., Verma, M., Millen, H.T., Borchardt, M.A., Field, K.G., Walters, W.A., Knight, R., Sivaganesan, M., Kelty, C.A., Shanks, O.C., 2014. Improved HF183 quantitative real-time PCR assay for characterization of human fecal pollution in ambient surface water samples. Appl. Environ. Microbiol. 80 (10), 3086–3094. Haas, C.N., Rose, J.B., Gerba, C., Regli, S., 1993. Risk assessment of virus in drinking water. Risk Anal. 13 (5), 545–552. van Heerden, J., Ehlers, M.M., Vivier, J.C., Grabow, W.O., 2005. Risk assessment of adenoviruses detected in treated drinking water and recreational water. J. Appl. Microbiol. 99 (4), 926–933. Hughes, B., Beale, D.J., Dennis, P.G., Cook, S., Ahmed, W., 2017. Cross-comparison of human wastewater associated molecular markers in relation to fecal indicator bacetria and enteric viruses in recreational beach waters. Appl. Environ. Microbiol. 83 (8) (pii: e00028-17). Johnston, C., Ufnar, J.A., Griffith, J.A., Stewart, J.R., 2010. A real-time qPCR assay for the detection of the nifH gene of Methanobrevibacter smithii, a potential indicator of sewage pollution. J. Appl. Microbiol. 109 (6), 1946–1956. Kauppinen, A., Al-Hello, H., Zacheus, O., Kilponen, J., Maunula, L., Huusko, S., Lappalainen, M., Miettinen, I., Blomqvist, S., Rimhanen-Finne, R., 2017. Increase in outbreaks of gastroenteritis linked to bathing water in Finland in summer 2014. Euro Surveill. 22 (8), 30470. Kildare, B.J., Leutenegger, C.M., McSwain, B.S., Bambic, D.G., Rajal, V.B., Wuertz, S., 2007. 16S rRNA based assays for quantitative detection of universal, human-, cowand dog-specific fecal Bacteroidales: a Bayesian approach. Water Res. 41 (16), 3701–3715. Lim, K.Y., Hamilton, A.J., Jiang, S.C., 2015. Assessment of public health risk associated with viral contamination in harvested urban stormwater for domestic applications. Sci. Total Environ. 523, 95–108. McBride, G.B., 2014. Norovirus dose-response in sewage-related QMRA: the importance of virus aggregation. In: 9th International Congress on Environmental Modelling and Software. 69. McBride, G.B., Stott, R., Miller, W., Bambic, D., Wuertz, S., 2013. Discharge-based QMRA for estimation of public health risks from exposure to stormwater-borne pathogens in recreational waters in the United States. Water Res. 47 (14), 5282–5297. McQuaig, S.M., Scott, T.M., Lukasik, J.O., Paul, J.H., Harwood, V.J., 2009. Quantification of human polyomaviruses JC virus and BK virus by TaqMan quantitative PCR and comparison to other water quality indicators in water and fecal samples. Appl. Environ. Microbiol. 75 (11), 3379–3388. Petterson, S.R., Ashbolt, N.J., Sharma, A., 2001. Microbial risks from wastewater irrigation of salad crops: a screening-level risk assessment. Water Environ. Res. 73 (6), 667–672. Pina, S., Puig, M., Lucena, F., Jofre, J., Girones, R., 1998. Viral pollution in the environment and in shellfish: human adenovirus detection by PCR as an index of human viruses. Appl. Environ. Microbiol. 64 (9), 3376–3382. Rosario, K., Symonds, E., Sinigalliano, C., Stewart, J., Breitbart, M., 2009. Pepper mild mottle virus as an indicator of fecal pollution. Appl. Environ. Microbiol. 75 (22), 7261–7267. Rusiñol, M., Fernandez-Cassi, X., Hundesa, A., Vieira, C., Kern, A., Eriksson, I., Ziros, P., Kay, D., Miagostovich, M., Vargha, M., Allard, A., Vantarakis, A., Wyn-Jones, P., Bofill-Mas, S., Girones, R., 2014. Application of human and animal viral microbial source tracking tools in fresh and marine waters from five different geographical areas. Water Res. 59, 119–129. Schoen, M.E., Ashbolt, N.J., 2010. Assessing pathogen risk to swimmers at non-sewage impacted recreational beaches. Environ. Sci. Technol. 44 (7), 2286–2291. Schoen, M.E., Ashbolt, N.J., jahne, M.A., Garland, J., 2017. Risk-based enteric pathogen reduction targets for non-potable and direct potable use of roof runoff, stormwater and greywater. Microbial. Risk Anal. 5, 32–43. Shanks, O.C., Kelty, C.A., Sivaganesan, M., Varma, M., Haugland, R.A., 2009. Quantitative PCR for genetic markers of human fecal pollution. Appl. Environ. Microbiol. 75 (17), 5507–5513. Shanks, O.C., White, K., Kelty, C.A., Sivaganesan, M., Blannon, J., Meckes, M., Varma, M., Haugland, R.A., 2010. Performance of PCR-based assays targeting Bacteroidales genetic markers of human fecal pollution in sewage and fecal samples. Environ. Sci. Technol. 44 (16), 6281–6288. Soller, J.A., Eisenberg, J., DeGeorge, J., Cooper, R., Tchobanoglous, G., Olivieri, A., 2006. A public health evaluation of recreational water impairment. J. Water Health 4 (1), 1–19. Soller, J.A., Schoen, M.E., Bartrand, T., Ravenscroft, J.E., Ashbolt, N.J., 2010a. Estimated human health risks from exposure to recreational waters impacted by human and non-human sources of fecal contamination. Water Res. 44 (16), 4674–4691.

5. Conclusions

• In this study, we translated the levels of five sewage-associated •

•

•

molecular markers (i.e., HF183, HPyV, PMMoV, HAdV, and nifH) into human health risks by undertaking an exploratory QMRA analysis based on the marker concentration in recreational water. HF183 concentration of 3.2 × 103 GC in 100 mL of recreational water represented a significant health risk to swimmers for both NoV and HAdV 40/41 when water was contaminated with untreated sewage, while concentration of 3.66 × 103 GC/100 mL of water represented risks for both NoV and HAdV 40/41 when water was contaminated with secondary treated sewage. For both NoV and HAdV, HPyV median concentrations ranging from 1.01 × 103 GC (water contaminated with untreated sewage) in 100 mL of water represented risk above the benchmark value. HAdV concentration as little as 4.11 × 101 GC/100 mL of water represented risk > 0.036 when water was contaminated with fresh secondary treated sewage. The risk framework described in the present study can also be implemented for other types of water such, as drinking water or stormwater. A similar strategy also can be used to determine the risk associated with animal fecal markers. Given that target, various scenarios can be assessed to provide water managers with a greater understanding of situations that may impact GI risk due to recreational activity.

Acknowledgement We gratefully acknowledge Julius Career Award funding from CSIRO and the School of Agriculture and Food Sciences, The University of Queensland. We also thank the Minnesota Agricultural Experiment Station for funding to Michael J. Sadowsky. Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.envint.2018.05.012. References Ahmed, W., Wan, C., Goonetilleke, A., Gardner, T., 2010. Evaluating sewage-associated JCV and BKV polyomaviruses for sourcing human fecal pollution in a coastal river in Southeast Queensland, Australia. J. Environ. Qual. 39 (5), 1743–1750. Ahmed, W., Sidhu, J.P.S., Toze, S., 2012. Evaluation of the nifH gene marker of Methanobrevibacter smithii for the detection of sewage pollution in environmental waters in Southeast Queensland, Australia. Environ. Sci. Technol. 46 (1), 543–550. Ahmed, W., Gyawali, P., Sidhu, J.P.S., Toze, S., 2014. Relative inactivation of faecal indicator bacteria and sewage markers in freshwater and seawater microcosms. Lett. Appl. Microbiol. 59 (3), 348–354. Ahmed, W., Triplett, C., Gomi, R., Gyawali, P., Hodgers, L., Toze, S., 2015. Assessment of genetic markers for tracking the sources of human wastewater associated Escherichia coli in environmental waters. Environ. Sci. Technol. 49 (15), 9341–9346. Arnold, B.F., Wade, T.J., Benjamin-Chung, J., Schiff, K.C., Griffith, J.F., Dufour, A.P., Weisberg, S.B., Colford Jr., J.M., 2016. Acute gastroenteritis and recreational water: highest burden among young US children. Am. J. Public Health 106 (9), 1690–1697.

248

Environment International 117 (2018) 243–249

W. Ahmed et al.

3461–3473. US EPA, 2012. 820-F-12-058. In: Office of Water (Ed.), Recreational Water Quality Criteria, Washington, D.C. Van Abel, N., Schoen, M.E., Kissel, J.C., Meschke, J.S., 2017. Comparison of risk predicted by multiple norovirus dose-response models and implications for quantitative microbial risk assessment. Risk Anal. 37 (2), 245–264. Viau, E.J., Lee, D., Boehm, A.B., 2011. Swimmer risk of gastrointestinal illness from exposure to tropical coastal waters impacted by terrestrial dry-weather runoff. Environ. Sci. Technol. 45, 7158–7165. Wang, D., Farnleitner, A.H., Field, K.G., Green, H.C., Shanks, O.C., Boehm, A.B., 2013. Enterococcus and Escherichia coli fecal source apportionment with microbial source tracking genetic markers − is it feasible? Water Res. 47 (18), 6849–6861. Zlot, A., Simckes, M., Vines, J., Reynolds, L., Sullivan, A., Scott, M.K., McLuckie, J.M., Kromer, D., Hill, V.R., Yoder, J.S., Hlavsa, M.C., 2015. Norovirus outbreak associated with a natural lake used for recreation-Oregon, 2014. Am. J. Transplant. 15 (7), 2001–2005.

Soller, J.A., Bartrand, T., Ashbolt, N.J., Ravenscroft, J., Wade, T.J., 2010b. Estimating the primary etiologic agents in recreational freshwaters impacted by human sources of faecal contamination. Water Res. 44, 4736–4747. Staley, C., Gordon, K.V., Schoen, M.E., Harwood, V.J., 2012. Performance of two quantitative PCR methods for microbial source tracking of human sewage and implications for microbial risk assessment in recreational waters. Appl. Environ. Microbiol. 78 (20), 7317–7326. Sunger, N., Haas, C.N., 2015. Quantitative microbial risk assessment for recreational exposure to water bodies in Philadelphia. Water Environ. Res. 87 (3), 211–222. Symonds, E., Sinigalliano, C., Gidley, M., Ahmed, W., McQuaig-Ulrich, S., Breitbart, M., 2016. Faecal pollution along the southeastern coast of Florida and insight into the use of pepper mild mottle virus as an indicator. J. Appl. Microbiol. 121 (5), 1469–1481. Teunis, P.F., Xu, F., Fleming, K.K., Yang, J., Moe, C.L., Lechevallier, M.W., 2010. Enteric virus infection risk from intrusion of sewage into a drinking water distribution network. Environ. Sci. Technol. 44 (22), 8561–8566. Teunis, P., Schijven, J., Rutjes, S., 2016. A generalized dose-response relationship for adenovirus infection and illness by exposure pathway. Epidemiol. Infect. 144,

249