Human health risk from consumption of aquatic species in arsenic-contaminated shallow urban lakes

Abstract

Arsenic (As) causes cancer and non-cancer health effects in humans. Previous research revealed As concentrations over 200 μg g⁻¹ in lake sediments in the south-central Puget Sound region affected by the former ASARCO copper smelter in Ruston, WA, and significant bioaccumulation of As in plankton in shallow lakes. Enhanced uptake occurs during summertime stratification and near-bottom anoxia when As is mobilized from sediments. Periodic mixing events in shallow lakes allow dissolved As to mix into oxygenated waters and littoral zones where biota reside. We quantify As concentrations and associated health risks in human-consumed tissues of sunfish [pumpkinseed (Lepomis gibbosus) and bluegill (Lepomis macrochirus)], crayfish [signal (Pacifastacus leniusculus) and red swamp (Procambarus clarkii)], and snails [Chinese mystery (Bellamya chinensis)] from lakes representing a gradient of As contamination and differing mixing regimes. In three shallow lakes with a range of arsenic in profundal sediments (20 to 206 μg As g⁻¹), mean arsenic concentrations ranged from 2.9 to 46.4 μg g⁻¹ in snails, 2.6 to 13.9 μg g⁻¹ in crayfish, and 0.07 to 0.61 μg g⁻¹ in sunfish. Comparatively, organisms in the deep, contaminated lake (208 μg g⁻¹ in profundal sediments) averaged 11.8 μg g⁻¹ in snails and 0.06 μg g⁻¹ in sunfish. Using inorganic As concentrations, we calculated that consuming aquatic species from the most As-contaminated shallow lake resulted in 4–10 times greater health risks compared to the deep lake with the same arsenic concentrations in profundal sediments. We show that dynamics in shallow, polymictic lakes can result in greater As bioavailability compared to deeper, seasonally stratified lakes. Arsenic in oxygenated waters and littoral sediments was more indicative of exposure to aquatic species than profundal sediments, and therefore we recommend that sampling methods focus on these shallow zones to better indicate the potential for uptake into organisms and human health risk.

Graphical Abstract

1. Introduction

Metal and metalloid contaminants (hereafter referred to collectively as metals) are a persistent threat to environmental and human health. These legacy metals can be taken up directly by aquatic organisms through sediment contact or ingestion (Anderson et al., 1997), or through remobilization and absorption of dissolved forms (Azizur Rahman et al., 2012; Barrett et al., 2018). Metals are also readily transferred through food webs from primary producers to top consumers (Chen and Folt, 2000; Foust et al., 2016; Revenga et al., 2012). Aquatic organisms are vulnerable to the uptake of metal contaminants (Cui et al., 2020), leading to population declines (Erickson et al., 2011), compromised ecosystem functioning (Chen et al., 2015), and decreased quality of food available for human consumption (Liang et al., 2011).

Exposure to arsenic remains a prevalent human health issue across the world (Luvonga et al., 2020; Yoshida et al., 2004). Elevated arsenic levels have been reported in water, soils, and sediments as a result of both natural and anthropogenic sources. In Asia, many locations have groundwater with extremely high levels of naturally occurring arsenic (Fendorf et al., 2010), with concentrations and spatial distributions of arsenic sensitive to human activities, such as groundwater pumping (Harvey et al., 2006; Neumann et al., 2010). Arsenic is also a prevalent pollutant associated with industrial practices such as mining, wood processing, herbicide applications, and smelting, and thus anthropogenic sources of arsenic are widespread (Ferguson and Gavis, 1972; Rodie et al., 1995; Smedley and Kinniburgh, 2002).

Exposure to arsenic, either through ingestion of contaminated drinking water (Karagas et al., 2002) or food (Banerjee et al., 2013; Kar et al., 2011) can lead to adverse effects on human health. Arsenic is a carcinogen and neurotoxin and known to cause skin disorders including Blackfoot disease, hyperpigmentation, and skin lesions through chronic exposure (Kazi et al., 2009; Smith et al., 2000; Tseng, 1977; Vega et al., 2001). Long-term exposure to arsenic, even at low-levels has been related to increased rates of morbidity (Argos et al., 2010) and elevated incidence of diabetes (Bräuner et al., 2014), hypertension (Zhang et al., 2013), decreased male fertility (Wang et al., 2016), respiratory problems, cardiovascular diseases, kidney and bladder disorders, and complications of the gastrointestinal tract (Sinha and Prasad, 2020).

Lake sediments are known to sequester metal contaminants, often long past the date of initial deposition (Gawel et al., 2014). During the summer, lakes can thermally stratify, forming a cooler, stagnant bottom layer of water known as the hypolimnion. Inorganic arsenic can be mobilized from sediments during periods of anoxia in a lake’s hypolimnion resulting from aerobic respiration by bacteria (Aggett and Kriegman, 1988; Aggett and O’Brien, 1985; Martin and Pedersen, 2002). Dissolution of iron hydroxide solids in sediments and diffusion from porewater into the water column has been well described and shown to be the primary mechanism driving arsenic mobilization under anoxic conditions (Barrett et al., 2019; Smedley and Kinniburgh, 2002; Vitre et al., 1991). Temperature-mediated effects (Barrett et al., 2019; Weber et al., 2010) and endemic bacterial populations (Barral-Fraga et al., 2020) affect the rate at which this process occurs. Dissolved inorganic arsenic within the water column (namely arsenite and arsenate) can subsequently be taken up and bioaccumulated by microorganisms, such as phytoplankton and periphyton, thus entering the food web (Barrett et al., 2018; Caumette et al., 2011; Lopez et al., 2016; Wang et al., 2017). The trophic transfer of arsenic may be enhanced by periodic mixing events that bring remobilized arsenic into oxic waters (Barringer et al., 2011) where a majority of aquatic organisms reside, as occurs in shallow lakes (Barrett et al., 2018). Benthic organisms, including crayfish (Mason et al., 2000; Williams et al., 2009) and snails (Whaley-Martin et al., 2013), may be doubly exposed to arsenic as they spend the majority of their time in or on contaminated sediment in oxygenated waters.

Elevated levels of arsenic, attributed primarily to the former ASARCO copper smelter located in Ruston, WA, have been reported in the Puget Sound lowlands. The smelter operated for over a century before being demolished in the early 1990s. Daily operations of the smokestack produced a plume covering over a thousand square miles, with the fallout depositing arsenic and lead across the landscape (Glass, 2003). Recent remediation efforts near the former smelter have focused on the quantification of arsenic concentrations within, and removal of, contaminated terrestrial topsoil. By contrast, the amount and behavior of arsenic that was deposited into aquatic systems remains a critical knowledge gap (US EPA, 2016).

While still largely understudied, previous research found arsenic concentrations of over 200 μg g⁻¹ in surface sediments collected from lakes within the Puget Sound lowlands as a result of aerial deposition from the ASARCO smelter (Gawel et al., 2014). Significant bioaccumulation of arsenic in phytoplankton and zooplankton in shallow lakes has also been measured in this region (Barrett et al., 2018). What has not yet been addressed is how much arsenic is transferred into organisms of higher trophic positions in these lakes, and what risk, if any, there may be to humans through direct consumption of aquatic life.

Arsenic is known to biodiminish at increasing trophic levels (Chen and Folt, 2000), with several studies finding arsenic concentrations in food sourced from contaminated water bodies to be below human health standards (Lorenzana et al., 2009; Petkovšek et al., 2012; Williams et al., 2006). However, arsenic-contaminated shallow lakes may pose a particularly problematic combination of physical, chemical, and biotic factors that result in greater risk to human health. Especially in the Puget Sound region, where lakes are both easily accessible and heavily contaminated with arsenic, the risk could be much higher for people whose diets are supplemented from lakes. In addition, recent surveys estimate that Northwest tribal members, Asian Americans, and Pacific Islanders may consume 10–100 times more seafood than the average consumer in Washington State (Harper and Walker, 2015; Washington State Department of Ecology, 2013; US EPA, 2002).

Our goal is to identify lake characteristics that lead to greater human health risk from the harvest of aquatic species. We hypothesize that shallow lakes pose a greater risk to human health due to enhanced arsenic uptake by organisms when elevated arsenic concentrations are present in sediments. Therefore, we set the following objectives: first, demonstrate the physical and chemical differences that dictate the extent of arsenic bioavailability through years of monitoring results from five lakes representing a gradient of arsenic contamination in sediments and differing water column mixing regimes. Second, present total arsenic concentrations in human-consumed tissues of three freshwater organisms, represented by five species: sunfish [pumpkinseed (Lepomis gibbosus) and bluegill (Lepomis macrochirus)], crayfish [signal (Pacifastacus leniusculus) and red swamp (Procambarus clarkii)], and snails [Chinese mystery (Bellamya chinensis)]. Third, calculate the average percent of inorganic arsenic in these species. Finally, use inorganic arsenic concentrations in a risk assessment model to determine cancer risks and daily dosages for non-cancerous health risks for mean, 90^th, and 99^th percentile consumption rates of these organisms to reflect the average U.S. adult diet as well as demographic populations that consume aquatic species at higher rates. All lakes included in the study are residential, have public access, and are commonly used for recreation, including the capture of live organisms. The organisms chosen are known to be consumed by humans (US EPA, 2002), and account for the majority of biomass in the littoral zone of their respective lakes (Larson and Olden, 2013; Twardochleb and Olden, 2016).

2. Methods

2.1. Site Characterization

Four study lakes in the Puget Sound Lowland region (Figure 1, Table 1) were chosen to constitute a gradient of arsenic concentrations in profundal (the deepest zone in a lake below the range of significant light penetration, usually between the thermocline and the lakebed) sediments and differing lake mixing behaviors: polymictic with high levels of arsenic contamination (206 μg g⁻¹; Lake Killarney); polymictic with moderate levels of arsenic contamination (49 μg g⁻¹; Steel Lake); polymictic with low levels of arsenic contamination (20 μg g⁻¹; Bonney Lake); and strong seasonal stratification with high levels of arsenic contamination (208 μg g⁻¹; Angle Lake) (Gawel et al., 2014). Killarney, Steel, and Angle lakes are located within 25 km of the former ASARCO copper smelter in Ruston, WA, in the direction of prevailing winds (out of the S/SW), and within the predicted deposition field for smelter emissions (Gawel et al., 2014). Bonney Lake is located outside the field of deposition but in the same Puget Lowlands glacial till geological setting, with initial evidence that it contained similar aquatic organisms to the other study lakes (e.g. snails and sunfish). Therefore, Bonney Lake represents a shallow, polymictic reference site for this study. Bonney Lake was found to be void of crayfish, therefore a fifth lake (Pine Lake) was sampled to represent a crayfish population in a lake with low levels of littoral sediment arsenic contamination (9 μg g⁻¹) (Figure 1, Table 1). Pine Lake thus serves as a reference site for crayfish alone.

Figure 1.

Location of five study lakes in the Puget Sound lowlands relative to the former ASARCO smelter (Commencement Bay/Nearshore Tideflats Superfund Site). Deposition zone predicted to have elevated arsenic in soils due to smelter emissions is indicated by the shaded area (Area-Wide Soil Contamination Task Force, 2003).

Table 1.

Characteristics of lakes included in this study.

Lake	Site location (lat/long)		Lake area (km2)	Max depth (m)	Selection criteria: [As] in sediments; Mixing regime	Mean [As] in profundal sediments (μg g−1 dry wt) a	Mean [As] in littoral sediments (μg g−1 dry wt) b	Mean dissolved [As] in oxic water column (μg L−1) c
Bonney	47.189	−122.186	0.07	5.8	Low [As]; polymictic	20	10	BDLd
Steel	47.327	−122.304	0.19	7.3	Moderate [As]; polymictic	49	50	1.75
Killarney	47.285	−122.292	0.12	4.6	High [As]; polymictic	206	213	20.42
Pine e	47.587	−122.045	0.36	11.9	Low [As]; seasonally stratified	- -	9	- -
Angle	47.427	−122.287	0.42	15.8	High [As]; seasonally stratified	208	35	1.47

^aTop 9–10.5 cm (Gawel et al., 2014)

^bSurface dredge samples collected in 2018 & 2019

^cBased on measurements taken in June 2019

^dBelow detection limit

^ePine Lake was not sampled for water nor profundal sediments.

2.2. Sample Collection

Dissolved oxygen measurements, phytoplankton tows, and water samples for dissolved arsenic were collected from a boat at approximately the deepest point in each lake as described previously (Barrett et al., 2018). Dissolved oxygen was measured every half meter (In-Situ smarTROLL MP equipped with an optical Rugged Dissolved Oxygen (RDO) sensor). Phytoplankton samples were collected using a vertical net tow (20 μm mesh) from 1–2 m above the lakebed and pre-filtered through a 153 μm sieve to remove zooplankton, then captured on 5.0 μm polycarbonate filters and stored at −80 °C. Water samples were collected using a peristaltic pump, filtered (0.45 μm Geotech cartridge filter), and then acidified with 1% HNO₃ (v/v) in the laboratory and allowed to stand for 14 days prior to analysis for dissolved arsenic. Littoral sediment samples were collected in 2018 and 2019 from nearshore (1–2 m depth), and profundal sediments samples in 2004 and 2005 from the deepest point in each lake. Between 2 to 6 samples were collected at each site using a dredge (Wildco Petite Ponar Stainless Steel Grab) as described in Gawel et al. (2014). In the laboratory, samples were dried at 60 °C for 72 h and then homogenized using a porcelain mortar and pestle.

Snail (Bellamya chinensis), crayfish (Pacifastacus leniusculus; Procambarus clarkii), and sunfish (Lepomis gibbosus; L. macrochirus) sampling occurred primarily over four dates in June 2019 in the entire nearshore (< 2 m depth) area of each lake. Additional frozen crayfish samples from Killarney, Steel, and Pine Lakes collected in 2004, 2009, 2017, and 2018 were also analyzed for this study (Larson and Olden, 2013; J.D. Olden, unpublished data). Table S1 details species collected from each lake. All crayfish were captured using minnow traps set overnight and baited with dog kibble (Larson and Olden, 2016). Snails were collected by hand via snorkeling or with a dip net from a boat. Crayfish and snails were sedated by placing on ice in sealed plastic bags in the boat, euthanized by placing in a freezer, then later thawed and dissected for edible tissues (whole snail, excluding shell, and crayfish tail meat). Sunfish were collected using beach seining and kept alive in oxygenated lake water in coolers and transferred back to the laboratory within 3 h of collection. Fish were euthanized, and immediately measured and dissected for muscle tissue (fillet) samples.

2.3. Laboratory Analyses

Duplicate portions of each edible tissue of snail, crayfish, or fish specimen were either dried at 60 °C overnight or stored at −80 °C. Immediately prior to digestion for total arsenic, oven dried tissues were homogenized using a porcelain mortar and pestle and dried to constant mass. For arsenic speciation preparation, all samples stored at −80 °C were placed in a freeze dryer for 72 h and homogenized using a porcelain mortar and pestle.

Oven-dried sediment and animal tissues were digested using a microwave-assisted (CEM MARS 5) total digestion protocol (modified EPA method 3015a) using trace metal grade HNO₃ in pressurized digestion vessels. After digestion, sample solutions were diluted to 2% (v/v) HNO₃. Efficacy of the digestion procedure was verified using certified reference material NIST 2711a (Montana Soil II), and DOLT-5 (dogfish liver) which yielded a recovery of 92 ± 10% (n = 6) and 99 ± 9% (n = 11), respectively. Concentrations of total arsenic in water and digested sediment and animal tissue samples were determined by inductively-coupled plasma mass spectrometry (ICP-MS) on an Agilent 7900 with a 0.25 μg L⁻¹ limit of detection (LOD) for arsenic. Calibration was performed using certified multi-element standards. Analytical accuracy of the ICP-MS method was assessed using certified reference material NIST 1640a (trace elements in natural water), which had a recovery of 87 ± 6% (n = 10) for arsenic.

The speciation of arsenic in edible tissues of snails, crayfish, and fish from Steel and Killarney, the two lakes with the highest total arsenic in aquatic organisms, was determined at the Trace Element Analysis laboratory at Dartmouth College, following a dilute acid heat assisted extraction (Kubachka et al., 2012) performed at the University of Washington Tacoma. Organisms from Angle and Bonney Lakes were not analyzed for speciation due to low levels of total arsenic in tissues. At Dartmouth, an aliquot of the extract was diluted 1:1 with 200 mM NH₄CO₃ and analyzed for arsenic species by ion chromatography coupled to inductively coupled plasma mass spectrometry (IC-ICP-MS). A Dionex AS14 anion-exchange column was used with an NH₄CO₃ gradient to separate arsenobetaine, arsenocholine, arsenite, dimethylarsinic acid, monomethylarsonic acid, and arsenate (run time of 7 minutes, flow rate of 1 mL min⁻¹, and injection volume of 20 μL). The system was calibrated using primary standard solutions of the arsenic species (Sigma Aldrich, Spex certiprep) and NIST 2669a urine level II was used for quality control. The extraction efficiency (sum of species / total As by ICP-MS) was 36 ± 26% (n = 6) for snails, 127 ± 28% (n = 4) for crayfish, and 67 ± 27% (n = 6) for fish. Other studies report similar ranges for extraction efficiencies across different environmental materials using equivalent methods (Foster et al., 2007; Lorenzana et al., 2009; Taylor and Jackson, 2016).

Arsenic speciation in plankton was also determined at the Trace Element Analysis laboratory at Dartmouth College using methods based on Taylor and Jackson (2016). Plankton samples were freeze-dried and removed from polycarbonate filters into sample vials with 10% methanol. Samples were sonicated in a 30 °C bath for 2 h, then filtered (0.2 μm) to remove residual solids. Extraction solutions were analyzed on an Agilent LC1120 liquid chromatograph using an anion-exchange column (Hamilton PRP-X100) coupled to an Agilent 8900 triple quadrupole ICP-MS eluted with 20 mM (NH4)2CO3 (1.1 mL min⁻¹) with the column temperature held at 35 °C. Analytical accuracy was verified using a secondary standard with a recovery of 104 ± 2% (n = 3). The extraction efficiency was 9 ± 3% (n = 9).

2.4. Statistical Methods

A one-way ANOVA with post-hoc Tukey’s test was used to compare sample means for arsenic concentrations in animal tissues by lake and to compare means of arsenic in sampled fish species (Lepomis gibbosus and L. macrochirus) or crayfish species (Pacifastacus leniusculus and Procambarus clarkii). A comparison of total arsenic concentrations between the two crayfish species (both collected in Steel Lake) and between the two fish species (both collected in Bonney Lake; Table S1) sampled produced no significant difference (data not shown). Therefore, species are combined in all statistical analyses presented below. Samples below the ICP-MS limit of detection were assigned a value of 0.125 μg L⁻¹ (half of the LOD) (Wendelberger and Campbell, 1994).

2.5. Risk Assessment

Adult (18+ years of age) human health risks associated with consuming snail, crayfish, and fish tissues were assessed using a modified ATSDR model provided by the Washington State Department of Health (ATSDR, 2005). The model only considers inorganic arsenic, therefore average percentages of inorganic arsenic in organisms from Killarney and Steel Lakes were applied to the edible tissues of organisms from all lakes in this study. The model assumes daily exposure over 30 years, an average body weight of 70 kg, and a cancer potency factor of 5.7 mg kg⁻¹ dy⁻¹. The United States Environmental Protection Agency (US EPA) recommended reference dose for determining the potential for non-cancer health risks is 3×10⁻⁴ (μg g⁻¹ dy⁻¹) for inorganic arsenic (ATSDR, 2006), assuming daily exposure. Cancer risks were determined based on a 70-year lifetime with a cancer occurrence threshold of 10⁻⁵ (1 additional cancer occurrence in 100,000 people) applied to represent an increased lifetime cancer risk (US EPA, 2000). Mean consumption rates for freshwater/estuarine shellfish (3.27 g dy⁻¹) and finfish (4.23 g dy⁻¹), and 90^th (17.5 g dy⁻¹) and 99^th (142.4 g dy⁻¹) percentile consumption rates for the combination of freshwater/estuarine shellfish and finfish, were taken from the US EPA (2002, 2000). The 90^th percentile (average for sport anglers) and 99^th percentile (average for subsistence fishers) consumption rates are to be used as default for the general and subsistence fishing populations, respectively, when States and Tribes do not have local consumption data (US EPA, 2000). The 90^th percentile consumption rate is the default recommended by the US EPA because it is more protective of the majority of consumers, rather than the mean consumption rate which also includes non-consumers.

3. Results and Discussion

3.1. Physical and chemical lake conditions

Despite having similarly high concentrations of arsenic in profundal sediments (~200 μg As g⁻¹), six times more arsenic was measured in the littoral sediments of the shallowest lake, Lake Killarney (213 μg As g⁻¹) compared to the deepest lake, Angle Lake (35 μg As g⁻¹; Table 1). In lakes with steeper slopes, like Angle Lake, fine sediment tends to be transported from littoral zones to profundal zones where it is less likely to be resuspended (Blais and Kalff, 1995). When compared to the deeper Angle Lake, arsenic flux estimates reported in Barrett et al. (2019) suggest an enhanced renewal of arsenic from profundal sediments may occur in shallow lakes like Lake Killarney. As arsenic is released from sediment porewaters during periods of anoxia, it diffuses upward into the water column, is periodically mixed into oxygenated surface waters, spread horizontally into littoral areas, and is then redeposited to surface sediments throughout the lake. The mixing of arsenic from the anoxic bottom waters into oxygenated surface waters in shallow lakes is also evidenced by water column monitoring over multiple seasons from 2015 to 2019 (Barrett et al., 2019, 2018; Figure 2). In June 2019, when organisms were collected for this study, the mean dissolved arsenic concentration in oxic waters (> 1 mg O₂ L⁻¹) in Lake Killarney was more than ten times the amount in the surface waters in Angle Lake (Figure 2b; Table 1). Compared to deeper lakes, the polymictic nature of shallow lakes increases arsenic exposure to aquatic organisms that reside primarily in oxygenated waters. In Lake Killarney, arsenic concentrations up to 970 μg As g⁻¹ have been measured in phytoplankton and 80 μg As g⁻¹ in zooplankton, compared to just 150 μg As g⁻¹ and 25 μg As g⁻¹ in Angle Lake, respectively (Barrett et al., 2018). The combination of elevated arsenic concentrations in oxic waters and littoral sediments in shallow lakes results in greater arsenic exposure for aquatic life compared to deeper, seasonally stratified lakes. This process may more greatly affect organisms in the oxygenated littoral zone and feeding at lower trophic levels, such as snails and crayfish, due to arsenic’s tendency to biodiminish as it passes from one trophic layer to the next (Maeda et al., 1992).

Figure 2.

(a) Dissolved oxygen concentrations (mg L⁻¹) and (b) dissolved arsenic concentrations (μg L⁻¹) in filtered water samples from the water column of four study lakes in June 2019 (closed symbols from shallow polymictic lakes, open symbols from the deep, seasonally stratified lake). Waters with dissolved oxygen concentrations above 1 mg O₂ L⁻¹ (dotted line; a.) are considered oxic. All dissolved arsenic measurements from Bonney Lake are below the LOD, and therefore reported as 0.125 μg L⁻¹.

3.2. Arsenic concentrations and speciation in organisms

Average concentrations of total arsenic in snails (46.4 μg As g⁻¹) and crayfish tissues (13.9 μg As g⁻¹) from Lake Killarney were statistically higher (p < 0.05) than all other lakes in this study (Figures 3a and 3b). Arsenic concentrations in fish fillets were also statistically highest (p < 0.05) in Lake Killarney (0.61 μg As g⁻¹), followed by Steel Lake (0.44 μg As g⁻¹), with both lakes having statistically greater concentrations in fish fillets than Bonney and Angle (Figure 3c). It is notable that Bonney, the shallow reference lake, and Angle, the highly contaminated deep lake, have roughly the same low arsenic concentration in fish fillet (~0.06 μg As g⁻¹). Overall, each organism showed a similar trend of increasing arsenic accumulation in edible tissues with increasing littoral sediment concentrations (Figure 3). Littoral sediments were chosen as the independent variable since snails and crayfish spend most of their lives buried in or feeding along the littoral sediment surface (Jokinen, 1982; Larson and Olden, 2016), and thus littoral sediments are more indicative of potential arsenic exposure for these organisms than profundal sediments (Figure S1). Sunfish also create and lay their eggs in nests along the littoral lakebed (Zięba et al., 2018) and feed on snails, crustaceans, and aquatic macroinvertebrates in the littoral zone (Andraso, 2005; Berchtold et al., 2015). However, a similar trend of increasing arsenic concentrations in edible tissues is also seen when using mean dissolved arsenic concentrations in oxic waters as the independent variable (Figure S2). As the number of lakes in the study is relatively small (maximum of n = 4), we did not attempt to formally correlate environmental covariates with tissue arsenic concentrations. Several other studies show that arsenic concentrations in organisms from sites of known arsenic contamination are similar to levels found in snails from Killarney, Steel and Angle Lakes (Lai et al., 2012), and crayfish and sunfish from Killarney and Steel Lakes (Otter et al., 2012; Williams et al., 2009, 2006).

Figure 3.

Mean arsenic concentrations in (a) Chinese mystery snail (Bellamya chinensis) whole tissues, (b) crayfish tail meat, and (c) fish (Lepomis sp.) fillet compared to littoral sediments of the study lakes (closed symbols from shallow polymictic lakes, open symbols from deep, seasonally stratified lakes). Error bars represent one standard deviation from the mean. Different letter labels denote significant difference (p < 0.05) between lakes.

The speciation of arsenic was determined for edible tissues of snails, crayfish, and fish from Steel and Killarney (Table S2), the two lakes with the highest total arsenic in aquatic organisms. The dominant species of inorganic arsenic in all tissues was arsenite, As(III). While arsenate, As(V), is more bioavailable to primary producers at the base of the aquatic food web (Lopez et al., 2016), As(III) is generally thought to be more toxic to humans and animals (Hughes, 2006). Most of the organic arsenic in all edible tissues was present as arsenobetaine (AsB), which is relatively non-toxic to humans and animals (Popowich et al., 2017). Snails from the two lakes contained an average of 33% As(III), 25% As(V), and 30% AsB. Snails may contain a higher percentage of inorganic arsenic than crayfish and fish due to their unique feeding behavior of being able to filter phytoplankton from the water column (Olden et al., 2013), which accumulate mostly inorganic arsenic forms (mean 87% inorganic As; Figure 4). Crayfish averaged 17% As(III), 3% As(V), and 55% AsB. The two crayfish species examined here feed on a wide range of items including plants, detritus, and aquatic animals, and thus they feed at a higher trophic level than snails, on average (Olden et al., 2009). Fish contained 7% As(III), 0.2% As(V), and 75% AsB, on average. A higher percentage of organic arsenic, in particular AsB, can be expected with increasing trophic positioning because arsenic is biotransformed with each trophic transfer along the aquatic food chain, concluding with AsB as the final storage form of arsenic in fish tissues (Zhang et al., 2016).

Figure 4.

Percent concentration of the sum of inorganic (arsenate and arsenite) and organic (dimethylarsinic acid, monomethylarsonic acid, arsenobetaine, and arsenocholine) arsenic species in edible aquatic species tissues. For phytoplankton, the organic fraction is comprised of only dimethylarsinic acid, monomethylarsonic acid, and arsenosugar. Samples are from Killarney and Steel Lakes for each organism. Error bars represent one standard deviation from the mean.

For estimating potential human toxicity, multiple studies have shown that, relative to organic arsenic species, the amount of inorganic arsenic ingested is of more importance (Muñoz et al., 2000; US EPA, 1988). Therefore, only the total amount of inorganic arsenic was considered in the risk assessment. On average, snails contained the highest percentage of inorganic arsenic (58%) while the percentage in crayfish (20%) and sunfish (7%) was much lower (Figure 4). These percentages, which were measured in organisms from Killarney and Steel Lakes, were applied to edible tissues of organisms from the other study lakes (Bonney, Angle, and Pine Lakes) to determine the associated human health risk.

3.3. Risk Assessment

Edible tissues of organisms from the shallow arsenic-contaminated lakes (Killarney and Steel) were found to result in consistently higher health risks compared to deep Angle Lake, despite this lake having the same concentration of arsenic as Killarney in profundal sediments (~200 μg g⁻¹; Table 1). Arsenic concentrations in profundal sediments of Steel Lake (49 μg As g⁻¹) are only about a quarter of that in Angle Lake, but concentrations in littoral sediments in Steel (50 μg As g⁻¹) are higher than in Angle (35 μg As g⁻¹). Thus, we find that arsenic bioavailability is reflected in levels of arsenic in littoral sediments, a product of arsenic transfer into the oxic water column. Arsenic concentrations in littoral sediments and the oxic water column are highest in the contaminated shallow, polymictic lakes in this study (Table 1).

Using the mean consumption rate for the general U.S. adult population, the increased cancer risk from consumption of snails and crayfish from Lake Killarney and snails from Angle and Steel Lakes is greater than one additional cancer occurrence in 100,000 (10⁻⁵, US EPA, 2000; Table 2). The 90^th percentile consumption rate (the default used by the US EPA), resulted additionally in an increased cancer risk (>10⁻⁵) for snails from Bonney and crayfish from Steel, with the risk level for snails from Angle, Steel, and Killarney now exceeding 1 additional cancer occurrence in 10,000 individuals (10⁻⁴). For the 99^th percentile consumption rate, used to represent subsistence fishing populations, an increased cancer risk in excess of 10⁻⁵ was calculated for snails and crayfish from all study lakes and for fish consumption in Lake Killarney. The most concerning results are an increased cancer risk above 10⁻³ (1 additional cancer occurrence in 1,000 individuals) for 99^th percentile consumers eating snails from Angle, Steel, and Killarney, with the risk level in Killarney over 4 additional cancer occurrences in 1,000 (4.3×10⁻³).

Table 2.

Estimated non-cancer health risk (μg As g⁻¹ body wt dy⁻¹) and cancer risk from consumption of aquatic species in study lakes for U.S. adult (18+ years of age) consumers.

Aquatic organism	Lake	Mean inorganic As (μg kg−1 fresh wt)	Non-Cancer Health Risk (μg As g−1 body wt dy−1)			Cancer Risk
			Mean consumer a	90th percentile b	99th percentile c	Mean consumer a	90th percentile b	99th percentile c
Snail	Bonney	202.97	9.48E-06	5.07E-05	4.13E-04d	6.10E-06	3.26E-05e	2.65E-04e
Snail	Angle	834.43	3.90E-05	2.09E-04	1.70E-03d	2.51E-05e	1.34E-04e	1.09E-03e
Snail	Steel	1072.64	5.01E-05	2.68E-04	2.18E-03d	3.22E-05e	1.72E-04e	1.40E-03e
Snail	Killarney	3267.25	1.53E-04	8.17E-04d	6.65E-03d	9.81E-05e	5.25E-04e	4.27E-03e
Crayfish	Pine	37.06	1.73E-06	9.26E-06	7.54E-05	1.11E-06	5.96E-06	4.85E-05e
Crayfish	Steel	107.46	5.02E-06	2.69E-05	2.19E-04	3.23E-06	1.73E-05e	1.41E-04e
Crayfish	Killarney	571.88	2.67E-05	1.43E-04	1.16E-03d	1.72E-05e	9.19E-05e	7.48E-04e
Fish	Angle	0.85	5.14E-08	2.12E-07	1.73E-06	3.30E-08	1.37E-07	1.11E-06
Fish	Bonney	0.99	5.99E-08	2.48E-07	2.02E-06	3.85E-08	1.59E-07	1.30E-06
Fish	Steel	6.23	3.77E-07	1.56E-06	1.27E-05	2.42E-07	1.00E-06	8.15E-06
Fish	Killarney	8.64	5.22E-07	2.16E-06	1.76E-05	3.36E-07	1.39E-06	1.13E-05e

^aThe consumption rate for the mean U.S. adult is 3.27 g dy⁻¹ for freshwater/estuarine shellfish (snails and crayfish) and 4.23 g dy⁻¹ for finfish (US EPA, 2002).

^bThe 90th percentile consumption rate is 17.5 g dy⁻¹ and includes both shellfish and finfish (US EPA, 2000).

^cThe 99th percentile consumption rate is 142.4 g dy⁻¹ and includes both shellfish and finfish (US EPA, 2000).

^dGreater than the US EPA recommended RfD (non-cancer health risk criterion) for inorganic arsenic of 3×10⁻⁴ (μg g⁻¹ dy⁻¹) (ATSDR, 2006).

^eGreater than an increased lifetime cancer risk of 10⁻⁵ over a 70-year lifetime (US EPA, 2000).

In addition to an increased cancer risk, for the 99^th percentile of consumers, snails from all lakes and Killarney crayfish are also above the US EPA’s chronic oral reference dose (RfD) for non-cancer health effects (3×10⁻⁴ μg g⁻¹ inorganic arsenic; ATSDR, 2006), meaning non-cancerous adverse health effects are possible. Killarney snails are also above this reference dose for the 90^th percentile of consumers (Table 2). The fact that the risk assessment for snails from the shallow reference lake (Bonney Lake) results in an increased cancer risk >10⁻⁵ (Table 2) using the US EPA’s default consumption rate (90^th percentile) suggests that unregulated consumption of Chinese mystery snails could be problematic even at background levels of natural arsenic in the geology of the region (Crecelius et al., 1975) and indicates that snails have an enhanced potential to accumulate arsenic (Coeurdassier et al., 2010; Lai et al., 2012; Madejón et al., 2013). However, the human health risk for consuming snails from Lake Killarney is more than ten times greater than that in Bonney Lake, and thus shallow lakes with anthropogenic arsenic inputs are of even greater concern for snail consumption.

In summary, arsenic levels in littoral zones (sediments and oxic waters) of lakes are more indicative of arsenic bioavailability to biota than profundal conditions. Cancer risks above 10⁻⁵ exist for consuming crayfish and snails from the arsenic-contaminated shallow lakes at the mean and 90^th percentile consumption rates, and for consuming fish from the most contaminated shallow lake for the 99^th percentile of consumers. The human health risk is greater for consumption of organisms from arsenic-contaminated shallow lakes compared to the deep lake with the same profundal sediment arsenic concentrations.

3.4. Rationale for consumption rates used

The mean consumption rate for the general U.S. adult population – 3.27 and 4.23 g dy⁻¹ (uncooked) for shellfish and finfish, respectively – is quite conservative from a risk standpoint as it also includes non-consumers (US EPA, 2002). Therefore, these low-end consumption rates were used to represent a conservative risk assessment for the adult U.S. population as a whole. On the other hand, it is well known that certain populations in the U.S., including indigenous peoples and immigrants from Asia, have much higher ingestion rates for snails, crayfish, and fish (Washington State Department of Ecology, 2013). Therefore, the 90^th percentile consumption rate (17.5 g dy⁻¹), based on average shell- and finfish consumption rates of sport anglers is recommended by the US EPA because it is more protective of the majority of the population, and the 99^th percentile consumption rate (142.4 g dy⁻¹) is to be used for representing consumption rates of subsistence fishers (US EPA, 2000).

Putting the US EPA consumption rates for shellfish (mean 3.27 g dy⁻¹) into perspective, we looked at consumption rates for snail and crayfish taken from international studies with the assumption that immigrants may maintain similar dietary practices (Washington State Department of Ecology, 2013). In fact, the Chinese mystery snail was first introduced to North America in the 1890’s after being imported and sold at Chinese markets in California (Mills et al., 1993). The estimated mean consumption rate of snails only, for the Taihu Lake region in China, where Bellamya sp. is a staple in daily meals, is 20.04 g dy⁻¹ (Kong et al., 2016), and the mean consumption rate for crayfish based on a survey in Nanjing City, China, is 10.1 g dy⁻¹ (Peng et al., 2016). These international ingestion rates specific to snails or crayfish alone are above the freshwater/estuarine shellfish consumption rate for the mean U.S. population used in Table 2 and well below the 99^th percentile U.S. consumption rate, which combines finfish and shellfish together. Thus, the low and high ingestion rates used in Table 2 serve as an acceptable bracketing of potential risk to U.S. consumers, including some of the most vulnerable populations of subsistence fishers. However, the highest reported fish consumption rate in Washington State is 865 g dy⁻¹ (Harper and Walker, 2015) for the Spokane Tribe of Indians in eastern Washington. This consumption rate is six times greater than the 99^th percentile consumption rate considered representative for subsistence fishers. Therefore, high consuming populations, including indigenous peoples and immigrants from Asian countries, may be at particular risk if relying on aquatic organisms from arsenic-contaminated shallow lakes for dietary needs.

3.5. The Need for Future Research on Shallow Lakes

Our results highlight the need for more work on metal uptake and human exposure in shallow freshwater bodies. One study by Yu et al. (2014) assessed the human health risk from consuming fish from Taihu Lake, which is impacted by surrounding agricultural and industrial development. It is one of the largest lakes in China with an average depth of only 2 m. The authors looked at the combination of several organic and inorganic pollutants in fish caught by commercial fishers for human consumption and found arsenic to be the main contributor to an increased cancer risk in both children and adults. Taihu Lake is notably shallow, like Lake Killarney in our study system, however, the link between lake depth and increased health risk seemingly has yet to be specifically explored.

Another study by Lin and Liao (2008) sampled milkfish from twelve groundwater-fed aquaculture ponds from four towns in an area of southwestern Taiwan affected by naturally occurring arsenic. According to the Food and Agriculture Organization of the United Nations, milkfish aquaculture ponds typically range between 0.3–3 m in depth (FAO, 2020). Findings from a survey of consumption rates of the residents of the four towns revealed consuming fish from all twelve ponds resulted in cancer risks considered to be elevated by the US EPA (>10⁻⁶, US EPA, 2000), with 8 exceeding 10⁻⁴. Like the arsenic-contaminated shallow lakes in our study, elevated human health risks were observed when there was a spatial overlap between biota and the presumably oxygenated waters of the culture ponds fed by high arsenic-containing groundwater. The relatively few risk assessment studies that have been published on the consumption of freshwater species collected from shallow lakes and ponds tend to focus on one type of system, whereas our study compares lakes with differing physical and chemical characteristics (i.e. depth, mixing regimes, and extent of arsenic contamination). More studies comparing human health risks resulting from consuming organisms from different types of lakes are needed to identify factors that make certain lakes candidates for enhanced uptake of arsenic into the aquatic food web.

4. Conclusions

The potential human health risks from consumption of aquatic species in shallow urban lakes contaminated with arsenic detailed in this study reinforce the need to consider the particular chemical, biological, and physical mechanisms of shallow lakes when assessing exposure scenarios. Our comparison of Killarney and Angle Lakes, with the same profundal sediment arsenic concentrations, shows that the dynamics of shallow, polymictic lakes can result in greater arsenic bioavailability compared to deeper, seasonally stratified lakes, and therefore greater human exposure and health risk. A more thorough search for arsenic contaminated shallow lakes is warranted and we recommend that monitoring the extent of arsenic contamination include oxic water and littoral sediment sampling, as concentrations in shallow, oxygenated zones may be a better indicator of the potential uptake into organisms compared to conditions in the deeper areas of lakes. We suggest that regional agencies consider regulating the consumption of aquatic species that are especially susceptible to the uptake of arsenic, like Chinese mystery snails. Several states in the northwestern United States, including Washington, do not require a permit to harvest snails or crayfish for consumption, or have very liberal or no limit on harvest numbers, and consequently, information on the demographics of those harvesting snails and crayfish is lacking (Larson and Olden, 2011). Surveying lake users for consumption practices and communication to potentially affected populations should be prioritized to minimize human health risks. Such efforts should involve a diverse group of stakeholders, including representatives from local and state agencies, non‐governmental organizations, sovereign tribal nations, academia, and the public, to ensure effective communication to at-risk populations.

Supplementary Material

Figure S1

Figure S1. Mean arsenic concentrations in Chinese mystery snail (Bellamya chinensis) whole tissues compared to profundal sediments of the study lakes (closed symbols from shallow polymictic lakes, open symbol from the deep, seasonally stratified lake). Error bars represent one standard deviation from the mean. Different letter labels (a, b) denote significant difference (p < 0.05) between lakes.

Figure S2

Figure S2. Mean arsenic concentrations in Chinese mystery snail (Bellamya chinensis) whole tissues compared to June 2019 mean oxic water column of the study lakes (closed symbols from shallow polymictic lakes, open symbol from the deep, seasonally stratified lake). Error bars represent one standard deviation from the mean. Different letter labels (a, b) denote significant difference (p < 0.05) between lakes.

Highlights.

• Littoral sediments more indicative of As bioavailability to biota than profundal

• Greater human health risk from harvest in As-contaminated shallow lakes

• Cancer risk >10⁻⁵ from eating crayfish and snails in As-contaminated shallow lakes

• Subsistence fishing cancer risk in shallow lake with same As level as deep lake