5

Exposure to Contaminants Associated with Consumption of Seafood

This chapter discusses the committee’s review and assessment of evidence on exposure to toxins and toxicants of concern in seafood, sources of contaminants in seafood, and biomarkers of exposure, including evidence published after the report Seafood Choices: Balancing Benefits and Risks (IOM, 2007). The chapter reviews all contaminants for which seafood consumption may be an important route of either chronic exposures or acute (episodic) exposures, although the latter were not reviewed in detail owing to being acute in nature. This chapter also provides further information about the extent to which seafood is likely to be a primary route of exposure, compared with other foods or nondietary exposure routes.

It was not within the scope of the committee’s work to review contaminants associated with consumption of other animal sources of protein or other foods beyond seafood. The committee recognizes, however, that multiple dietary and nondietary exposures exist for many of the contaminants reviewed.

TOXINS AND TOXICANTS OF CONCERN IN SEAFOOD

As noted in Chapter 4, the Dietary Guidelines for Americans 2020–2025 includes recommendations for including seafood in a healthful dietary pattern. Some seafood may contain contaminants or microorganisms that could pose a health risk, particularly to pregnant and lactating women and children. The following sections discuss evidence on contaminants of concern, including variant isoforms identified by the committee as potentially relevant to its task.

Estimates of exposure to contaminants of concern through consumption of seafood depend principally on two factors: the amount of the seafood consumed (discussed in Chapter 3), and the amount of the contaminant in seafood. Using the reported consumption rate from national surveys such as the National Health and Nutrition Examination Survey (NHANES) and the Canadian Community Health Survey (CCHS), along with data on evaluation of exposure or risk assessment reported in the literature, it is possible to estimate quantitatively the exposure to different contaminants from seafood consumption among women of childbearing age and children and adolescents. The concentrations of contaminants in seafood depend on many factors, including species, age, region of origin, preparation method, and parts of the species consumed. These factors are discussed in this chapter as they relate to different classes of contaminants, but it is beyond the committee’s task to perform quantitative exposure estimates of how concentrations of contaminants could change based on variations in such factors.

CONTAMINANTS RESULTING IN CHRONIC EXPOSURES

Metals, Metalloids, and Other Elements

Both essential and non-essential elements “are found naturally in the Earth’s crust, and their compositions vary among different localities, resulting in spatial variations of surrounding concentrations. Metals are substances with high electrical conductivity, malleability, and luster, which voluntarily lose their electrons to form cations” (Jaishankar et al., 2014, p. 60). A metalloid is a chemical element with a predominance of properties that are a mixture of those found in metals and nonmetals. Typical metalloids are metallic in appearance but are brittle and are only fair conductors of electricity that behave chemically mostly as nonmetals (Vernon, 2013). The metals, metalloids, and other trace elements of human health concern that are discussed in this chapter are mercury, cadmium, lead, arsenic, and selenium.

Mercury

Mercury (Hg) is generated as a gaseous element that is released into the atmosphere from natural (volcanic) and anthropogenic (fossil fuel combustion) sources and deposited onto land or water (Driscoll et al., 2013).

Ocean fish and marine mammals are major sources of dietary intakes of mercury. The potential for human exposure from seafood is related to the magnitude of regional and global emissions and deposition but also the ability of watersheds and oceans to convert Hg to methylmercury (MeHg) and biomagnify up the food chain. In general, fish that are higher on their respective food chains (e.g., marlin, sea bass, shark, orange roughy, swordfish, some tuna), tend to have higher total Hg levels.

Hg accumulates in the muscle of fish and is assumed to exist primarily as MeHg as an earlier study reported that, on average, 95 percent of the Hg detected in muscle tissue from 12 fish species was MeHg (Hight and Cheng, 2006). In contrast, only an estimated 45 percent of shellfish Hg is in the form of MeHg (FDA, 2014). The percentage of MeHg in seafood can vary by species and size of the organism (Lescord et al., 2018); for instance, the percentage can be lower in smaller and younger fish, as well as freshwater fish, accounting for approximately 60–80 percent of total Hg (Lescord et al., 2018). Therefore, using total Hg concentrations to estimate MeHg concentrations would likely result in an overestimation of exposure. In addition, the analytical methods for detecting total Hg are simpler and less expensive than for MeHg. Therefore, it is common to use total Hg concentrations as a proxy for MeHg concentrations in fish muscle tissues. Most Hg databases for fish and shellfish report only total Hg; therefore, total Hg is often used as a proxy for MeHg. Because industrial processing of fish and domestic cooking generally do not alter Hg concentrations (Goyer and Clarkson, 2001), the total concentration in raw fish serves as a reasonable approximation of that in prepared fish (Health Canada, 2007).

Although not covered in this report, consumption of marine mammals is also a source of exposure to MeHg, particularly in regions with high consumers such as populations living in western and eastern Canadian Arctic regions (Dietz et al., 2021). Hg concentrations have been detected in numerous marine mammals from these areas, including beluga whales, narwhal, white-toothed dolphins, pilot whales, walruses, harp seals, and ringed seals (Wagemann et al., 1995).

Karimi et al. (2012) created a seafood database of Hg data from federal and state governmental reports as well as the peer-reviewed scientific literature. The database focused on fish and shellfish from sources marketed in the United States, but it is not an exact model of the composition of the seafood market. Furthermore, although Hg is primarily associated with muscle tissue, the data compiled include fillets as well as whole fish. In the analysis, MeHg was used where available instead of total Hg. Overall, the levels of Hg detected varied widely across individual seafood samples, and the amount of variability differed by species. Hg concentrations were generally higher in wild-caught than farmed fish owing to differences among species and other factors.

Hg levels were understudied in farmed seafood and in some major imported fisheries as well as Asia and South America. The mean Hg values in the database tended to be greater than the values reflected in the U.S. Food and Drug Administration (FDA) Hg Monitoring Program data (FDA, 2011; Karimi et al., 2012).

Data on Hg concentrations in popular fish sold in Canada are reported by the Canadian Food Inspection Agency. It found most fish species (e.g., oyster, clams, scallops, mussels, shrimp, salmon, cod, flounder, trout, herring, lobster, crab, lake whitefish) to contain total average Hg levels less than 0.2 parts per million (ppm) (Health Canada, 2007).

FDA reported results from its Total Diet Study (TDS) from 2018 to 2020 for a range of contaminants including Hg (FDA, 2022a). Total Hg concentrations in the collected samples were below 1 ppm, the FDA action level for MeHg in fish, shellfish, crustaceans, and other aquatic animals. The five fish with the highest Hg concentrations were canned tuna (0.230 ppm),¹ baked cod (0.084 ppm), baked salmon (0.021 ppm), pan-cooked catfish, and precooked shrimp.

As discussed in Chapter 3, the most frequently consumed seafood species among women of childbearing age and children and adolescents in the United States are shrimp, tuna, salmon, other fish, crab, breaded fish, catfish, cod, scallops, lobster, and clam (Table 3-4 and Table 3-7). All of those have relatively low Hg concentrations and are classified by FDA in the good choice category (less than 0.23 µg/g [ppm]) or the best choice category (0.15 µg/g [ppm]). Thus, consumers can eat up to two servings of fish per week from the good choice category, or three servings per week from the best choice category without exceeding the reference dose.

Tuna is the only exception in that average Hg concentrations range from 0.13 µg/g [ppm] for canned light tuna to 0.35 µg/g [ppm] for yellowfin tuna and canned white tuna, to 0.36 µg/g [ppm] for fresh or frozen albacore tuna, and to 0.69 µg/g [ppm] for bigeye tuna. Bigeye tuna is in the avoid category (higher than 0.46 µg/g [ppm]) or not recommended to be consumed by women of childbearing age and children. Other fish are in the good choice category (less than 0.46 µg/g [ppm]) and can be consumed once per week (FDA, 2022b). As discussed in Chapter 3, only 19 percent of women of childbearing age (Table 3-6) and 6.4 percent of children (Table 3-12) consume more than two meals of fish per week. Therefore, these groups are estimated to have a low risk from Hg exposure through consumption of fish. Groups that frequently consume tuna make up the primary at-risk group.

In Canada, the most frequently consumed species among women of childbearing age and children are clam, cod, crab, haddock, lobster, salmon, shrimp, tilapia, trout, and tuna (Table 3-14). All of these species except tuna had relatively low Hg concentrations of less than 0.2 µg/g [ppm] according to the Canadian Food Inspection Agency and reported by Health Canada (2007). The concentrations of Hg in tuna ranged from 0.14 µg/g [ppm] in canned light tuna and 0.26 µg/g [ppm] in canned albacore (white) tuna to 0.65 µg/g [ppm] in bigeye tuna, and 1.27 µg/g [ppm] in fresh/frozen tuna.

As discussed in Chapter 3, the average fish consumption rates for women of childbearing age and children and adolescents in Canada are low. The daily consumption rate ranged from 0.08 g of clam per day for children to 5 g of salmon per day for women of childbearing age (Table 3-14). Therefore, like in the United States, most consumers will not be at risk from Hg exposure through the consumption of fish although frequent tuna consumers are at increased risk.

Health Canada advises women who are or may become pregnant and breastfeeding mothers to consume up to 150 g of frozen tuna per month (Health Canada, 2019a). Young children between age 5 and 11 years can consume up to 125 g per month. Younger children (between age 1 and 4 years) are advised to consume no more than 75 g per month. Health Canada issued separate advice for canned albacore (white) tuna and light (e.g., skipjack or yellowfin) tuna. Breastfeeding women and those who are or may become pregnant may consume up to 300 g per week of albacore tuna, the equivalent to about two 170-g cans of albacore tuna per week. Children between age 5 and 11 years may consume up to 150 g (about one 170-g can per week), and younger children age 1–4 years may consume up to 75 g (about half of a 170-g can per week). This advice does not extend to non-albacore, canned tuna, as those species contain significantly less mercury (Health Canada, 2019b).

It is important to note that the above-mentioned hazard identification for Hg exposure applies only to the general population. Sex, age, and proximity to coastal regions influences seafood consumption. Ethnicity is also an important factor in consumption rates, with Asians, Native Americans, Pacific and Caribbean Islanders, and mixed races having the highest consumption rates of fish and shellfish (Mahaffey et al., 2004).

As discussed in Chapters 3 and 4, certain populations such as Native American tribes and Indigenous peoples, as well as subsistence and sport fishers and their families, can consume different fish species or fish parts, or

___________________

¹ The FDA Total Diet Study Food/Analyte Matrix identifies tuna as canned in water or canned in oil.

consume species from locations with potentially higher Hg. In a national study, Chan et al. (2021) found that 1.9 percent of First Nations women of childbearing age, mostly living in northern Ontario and northern Quebec, who consumed large predatory fish had Hg intake levels exceeding the reference dose.

Hg concentrations in fish are not static and can vary over time. For example, trend data for North American freshwater species collected and analyzed by Grieb et al. (2020) identified an overall trend for decreasing Hg concentrations in fish tissue sampled from North American lakes between 1972 and 2016. This trend is consistent with reported Hg emission declines and soil and water deposition trends across the United States and Canada. More recently, a plateau in the rate of change in Hg concentrations in fish tissue has been reported, possibly caused by increased emissions from global sources between 1990 and 1995.

Schartup et al. (2019) collected data for more than 30 years in ecosystem modeling of MeHg concentrations in Atlantic cod. The model estimated a 23 percent increase in MeHg concentrations detected in Atlantic cod, and a 56 percent increase in Atlantic bluefin tuna, in part attributable to increases in seawater temperatures since 1969. This increase in MeHg concentrations in fish tissue is a trend in the opposite direction of the 22 percent reduction that was modeled in the 1990s. Taken together with the plateau that was reported in 2020 for global Hg emissions, these data suggest that ocean warming may become a driver of MeHg concentrations in fish species at the high end of the food chain.

The pattern of Hg exposure can change over time because of changes in the pattern of fish consumption. For example, Sunderland et al. (2018) found that shifts in the edible seafood supply between 2000–2002 and 2010–2012 affected MeHg exposure. These shifts resulted in changes in consumer preference (e.g., away from canned light tuna), global ecosystem shifts (e.g., northern migration of cod stocks), and an increase in the supply of fish from aquaculture (e.g., shrimp and salmon). The data showed that 37 percent of the U.S. population-wide exposure to MeHg in this time frame was primarily from domestic coastal systems: 45 percent was from open ocean ecosystems and 38 percent was from fresh and canned tuna. The Pacific Ocean is estimated to supply more than half the total MeHg exposure. In the United States, aquaculture and freshwater fisheries account for an estimated 18 percent of the total MeHg intake (Sunderland et al., 2018).

Cadmium

Cadmium (Cd) is naturally found in soil, water, and air; activities such as mining, smelting, and fuel combustion can increase human exposure through food chain biomagnifications processes.

While generally low in finfish muscle, shellfish such as oysters and scallops have higher Cd concentrations, ranging from 1 to 4 µg/g in the Pacific Northwest (Bendell, 2010; Pacific Shellfish Institute, 2008). Additionally, Cd can bioaccumulate in the hepatopancreas of crustaceans, including crab and lobster, owing to this organ’s detoxifying function (Chavez-Crooker et al., 2003; Lordan and Zabetakis, 2022). Therefore, heavy consumers of oysters, scallops, and the hepatopancreas may have higher risk of Cd exposure.

Using the data from FDA’s Total Diet Study (TDS, 2014–2016) and food consumption data from What We Eat In America (WWEIA)—the food survey portion of the NHANES—Spungen (2019) estimated that mean Cd exposures for children (age 1–6 years) ranged from 0.38 to 0.44 µg/kg body weight per day, with primary contributions from grains, mixtures (e.g., hamburgers, pizza, soups), and vegetables. The combined daily intake of meat, poultry, and seafood (66 g/day) contributed to 0.25 µg/day of Cd exposure or 3.8 percent of the total cadmium intake of 6.6 µg/day. These estimates are limited because Cd concentrations were only measured in 268 TDS foods compared to the approximately 8,000 foods reported to be consumed by WWEIA/NHANES respondents.

Lead

Pb is ubiquitous in aquatic environments, and it is one of the most accumulative toxic metals owing to its ability to easily bind oxygen and sulfur atoms in proteins to form a stable complex, leading to accumulation in fish tissues (Lee et al., 2019). The Pb concentrations in fish in the United States vary depending on the species, region of origin, and environmental factors (Frank et al., 2019). Fish usually have higher Pb levels than shellfish.

Using data from FDA’s TDS (2014–2016) and food consumption data from WWEIA, Spungen (2019) estimated the mean Pb exposures for children (age 1–6 years) ranged from 1.0 to 3.4 µg/day, with major contributions from fruit, grains, dairy, and mixtures (e.g., hamburgers, pizza, soups). The combined daily intake of meat, poultry, and seafood (66 g/day) was estimated to contribute 0.03 µg/day of lead exposure or 3 percent of the total Pb intake of 1.2 µg/day.

Gavelek et al. (2020) used the same approach and estimated the mean Pb exposure to range from 1.4 to 4.0 µg/day for older children (age 7–17 years), from 1.6 to 4.6 µg/day for women of childbearing age (16–49 years), and from 1.7 to 5.3 µg/day for adults (age 18 years and older). The estimated 90th percentile for lead exposures ranged from 2.3 to 5.8 µg/day for older children, 2.8 to 6.7 µg/day for women of childbearing age, and 3.2 to 7.8 µg/day for adults. Meat, poultry, and fish together contributed to only about 3 percent of the Pb intake. These results suggest that Pb exposure from fish consumption is lower than the interim reference level of 3 µg/day.

Arsenic

Arsenic (As) is a metalloid element globally present in both natural and anthropogenic sources, including commercial uses. Seafood and seaweed are primary dietary sources of total As and are present in marine-derived foods. Seafood consumption is estimated to account for 90 percent of total As exposure in the United States (Borak and Hosgood, 2007).

As is present in seafood primarily in the form of organic compounds. Two exceptions are freshwater fish from Thailand (Jankong et al., 2007), and blue mussel from Norway (Sloth et al., 2016), which are not known to contain As although the element may be found in other types. Arsenobetaine is the major As compound in most fish. It is generally nontoxic and not metabolized. Other organic As compounds include arsenosugars and arsenolipids, which are also present at significant quantities in some types of seafood and are metabolized in humans. Whereas taxonomic group affects the proportion of inorganic As in seafood, elevated levels have been found in bivalves and gastropods (Taylor et al., 2017). These group effects have also been the source of consumption guidelines in the Pacific United States (OHA, 2015).

Luvonga et al. (2020) analyzed total As and As compounds that are commonly consumed in aquatic species to better understand As speciation. The fish and shellfish samples collected were geoduck clam (Panopea generosa) from Alaska, wild-caught brown shrimp (Farfantepenaeus aztecus) from South Carolina, aquacultured white leg shrimp (Litopenaeus vannamei) from Alabama, and wild-caught and aquacultured coho salmon from Alaska and Washington state, respectively. The overall study results identified kelp, wild-caught shrimp, and geoduck clam as the species with the highest total As content.

In a review by Taylor et al. (2017) the presence and distribution of organic As compounds was evaluated in seafood and in combination with human consumption data. They reported that arsenosugars are associated largely with marine algae. The algae accumulate As from seawater and store it as arsenosugars, usually at high concentrations. Mollusks and crustaceans that are predominantly filter feeders are the marine species that contain arsenosugars, although at lower concentrations than algae. Arsenobetaine was reported as the major source in finfish and shellfish.

A paucity of information exists on the distribution of arsenolipids in seafood; however, levels of 50–62 percent have been found in fatty fish, with the higher concentrations in pelagic fish (Lischka et al., 2013). Methylated As compounds in seafood occur from the enzymatic methylation of inorganic As, such as arsenosugars. These compounds generally occur at low levels in seafood, with dimethylarsinate (DMA) the most prominent. Mollusks can contain from 3 to 46 percent of DMA, which is more than usually measured in finfish or algae (Taylor et al., 2017).

Selenium

As discussed in Chapter 4, selenium (Se) is an essential nutrient that functions through selenoproteins, several of which are oxidant defense enzymes. Se can be taken up through various dietary sources, but in quantities greater than the Tolerable Upper Intake Level (UL), potential for toxicity exists (see Chapter 4). Typical total Se concentrations in marine fish species from various origins range from about 0.1 mg Se per kg wet mass to almost

1 mg Se per kg wet mass (Cabañero et al., 2005; Murphy and Cashman, 2001; Navarro-Alarcon et al., 2008; Olmedo et al., 2013). Naturally occurring selenoneine (SeN) occurs in abundance in several marine fish species.

Species of fish that are sources of selenoamino acids and SeN include tuna, swordfish, and deep-sea greeneye (Cabañero et al., 2005; Yamashita et al., 2011). Kroepfl et al. (2015) reported that a methylated form of Se, Semethylselenoneine, has been reported in human blood and urine where it was thought to result from methylation of SeN ingested from fish and was also found in muscle tissue of mackerel, sardine, and tuna. Kroepfl et al. (2015) also identified a selenosugar (selenosugar 1, methyl-2-acetamido-2-deoxy-1-seleno-β-D-galactopyranoside) in marine fish.

The characterization of Se compounds in fish and how they are stored and metabolized is important for understanding the dose–response from fish consumption. For example, SeN, found in tuna and marine mammals, is a powerful antioxidant (Achouba et al., 2016). SeN may protect against MeHg toxicity by increasing its demethylation in red blood cells and in turn decreasing its distribution to target organs. Moreover, Drobyshev et al. (2021) demonstrated that SeN can cross the blood–brain barrier and may reach brain tissue in humans, potentially making it more significant in protecting against MeHg toxicity.

Persistent Organic Pollutants

Persistent organic pollutants (POPs) are a class of synthetic carbon-based chemicals that are characterized by their persistence in the environment owing to their resistance to degradation. They are semi-volatile and accordingly, move long distances and are ubiquitous throughout the environment. POPs are lipophilic substances that sequester in fatty tissue, allowing them to bioaccumulate and biomagnify through the food chain (WHO, 2010). Therefore, organisms at the top of the food chain, including some pelagic fish and humans, have the highest concentrations of these chemicals (EPA, 2023a). POPs traditionally included polychlorinated biphenyls (PCBs), polybrominated diphenyl ethers (PBDEs), and some pesticides (e.g., dichlorodiphenyltrichloroethane [DDT]). However, in 2023, per- and poly(fluoroalkyl) substances (PFAS, e.g., perfluorooctane sulfonic acid [PFOS] and perfluorooctanoic acid [PFOA]) were included as POPs by the Stockholm Convention,² the global treaty targeting POPs. Additionally, perfluorinated compounds share many of the same chemical properties, including their persistence and propensity to bioaccumulate.

Polychlorinated Biphenyls

PCBs consist of 209 congeners, which are composed of two linked benzene rings with between 1 and 10 chlorine substitutions, which dictates their chemical structure and properties (von Stackelberg, 2011). Because of their stability and insulating properties, PCBs were used in a variety of industrial applications until the late 1970s when they were regulated and restricted by the Toxic Substances Control Act (updated in 2016) and the U.S. Environment Protection Agency (EPA) rulemaking.³ However, because of their environmental persistence, PCB contamination is still widespread. Whether released in water from industrial waste or leaching into water from landfills, PCBs sequester in sediment owing to their lipophilicity. PCBs bioaccumulate in aquatic food chains and biomagnify (i.e., the increase in concentration of a substance) in predators that consume contaminated prey (Petersen and Kristensen, 1998).

In humans, the rate of individual⁴ congener metabolism depends on the number and position of chlorine atoms. For example, a study of initial body burden, low serum levels, ongoing environmental exposure, and congeners with very long half-lives found that for congeners Aroclor 1242 and Aroclor 1254, the estimated half-lives during a period of high internal dose were 1.74 years and 6.01 years, respectively. During a period of low internal dose,

___________________

² Available at https://www.pops.int/TheConvention/ThePOPs/TheNewPOPs/tabid/2511/Default.aspx (accessed February 27, 2024).

³ Public Law 114-182, June 22, 2016. Available at https://www.congress.gov/114/plaws/publ182/PLAW-114publ182.pdf (accessed February 27, 2024).

⁴ Bioaccumulation occurs when chemicals become concentrated at levels that are higher in an organism than in open water. Biomagnification occurs when the chemicals become concentrated at higher levels in an organism than in the food chain. See https://www.epa.gov/sites/default/files/documents/bioaccumulationbiomagnificationeffects.pdf (accessed February 27, 2024).

the half-lives for Aroclor 1242 and Aroclor 1254 were estimated to be 21.83 years and 133.33 years, respectively (Hopf et al., 2013).

While occupational exposure routes include inhalation and dermal exposure, the majority of nonoccupational exposure to PCBs comes from dietary sources. A targeted literature review of PCB concentrations in environmental media examined the relative contribution of PCB exposure from different exposure pathways based on studies published since 2007 (Weitekamp et al., 2021). Analysis of data from the studies reviewed indicated that for adults, dietary intake accounted for 88 percent of exposure across the population (ranging from 1.8 to 3.6 ng/kg/day).

Among populations who consume large amounts of fish, particularly from the North American Great Lakes region, PCB exposure from fish may be more substantial (EPA, 2023b). This region supports one of the world’s largest freshwater fisheries, which supplies fish to more than 4 million adults (and their children) in the United States and approximately 1 million adults in Canada, including Native American populations who rely on fish as a primary food source (Gandhi et al., 2017). Currently, concern about PCB exposure accounts for the majority of fish consumption advisories for the Great Lakes (Gandhi et al., 2017). Governmental agencies typically monitor contaminants in fatty fish (e.g., lake trout, lake whitefish, rainbow trout, and chinook salmon) and bottom feeders (e.g., brown bullhead, white sucker, and common carp), as these are the fish that are most likely to accumulate these hydrophobic chemicals (Gandhi et al., 2017). While concentrations of PCBs have been declining over time (Gandhi et al., 2017), for those who consume substantial quantities of fish from the Great Lakes, PCB exposure may still be an important concern (EPA, 2023b).

The National Study of Chemical Residues in Lake Fish Tissue (EPA, 2009), a national freshwater fish contamination survey, estimated the national distribution of selected persistent, bioaccumulative, and toxic chemical residues in fish tissue from 500 lakes and reservoirs in the contiguous 48 states. This 4-year study provided the first national estimates of median concentrations for 268 chemicals in lake fish, defined a baseline for national fish contamination to track progress of pollution control activities, and identified areas where further investigation of contaminants is warranted.

Specifically, 486 predator (fillet) and 395 bottom-dweller (whole-body) samples were collected and analyzed for chemicals including Hg, 17 dioxins and furans, five forms of As, 159 PCB congeners, 46 pesticides, and 40 semi-volatile organic compounds. Overall, the analysis showed that mercury, PCBs, and dioxins and furans are widely distributed in lakes and reservoirs in the contiguous 48 states. However, 43 of the 268 selected chemicals were not detected in any samples, including all nine organophosphate pesticides, 1 PCB congener, and 16 of the 17 polycyclic aromatic hydrocarbons analyzed as semi-volatile chemicals. Forty-eight percent of the sampled lakes had predator Hg tissue concentrations above the 0.3 ppm human health screening value (SV), and 16.8 percent had total PCB concentrations above the 12 parts per billion (ppb) SV. More than 7 percent of the sampled lakes had dioxin and furan concentrations greater than the 0.15 parts per trillion (ppt) SV, 1.7 percent had DDT concentrations over 69 ppb SV, and 0.3 percent had chlordane concentrations greater than 67 ppb SV.

Polybrominated Diphenyl Ethers

Polybrominated diphenyl ethers (PBDEs) are also a family of 209 congeners, which were primarily used as flame retardants in a variety of consumer products, including furniture, electronics, and clothing. PBDEs were phased out of use from 2004 through 2013, largely owing to their persistence and neurotoxicity (Washington State Department of Health, 2024a). Many consumer products that contain PBDE are still in use, however, and when these products are discarded, PBDE may leach from landfills into the watershed and bioaccumulate in the food chain (similar to PCBs). Therefore, it is plausible that as more of these PBDE-containing products are discarded, food, including fish, will become an increasingly important source of exposure (Schecter et al., 2010a). For example, in a recent report of seafood obtained from Puget Sound (Washington State) including both locally captured and nonlocal sources, a variety of finfish as well as bivalves contained the highest levels of PBDE and their metabolites. Finfish including English sole, sablefish, and trout, and shellfish including calamari and shrimp had the highest levels of parent PBDE while bivalves had higher levels of metabolites than parent compounds (Cade et al., 2018).

Human PBDE exposure, like PCB, can be measured in fat-containing tissue and fluids (e.g., blood and breast milk) (ATSDR, 2017; CDC, 2017). Current literature suggests that most important routes of PBDE exposure vary by amount of exposure and age, such that among young children, inhalation and inadvertent ingestion of house dust is likely the most important source of exposure owing to high hand-to-mouth activity. Among children and adolescents with relatively low exposure, ingestion of PBDE-containing food is estimated to be the most important. Among children and adolescents with high exposure, ingestion of dust is the dominant exposure route (ATSDR, 2017). Although consumption of PBDE-containing seafood is an important source of dietary exposure among high fish consumers, one study estimated dietary intake of PBDE using the 2007 U.S. Department of Agriculture loss-adjusted food availability report and a market basket survey estimating PBDE exposure in 31 food types (in 310 samples) (Schecter et al., 2010b). The study found that fish contain relatively high amounts of PBDE, with canned sardines containing the highest concentration of all foods measured (on a pg/g wet weight basis) excepting butter, which contained roughly four times more PBDE. In addition to sardines, salmon also had high concentrations of PBDE, followed by catfish fillets. Despite these figures, the study estimated that on average, fish consumption accounts for a small fraction of total dietary PBDE exposure, which is primarily contributed by dairy and meat.

Per- and Poly(fluoroalkyl) Substances

Per- and poly(fluoroalkyl) substances (PFAS) are a class of synthetic chemicals widely used in various industrial and consumer products. They are more persistent than PCBs in the aquatic and terrestrial environment where they can contaminate drinking water and seafood, among other pathways, resulting in dietary exposure. In the 2020 scientific opinion released by the European Food Safety Authority (EFSA), seafood was identified as the most important contributor to dietary exposure for perfluorooctane sulfonic acid (PFOS), which biomagnifies in aqueous and marine food chains, and for perfluorooctanoic acid (PFOA) (EFSA, 2020).

Young et al. (2022) examined PFAS levels in eight seafood types, including both wild-caught (canned tuna, pollock, cod, crab, and clam) and aquacultured (shrimp, salmon, and tilapia) species. Packaging materials were also analyzed but were not found to contribute to PFAS concentrations observed in seafood. Clams had the highest total PFAS values across all seafood analyzed with PFOA ranging from 4 to 23 µg/kg. All samples were labeled as products of China. Crabs were found to have the second highest total PFAS levels. Total PFAS levels in cod ranged from below the method detection limit (MDL) to 0.96 µg/kg, while in salmon perfluorododecanoic acid was the only analyte detected above the MDL with all values fewer than 0.045 µg/kg. In pollock the sum of total PFAS ranged from not detected to 0.73 µg/kg. The sum of PFAS in tuna (primarily canned and one soft package or pouch) ranged from 0.083 to 1.75 µg/kg, while in tilapia the lowest values with sum of PFAS ranged from not detected to 0.09 µg/kg. Shrimp was found to have the lowest PFAS levels among all types of seafoods tested with only one sample detected at 0.027 µg/kg. This study showed similar trends as a Netherlands study that detected the highest PFAS levels among clams and crab (Zafeiraki et al., 2019).

While levels of PFAS in freshwater fish typically exceed those of fish from commerce, these levels have declined 30 percent (using PFOS as an indicator) between 2008–2009 and 2013–2014 (Barbo et al., 2023). Data from the EPA 2018–2019 National Rivers and Streams Assessment further indicate that PFOS in fish fillet composite samples declined an additional 6.7 percent from 2013–2014 (EPA, 2023c). Given these data were just posted at the time of this report preparation (December 2023), future work examining the data in detail will permit a deeper understanding of current PFAS exposure in fish from freshwater sources.

Microplastics

Microplastics are complex materials with chemical and physical characteristics that can be classified as either contaminants or toxicants. Emerging evidence indicates that microplastic particles may be toxic because of their physical and toxicologic effects as well as acting as vectors that transport toxic chemicals, such as PBDEs and bacterial pathogens, into tissues and cells (MacLeod et al., 2021). At present, no epidemiological evidence exists to indicate that microplastics are toxic to human populations. A summary of current evidence on microplastics from seafood sources is shown in Box 5-1.

Summary of Evidence on Contaminants Resulting in Chronic Exposures

Seafood contains a broad range of toxicants and toxins. Concentrations of metals, metalloids, and other trace elements along with organic compounds such as PCBs, PBDEs, and PFAS can result in chronic exposure through seafood consumption. Concentrations vary widely between species, geographic region, size, and age of the organism, and whether they are wild caught or cultivated, among other factors. Hg is the most studied contaminant in seafood. Because levels of seafood consumption are generally lower than recommended levels, and concentrations of Hg for commonly consumed seafood tend to be relatively low, except for tuna, human exposure is not expected to exceed guideline values. Certain subgroups of the population, owing to their pattern of seafood intake or source of seafood, could be at greater risk of exposure to seafood toxicants.

CONTAMINANTS RESULTING IN ACUTE OR EPISODIC EXPOSURE

Infectious Organisms

Many different toxins of biological origin that can have potentially deleterious effects on consumers have been detected in various types of seafood. Contamination with such infectious organisms tends to be specific to geographic location and time of year, and risk areas or periods are captured as state and local advisories.

Algal Blooms

A number of different toxins are associated with harmful algal blooms that naturally occur when environmental conditions are conducive for the organisms to proliferate at high densities and produce toxic metabolites that may enter various marine organisms destined for human consumption. Although the specific combinations of water and other environmental conditions that promote excessive algal blooms are not well known, bivalve molluscan shellfish such as clams, geoduck, mussels, oysters, and scallops are most likely to incorporate toxic compounds produced by microorganisms due to their filter feeding behavior (Washington State Department of Health, 2024b).

Amnesic Shellfish Poisoning from Domoic Acid

Human exposure to domoic acid (DA), an excitatory neurotoxin produced by marine algae, can lead to amnesic shellfish poisoning. This condition is caused by diatoms of the genus Pseudonitzchia that produce DA. DA becomes concentrated in filter feeding bivalves, especially razor clams, but they also can occur in crustaceans and fish such as sardines and anchovies (OEHHA, 2017). DA is not destroyed by cooking or freezing and inhibits neurochemical processes, resulting in short-term memory loss, and potentially brain damage in humans consuming contaminated seafood (Ansdell, 2019; Grant et al., 2010). Recent studies have reported that year-round consumption of large quantities of razor clams has been associated with persistent memory problems. FDA and EPA established an action level for notifying consumers of 20 mg/kg [ppm] or greater in shellfish tissue (FDA, 2021).

The committee’s evidence review identified one study in humans and one in nonhuman primates relevant to exposure to DA from seafood consumption. Ferriss et al. (2017) examined consumption rates of razor clams by recreational razor clam harvesters. They used food frequency surveys to determine if the harvesters were exposed to DA levels above the regulatory reference levels and/or chronically exposed to low levels of DA. The survey collected data on the daily consumption of clams and the frequency over the past 2 years. Each survey included up to six members of a household and recorded age, sex, and race of household members. The survey results showed that children and young adults (ages 10–20 years) had the highest predicted DA exposure as a result of their lower body weights, and despite lower consumption rates than other age groups. Collectively, about 7 percent of total acute exposures calculated exceeded the regulatory reference dose (0.075 mg DA/kg body weight/day. These exposures were attributed to higher than previously reported consumption rates, lower body weights, and/or consumption of clams at the upper range of the regulatory reference levels for DA.

Petroff et al. (2019) explored how tremors in female Macaca fascicularis monkeys with chronic, low-level oral exposure to DA were related to changes in brain structure and neurochemistry. The exposure period included a pre-pregnancy, pregnancy, and postpartum period. Although the study found considerable variability in DA-related tremors among individual animals, overall results suggested that chronic, low-level DA oral exposure at levels below those previously shown to be asymptomatic were related to significantly greater behavioral tremors compared to nonexposed control animals.

Ciguatera Poisoning

Ciguatera fish poisoning (CFP) is a foodborne disease that results from consuming predatory ocean fish contaminated with ciguatoxins. Similar to ciguatera, neurotoxic or paralytic shellfish poisoning is a neurological disturbance that arises from consumption of tropical reef fish and shellfish. The condition is produced by naturally occurring marine biotoxins such as saxitoxin and other potent neurotoxins associated with certain species of microalgae, which are consumed by fish, as well as single-cell dinoflagellates.

In the United States, up to 5 to 70 cases per 10,000 are estimated to occur annually in endemic states and territories. Symptoms of CFP include nausea, vomiting, abdominal cramps, or diarrhea and occur within a few hours after fish consumption (CDC, 2009). Exposure to these toxins by consuming contaminated seafood affects the nervous system and results in temporary paralysis. Neurologic symptoms can include fatigue, muscle pain, tingling, itching, and reversal of hot and cold sensation (Dickey and Plakas, 2009). In the United States, ciguatera-contaminated fish have been found in coastal waters as far north as North Carolina. Research from previous decades demonstrated CFP is considered one of the most common illnesses related to fish consumption in the United States (Anderson et al., 2021; Pennotti et al., 2013).

Detection of ciguatera is based on clinical presentation as diagnostic tests do not exist. Lopez et al. (2016) conducted a study to identify relevant biomarkers and investigate factors that may contribute to clinical presentation and gene expression in peripheral blood leukocytes. This study found significant differences in plasmablastic lymphoma gene expression patterns among participants with ciguatera poisoning compared with controls, but significant differences in gene expression were not seen when participants with recurrent symptoms were compared with those with acute symptoms.

Based on the proximity of Florida to the geographic distribution of ciguatera, Radke et al. (2015) administered a survey to recreational fishermen and carried out an analysis of reported ciguatera to identify high-risk

demographic groups, high-risk fish types, and catch locations that are linked to CFP in Florida. The survey results identified barracuda, grouper, and amberjack as the most frequently consumed species linked to CFP in Florida. The study further found that Hispanics had higher rates of CFP than non-Hispanics, which was likely attributable to more frequent consumption of barracuda. The majority of CFP cases identified were caused by fish caught in the Bahamas and the Florida Keys.

Vibriosis

Twelve different species of bacteria of the genus Vibrio cause human illness known as vibriosis, which is characterized by gastrointestinal problems marked by symptoms such as watery diarrhea, fever, nausea, and vomiting that may either precede or follow septicemia. This condition is most encountered when raw oysters or undercooked seafood having high concentrations of Vibrio are consumed (Baker-Austin et al., 2018; Leng et al., 2019).

Norovirus

This single-stranded RNA virus is commonly associated with the stomach flu. Shellfish are one of the most common types of food that have been associated with norovirus outbreaks.

Micro-organisms

The committee’s commissioned evidence review of primary studies identified evidence about various microorganisms in seafood, but most of the studies were conducted outside of the United States and Canada. The following sections discuss current evidence from exemplar studies that examined salmonella, Escherichia coli (E. coli), and hepatitis A in seafood and aquatic food sources. Foodborne pathogens of special concern for pregnant women include Listeria monocytogenes and Salmonella enterica because of the increased risk of adverse pregnancy outcomes (Smith, 1999).

Salmonella

The majority of studies in the committee’s evidence review, many of them case study reports, identified salmonella as a microorganism of concern in seafood. From 2011 to 2015, up to three cases per month of Salmonella javiana were reported in restaurants sporadically across Arizona. Venkat et al. (2018) investigated these outbreaks and conducted a case–control study to assess risk factors for the infection. The analysis included 21 laboratory-confirmed cases. Whole genome sequencing demonstrated that all Salmonella isolates were genetically related, signifying that the illnesses were linked to a common source. The analysis further showed that outbreaks could be propagated in a restaurant even when no health violations are noted. The study concluded that the outbreak of Salmonella javiana could have been caused by Salmonella contamination of prepared, uncooked shrimp and that the Salmonella survived the minimum cooking temperatures that are required for seafood in the Arizona Food Code and the FDA Food Code.

Huang et al. (2012) carried out a study to evaluate the prevalence of Salmonella and Vibrio species in seafoods obtained in Singapore. Seafood samples were purchased from three major supermarkets and nine wet markets in Singapore. Results of bacterial counts showed that the highest mean counts were found in thawed-frozen shellfish, although this was not significantly different from fresh shellfish, and fresh prawn. Overall, the study found that seafood sold in Singapore had the potential to be contaminated with Vibrio parahaemolyticus and Salmonella lexington, implicating unsanitary conditions in the markets as a causal factor. However, the final product may also have been contaminated through cross-contamination from raw materials due to mishandling by both vendors and customers.

A study by Hamilton et al. (2018) carried out a quantitative microbial risk assessment for wastewater-fed aquaculture to examine the relative importance of aquaculture practices for microbiological health risks. The premise of the study was that aquaculture-produced seafood products are more likely to contain Salmonella species

than wild-caught seafood. A variety of industrial and consumer processing scenarios were identified to assess the relative risks caused by Salmonella species exposure from consumption of shrimp raised in aquaculture ponds in the United States. The United States has a zero-tolerance policy for the presence of Salmonella species on raw, ready-to-eat, and cooked shrimp.

The results of this analysis found that improper cooking times in non-gamma-irradiated shrimp were associated with the highest risk of infection. Importantly, in each risk scenario, Salmonella species levels in aquaculture ponds had only a moderate effect, indicating that other management points for reducing risks may be more effective. The largest difference between microbiological risks for the scenarios tested was seen for proper versus improper cooking and gamma irradiation of shrimp. The scenarios identified less than one order of magnitude of risk for peeling and deveining versus peeling only.

Escherichia coli

Viganò et al. (2007) evaluated the microbiological quality of ready-to-eat foods in Pemba Island, Tanzania. This study identified thermotolerant coliforms in 34 percent of seafoods and 58 percent of household meals that were tested. Among these, E. coli was the most frequently isolated species found in seafood (15 percent), except for boiled or fried seafood.

Hepatitis A

Filter feeding bivalve shellfish are common vehicles for the transmission of enteric viruses, including hepatitis A virus (HAV), primarily via the fecal–oral route. In 2005, an outbreak of HAV occurred across a four-state region of the United States among individuals who had consumed oysters. Shieh et al. (2007) described an approach using reverse transcription polymerase chain reaction (RT-PCR) to identify a single HAV strain among infected consumers who consumed oysters identified as the outbreak source. This was the first direct evidence to link infected consumers from the outbreak to HAV.

To understand the relationship between two subsequent HAV outbreak incidences from imported coquina clams and the likelihood of the infection coming from the contaminated clams, Pintó et al. (2009) used RT-PCR analysis to estimate the genome copy number of virus particles per gram of tissue in clams associated with the outbreaks. This analytical approach showed a dose–response relationship between the number of infectious particles detected and the probability of infection.

Summary of Evidence on Contaminants Resulting in Acute or Episodic Exposure

Exposure to pathogens and microbial toxins tends to occur episodically as “outbreaks” at a specific time and location or as food poisoning cases among individuals who consume contaminated seafoods. Such risks are often mitigated by closure of harvest at specific time or locations, withdrawing the sales of contaminated seafood from the market, or public health advice to avoid the contaminated seafood.

HUMAN BIOMARKERS OF TOXICANT EXPOSURE ASSOCIATED WITH SEAFOOD CONSUMPTION

Metals and Metalloids

Davis et al. (2014) analyzed NHANES data from 2007 to 2008 to examine dietary intake of mercury, cadmium, lead, and arsenic. The study populations included children younger than 18 years of age and adults. The analysis included whole blood for mercury, cadmium, and lead; urinary lead; and total arsenic and arsenic speciation. Dietary intake was assessed from the NHANES 24-hour recall questionnaire. The study results found significant associations between seafood consumption and mercury, total As, arsenobetaine, and dimethylarsinate (DMA)

among both adults and children. Placental tissue and human milk both contain metals and metalloids, but these biomarkers are not used extensively in studies relating exposures to seafood consumption.

Mercury

Blood Hg is a mixture of both methylmercury and inorganic mercury but is primarily composed of MeHg among seafood-eating populations. Umbilical cord blood Hg concentrations are almost exclusively MeHg. Likewise, hair is largely MeHg and can be subject to external contamination. Nails, like hair, also accumulate MeHg. Inorganic Hg is the dominant type of urinary Hg reflecting exposure to primarily elemental Hg from inhalation and ingestion (i.e., silver-mercury dental amalgams). The transport of MeHg from the maternal bloodstream to human milk is lower than what is transferred via the placenta.

Karagas et al. (2012) reviewed the published literature to assess current evidence on the human health effects of low-level exposure to MeHg, including biomarkers of exposure. As noted by other investigators, total hair or blood Hg levels are regarded as more accurate measures of exposure than dietary assessment. While cord blood Hg may be a better marker of fetal exposure than maternal hair, variability remains a concern.

The Agency for Toxic Substances and Disease Registry and EPA, in their most recent draft “Toxicological Profile for Mercury” (ATSDR, 2022), reviewed Hg levels in blood and urine reported in NHANES, which draws from survey data collected in 2015–2016. Overall, the mean total blood Hg level in the adult U.S. population was estimated to be 0.810 µg/L (95% confidence interval [CI] 0.740, 0.886), whereas the 50th percentiles for total blood Hg in children age 1–5 years was less than 0.28 µg/L. For the 2015–2016 collection period, the 50th percentiles of total urinary mercury were below the detection limit (0.13 µg/L) in children ages 3–5 years.

In Canada, the geometric mean of blood Hg concentrations for women of childbearing age (18–49 years) in the general population (2009–2011) was 0.67 µg/L (Health Canada, 2021), and the geometric mean for children (3–19 years) ranged from 0.27 to 0.3 µg/L from 2007–2017, depending on age and sex (Health Canada, 2019c).

The EPA reference level of 5.8 µg/L Hg in blood is the equilibrium blood Hg level that is associated with a dietary intake of Hg at the current reference dose of 0.1 µg/kg body weight/day. Health Canada uses a blood guideline of 8.0 µg/L, based on the existing provisional tolerable daily intakes for children, pregnant women, and women of childbearing age.

Seafood consumption has been specifically related to biomarkers of total Hg (THg) in pregnancy. Schaefer et al. (2019) examined concentrations of THg from pregnant women in coastal Florida along with a validated dietary questionnaire with a 3-month recall period. Among the 299 women examined, 19 (8.3 percent) had hair Hg concentrations exceeding the EPA reference dose of 1.0 µg/g. Seafood was consumed once per week or more by 35.5 percent of respondents, while 17.1 percent did not consume seafood in the previous 3 months. The highest concentration of THg in hair was observed in women who reported eating seafood three times per week. In a pregnancy cohort study reported by Emeny et al. (2019), prenatal toenail Hg concentrations correlated with the amount of fish/seafood consumption reported on a validated food frequency questionnaire.

Cusack et al. (2017) used data from six consecutive cycles of NHANES (1999–2010) to determine trends in blood Hg levels among women aged 16–49 years and residing in different U.S. regions. Population characteristics examined included age, race/ethnicity, income level, and fish consumption using geographic variables. This study found that women of childbearing age living in coastal regions consumed more fish per month and had higher whole blood Hg concentrations compared with women living in the Midwest after controlling for other confounders. Compared with the results of a previous study by Mahaffey et al. (2009), who examined women of childbearing age using NHANES data from 1999–2004, a modest decrease in the geometric mean blood mercury concentrations was found for women residing in the Atlantic coast (1.55 to 1.35 μg/L) and the Gulf of Mexico (0.96 to 0.88 μg/L). However, after including data from the 2005–2010 NHANES survey cycles, a modest increase in blood Hg concentration was found for women residing in the Inland Northeast (0.77 to 0.85 μg/L) and no change was identified in other regions.

In addition to the NHANES study that included children (Davis et al., 2014), studies of preschool to adolescent children have been conducted in Spain where seafood consumption is higher than the United States. Among 4-year-old children, fish intake was associated with hair Hg concentrations, especially swordfish, lean fish, and

canned tuna (Llop et al., 2014). Similar findings were observed among 9-year-old children in Spain (Soler-Blasco et al., 2019). Likewise, fish intake was associated with hair mercury concentrations among 11-year-old children, with the highest hair Hg concentrations among those who reported eating swordfish (López-González et al., 2023).

Cadmium

Cadmium accumulates in the kidneys; therefore, urinary Cd reflects long-term or cumulative exposure. In contrast, blood Cd represents recent exposure. Cd also binds to nail tissue, and as with urine, it correlates with smoking exposure during pregnancy, the primary source of Cd exposure. Epidemiologic studies of Cd biomarkers and seafood consumption are scant. In a randomized clinical trial testing the Mediterranean diet versus a conventional diet, fish intake was associated with urinary Cd concentrations (Rempelos et al., 2022). However, no association was found between urinary Cd and seafood intake in NHANES (Davis et al., 2014) in adults or children, nor in a Spanish cohort of children (Notario-Barandian et al., 2023).

Lead

Measurements of Pb in blood, urine, and tissues are used to assess exposures among individuals, although blood is most widely used. However, blood makes up less than 2 percent of the total Pb body burden (Levin-Schwartz et al., 2020). Because Pb is eliminated from blood faster than from bone, blood Pb reflects the previous few months of exposure (ATSDR, 2020). Additionally, slow release of Pb from bone can add to blood Pb levels after exposure ends. The use of time-integrated blood Pb measures, however, can mitigate concerns about the magnitude of exposure for both acute and chronic exposure, and can measure long-term exposure.

Urine reflects recent Pb exposure, but interpretation requires measuring the glomerular filtration rate and estimating volume. Urine Pb measures can also exhibit high intraindividual variability. Other tissues that can be used include tooth, saliva and sweat, hair, nails, and semen, although these tissues are not widely used (ATSDR, 2020).

While Pb exposure was unrelated to seafood consumption in the NHANES study by Davis et al. (2014), Pb concentration in urine was associated with shellfish but not with other seafood consumed in a study of 4-year-old children from Spain (Junqué et al., 2022). Thus, Pb exposure is a potential concern in populations with higher intakes of fish.

Arsenic

Urine is the primary biomarker of As. Identification of specific As metabolites in urine is required to measure exposure to inorganic As in fish and seafood from the metabolites of As, monomethylarsonic acid (MMA) and DMA which may be present in seafood and other foods as mentioned previously. Arsenobetaine had not been considered toxic because it is generally excreted unchanged. Evidence suggests, however, that metabolism to inorganic As (iAs) may occur based on mouse studies and in vitro studies which suggests metabolism to other arsenocompounds by the microbiome. Furthermore, both arsenolipids and arsenosugars present in seafood break down to form DMA as the major metabolite in urine (Raml et al., 2009; Schmeisser et al., 2006).

Studies frequently sum As, MMA, and DMA as the measure of total urinary iAs. While urine reflects short-term As exposure, urinary As can be consistent over time among those individuals with consistent exposures (i.e., through drinking water or diet). Blood is a potential exposure marker, but because As is rapidly cleared from the blood and is more complicated to analyze it is rarely used in epidemiologic studies.

In addition to the work mentioned previously from NHANES (Davis et al., 2014), studies in the draft Scientific Opinion on the Update of the EFSA Scientific Opinion on Inorganic Arsenic in Food (EFSA Panel on Contaminants in the Food Chain, 2024) reported elevated urinary total As, arsenobetaine, and DMA concentrations among those who consumed cod, salmon, and mussels, and elevated MMA among those who consumed mussels.

Hair and nails are used as a measure of longer-term As exposure in epidemiologic studies, although hair is susceptible to external contamination, whereas carefully washed nails are typically not. Hair provides an estimate of As exposure over time whereas toenails represent exposure over the previous 6 to 12 months (in adults) and thus

are considered a reliable long-term biomarker of exposure (Signes-Pastor et al., 2021). Cottingham et al. (2013) examined associations between the As body burden, as indicated by toenail As concentration, and potential dietary exposure from seafood with exposure through drinking water. The investigators found elevated toenail As, which is primarily inorganic, in those who consumed greater amounts of dark meat fish (tuna steak, sardines, mackerel, salmon, bluefish, or swordfish), but not more fish overall.

Currently, the dominant compounds of As in human milk are not fully understood, and the concentrations of As are typically considered relatively low (Carignan et al., 2016; Tillett, 2008). A study of repeated measures of human milk following salmon consumption found that As levels peaked in about 8 hours following consumption and comprised arsenolipids (mainly As hydrocarbons for lipid fraction) and arsenobetaine (Xiong et al., 2020).

Persistent Organic Pollutants

As described above, POPs including PCBs, dioxins and dioxin-like compounds (DLCs), and PBDEs can be measured in lipophilic biomatrices (e.g., blood serum or plasma and breast milk) to assess human exposure. PFAS can also be measured in blood and breast milk. Many studies describe human exposure, but fewer studies in North America elucidate exposure attributable to fish and seafood consumption. A notable exception includes some Indigenous populations who, owing to a combination of environmental injustice and cultural practices that include dietary reliance on and importance of local fish, are generally more highly exposed (Cordier et al., 2020).

Polychlorinated Biphenyls

While prior research indicated that fish consumers generally have higher serum PCBs than fish nonconsumers (ATSDR, 1996; Humphrey et al., 2000; Tee et al., 2003), updated analyses indicate that this association is strongest in ethnic groups that consume higher amounts of fish (e.g., Asian, Pacific Islander, Native American, other multiracial ethnicities) (Xue et al., 2014).

In an analysis applying EPA’s Stochastic Human Exposure and Dose Simulation (SHEDS) dietary exposure model to NHANES 2001–2002 and 2003–2004 cycle data, predictors of higher concentrations of PCB include age (older participants have higher concentrations than younger participants ages 12–30 years), ethnicity (individuals from Asian, Pacific Islander, Native American, or other multiracial ethnicities have higher concentrations than other ethnicities), and fish consumption (Xue et al., 2014).

An important limitation of this SHEDS-NHANES analysis is it was unable to estimate PCB from other dietary sources, including meat, skin, fat, and milk. Despite the observation that fish and seafood contain the highest levels of PCBs relative to other food sources, in a recent study of mothers and children from the Midwest, researchers found that among dietary sources of PCB, meat accounts for the majority of exposure and fish and seafood accounts for fewer than a quarter of dietary PCB exposure among both mothers and children (Saktrakulkla et al., 2020).

Dioxins

As described above, limited data are available to estimate dioxin body burden attributable to fish and seafood in the United States and in Canada, with the notable exception of populations relying on fish from the Great Lakes or from contaminated sites. For example, Wattigney et al. (2019) found that urban anglers who consume fish from the Detroit River, which flows between the Great Lakes, did not have higher body burdens of dioxins. However, they did have elevated levels of some PCBs and total blood Hg.

Polybrominated Diphenyl Ethers

Fraser et al. (2009) conducted an analysis evaluating dietary sources of PBDEs using both food frequency questionnaire data, which is meant to approximate a broad picture of average intake, as well as 24-hour recall data prior to blood concentrations in the NHANES 2003–2004 cycle, in which PBDE concentrations were measured.

The study found that poultry and red meat consumption are significant sources of PBDE body burdens, whereas other dietary sources including seafood were not associated with PBDE exposure.

A systematic review (Bramwell et al., 2016) identified two studies, both in European populations, that used duplicate diets to relate dietary exposure to body burden. Neither study found an association between diet and exposure but noted that stronger associations may have been observed only in cases where contaminated food is a regular or major dietary component (e.g., consumption of fish from a contaminated lake).

Per- and Poly(fluoroalkyl) Substances

Christensen et al. (2017) evaluated the association between fish consumption in the past 30 days (assessed via dietary interview) and blood concentration of PFAS in the NHANES 2007–2008, 2009–2010, 2011–2012, and 2013–2014 cycles. The study found that while fish consumption was generally low, consumption of both fish and shellfish was associated with higher levels of almost all PFAS compounds measured, though associations varied by each PFAS and by specific types of fish and shellfish, and those relationships were stronger in higher-income households. While this investigation focuses on recent fish consumption and blood concentrations, PFAS bioaccumulate in human blood and tissue and likely reflect past as well as recent exposure.

A review by McAdam and Bell (2023) focused on the determinants of PFAS exposure among pregnant mothers and neonates. Out of 35 studies reviewed, 14 evaluated dietary determinants for fish or shellfish. Findings from the review were inconsistent for individual PFAS compounds. Approximately half of the studies reported positive findings and half reported null findings, although most of the studies in the review were not from North America.

Among the U.S. and Canadian studies identified in the review that examined fish as a predictor of exposure, the two Canadian studies (Caron-Beaudoin et al., 2020; Fisher et al., 2016) and one U.S. study (Kingsley et al., 2018) reported no associations. Another U.S. study used a principal components analysis and reported a positive association between fish consumption and the principal component characterized by the high concentrations of PFOS and perfluorononanoic acid (Kalloo et al., 2018). Since the time that the review was completed, an additional study of children living in the Boston area used a food frequency questionnaire to evaluate dietary intake and patterns (Seshasayee et al., 2021). This study reported that individual food items were not associated with PFAS exposure. However, a dietary pattern that included high consumption of packaged foods and fish (excluding canned tuna and fried fish) was associated with higher concentrations of all PFAS measured.

Summary of Evidence on Biomarkers of Exposure

Epidemiological studies relating seafood intake during pregnancy to biomarkers of contaminant exposure have largely focused on Hg, with findings of higher blood, hair, and toenail Hg concentrations. Using NHANES data, higher seafood food intake was positively correlated with blood levels of Hg and urinary concentrations of total As, DMA, and arsenobetaine both among adults and children. Studies of the biomarker concentrations of other contaminants associated with seafood consumption are relatively scarce, and for all contaminants very little data exist on biomarker associations with seafood intake during lactation, infancy, and childhood.

FINDINGS AND CONCLUSIONS

Findings

Contaminants of Concern

1. Toxins, toxicants, and microbes, including persistent bioaccumulative chemicals, metals and metalloids, infectious organisms, microplastics, and microorganisms, may be present in seafood at levels hazardous to consumers. The concentration of these various contaminants in seafood depends on many factors, including species, trophic position, size, age, geographic location, and origin—wild caught or farm raised.

2. With the exception of some types of tuna, the most commonly consumed seafood species in the United States and Canada contain relatively low concentrations of methylmercury, and concentrations of other metals and metalloids tend to be limited to certain species and geographic areas.⁵

Acute and Episodic Exposure to Contaminants in Seafood

3. Among adults and children, seafood consumption is associated with higher blood, hair, and toenail levels of Hg, and urinary concentrations of certain forms of As, particularly those common to seafood such as arsenobetaine.
4. Average intake levels of MeHg from seafood are below the FDA Closer to Zero recommended limits among women of childbearing age, infants, and children, except for those who frequently consume tuna.⁶
5. Polychlorinated biphenyls (PCBs) and Hg are the key drivers for fish consumption advisories and PCBs are particularly relevant in the Great Lakes region.
6. Certain population groups, in particular Native Americans and Indigenous peoples, as well as subsistence and sport fishers and their families, may consume more seafood species or seafood components from geographic locations that could have high concentrations of Hg and PCBs than individuals in the general population, thereby exceeding recommended limits.⁷

Conclusions

1. The committee found insufficient evidence to assess exposure to most consumers associated with emerging contaminants in seafood, including PFAS, microplastics, and domoic acids.
2. PBDE exposure is not due primarily to seafood consumption; however, as PBDEs migrate into aquatic environments, risk of increased exposure through seafood may emerge.
3. Although seafood consumption is generally low across population groups, it continues to be an important predictor of MeHg, As, and PCB exposure and may be important for assessing PFAS exposure, where evidence is beginning to emerge. Therefore, if fish intake were to increase to DGA recommended levels, then exposures would likely increase.

RESEARCH GAPS

• Additional research is needed to assess geographic and temporal trend data for levels of MeHg, Hg, and other contaminants in seafood, and to monitor intake levels of these contaminants among women of childbearing age, infants, and children. Special attention should be given to at-risk population groups such as Native Americans and Indigenous peoples, and to other at-risk groups, such as subsistence and sport fishers and their families.
• Research is needed for specific studies that examine As and Se in fish. Additional research is needed to assess the potential protective role of nutrients and other factors, such as selenoneine effects on Hg toxicity.
• More quantitative characterization is needed to assess the risk of chronic exposure to less studied contaminants such as PFAS, arsenic species, microplastics, and DA to assess bioaccumulation in food chains at the levels of exposure and toxicity.
• Research is needed to characterize biomarkers of exposure to contaminants in seafood among women of childbearing age, infants, children, and adolescents. This research is needed to identify and characterize dose–response relationships between contaminants and contaminant mixtures in seafood and adverse outcomes among the children of women exposed during pregnancy and lactation.

___________________

⁵ The sentence was revised after release of the report to the study sponsor to clarify that concentrations of methylmercury vary in different types of tuna.

⁶ This sentence was modified after release of the report to the study sponsor to clarify that recommended limits of contaminant exposure are based on the Closer to Zero Action Plan.

⁷ This phrase was modified after release of the report to the study sponsor to reference recommended limits rather than acceptable risk levels.

REFERENCES