Skip directly to search Skip directly to A to Z list Skip directly to site content




A. Introduction

The following section discusses various COCs, how people might come into contact with these contaminants and the potential health effects that may result. In order for an exposure to these contaminants to occur, all the elements of an exposure pathway must be in place. Exposure pathways are divided into "completed" and "potential" and can be current, past or future. A completed exposure pathway consists of five elements: 1) source, 2) environmental media/transport, 3) point of exposure, 4) route of exposure, and 5) receptor population. A potential exposure pathway exists when some but not all of these five elements are present and the potential exists that the missing elements have been present (past), are present (current) or will be present (future). The completed and potential exposure pathways for the LDW site are given in Tables 5 and 6 below. Each pathway is then discussed in terms of the contaminants of concern and the potential health hazard posed.

Evaluating Noncancer Risk

In order to evaluate the potential for noncancer adverse health effects that may result from exposure to contaminated media (i.e., air, water, soil, and sediment), a dose is estimated for each contaminant of concern. These doses are calculated for situations (scenarios) in which nearby residents might come into contact with the contaminated media. The estimated dose for each contaminant under each scenario is then compared to ATSDR's minimal risk level (MRL) or EPA's oral reference dose (RfD). MRLs and RfDs are doses below which noncancer adverse health effects are not expected to occur (so called "safe" doses). They are derived from toxic effect levels obtained from human population and laboratory animal studies. These toxic effect levels can be either the lowest-observed adverse effect level (LOAEL) or a no-observed adverse effect level (NOAEL). In human or animal studies, the LOAEL is the lowest dose at which an adverse health effect is seen, while the NOAEL is the highest dose that did not result in any adverse health effects.

Due to uncertainty in these data, the toxic effect level is divided by "safety factors" giving the lower and more protective MRL or RfD. If a dose exceeds the MRL or RfD, this indicates only the potential for adverse health effects. The magnitude of this potential can be inferred from the degree to which this value is exceeded. If the estimated exposure dose is only slightly above the MRL or RfD, then that dose will fall well below the toxic effect level. The higher the estimated dose is above the MRL or RfD, the closer it will be to the actual toxic effect level. This comparison is known as a hazard quotient (HQ) and is given by the equation below:

HQ equals Estimated Dose (mg/kg-day) divided by RfD (mg/kg-day)

Noncancer effects from exposure to multiple chemicals is evaluated by summing the hazard quotients to calculate a hazard index. This approach attempts to account for chemical interactions and is discussed further on page 24.

Evaluating Cancer Risk

Cancer Risk: Cancer risk estimates do not reach zero no matter how low the level of exposure to a carcinogen. Terms used to describe this risk are defined below as the number of excess cancers expected in a lifetime.Some chemicals have the ability to cause cancer. Cancer risk is estimated by calculating a dose similar to that described above and multiplying it by a cancer potency factor, also known as the cancer slope factor. Some cancer potency factors are derived from human population data. Others are derived from laboratory animal studies involving doses much higher than are encountered in the environment. Use of animal data require extrapolation of the cancer potency obtained from these high dose studies down to real-world exposures. This process involves much uncertainty.

Current regulatory practice suggests that there is no "safe dose" of a carcinogen and that a very small dose of a carcinogen will give a very small cancer risk. Cancer risk estimates are, therefore, not yes/no answers but measures of chance (probability). Such measures, however uncertain, are useful in determining the magnitude of a cancer threat since any level of a carcinogenic contaminant carries an associated risk. The validity of the "no safe dose" assumption for all cancer-causing chemicals is not clear. Some evidence suggests that certain chemicals considered to be carcinogenic must exceed a threshold of tolerance before initiating cancer. For such chemicals, risk estimates are not appropriate. More recent guidelines on cancer risk from EPA reflect the potential that thresholds for some carcinogenisis exist. However EPA still assumes no threshold unless sufficient data indicates otherwise. [18]

This document describes cancer risk that is attributable to site-related contaminants in qualitative terms like low, very low, slight and no significant increase in cancer risk. These terms can be better understood by considering the population size required for such an estimate to result in a single cancer case. For example, a low increase in cancer risk indicates an estimate in the range of one cancer case per ten thousand persons exposed over a lifetime. A very low estimate might result in one cancer case per several tens of thousands exposed over a lifetime and a slight estimate would require an exposed population of several hundreds of thousands to result in a single case. DOH considers cancer risk to be not significant when the estimate results in less than one cancer per one million exposed over a lifetime. The reader should note that these estimates are for excess cancers that might result in addition to those normally expected in an unexposed population.

Cancer is a common illness and its occurrence in a population increases with age. Depending on the type of cancer, a population with no known environmental exposure could be expected to have a substantial number of cancer cases. There are many different forms of cancer that result from a variety of causes; not all are fatal. Approximately one-quarter to one-third of people living in the United States will develop cancer at some point in their lives. [19]

Multiple Exposure and Toxicological Mixtures

A person can be exposed by more than one pathway and to more than one chemical. Exposure to multiple pathways occurs if a contaminant is present in more than one medium (i.e., air, soil, surface water, groundwater, and sediment). For example, the dose of a contaminant received from fish consumption may be combined with the dose received from contact with that same contaminant in sediment.

It is much more difficult, however, to assess exposure to multiple chemicals. In almost every situation of environmental exposure, there are multiple contaminants to consider. The potential exists for these chemicals to interact in the body and increase or decrease the potential for adverse health effects. The vast number of chemicals in the environment make it impossible to measure all of the possible interactions between these chemicals. Individual cancer risk estimates can be added since they are measures of probability. When estimating noncancer risk, however, similarities must exist between the chemicals if the doses are to be added. Groups of chemicals that have similar toxic effects can be added such as volatile organic compounds (VOCs) which cause liver toxicity. Polycyclic aromatic hydrocarbons (PAHs) are another group of chemicals that can be assessed as one added dose based on similarities in chemical structure and metabolites. In the case of the LDW, PCBs and mercury have similar developmental effects. Although some chemicals can interact to cause a toxic effect that is greater than the added effect, there is little evidence demonstrating this at concentrations commonly found in the environment.

There were hundreds of different contaminants reported in the data sets for fish/shellfish tissue and sediments from the LDW. Most of these contaminants were screened out because they were not at levels that caused health concern, or they lacked comparison values or quantitative toxicological information with which decisions can be made. For the purpose of this health assessment, the consideration and evaluation of the seven contaminants of concern in fish/shellfish and sediments were assumed to protective of human health.

ATSDR's interaction profile for persistent chemicals found in fish looked specifically at the interaction between polychlorinated biphenyls (PCBs), methylmercury, p',p'-DDE, chlorinated dibenzo-p-dioxins (CDDs), and hexachlorobenzene. [20] The profile concluded that data was inadequate to determine if these compounds act independently of one another or in unison with regard to similar toxicological effects. Therefore, it was recommended that additivity be assumed as a public health protective measure in exposure-based assessments of the health hazards associated with exposure to mixtures of these components. In this health assessment, PCBs, mercury, and DDE were identified as contaminants of concern in LDW fish. The additive developmental hazards for these chemicals are considered, and as a result, consumption messages to women/pregnant women are emphasized.

The following evaluations do not rely solely on whether the estimated dose of a contaminant exceeds its health comparison value (i.e., MRL, RfD, cancer risk levels). Factors such as background exposure, a growing scientific data base and the inherent uncertainty in assessing health risk are considered when formulating conclusions. These evaluations are based on current data and subject to change should more data become available relative to the site and/or the toxic potential of the contaminants.


Assessment of risks attributable to environmental exposures is filled with many uncertainties. Uncertainty with regard to the health assessment process refers to the lack of knowledge about factors such as chemical toxicity, human variability, human behavior patterns, and chemical concentrations in the environment. Uncertainty can be reduced through further study.

The majority of uncertainty comes from our knowledge of chemical toxicity. For most chemicals, there is little knowledge of the actual health impacts that can occur in humans from environmental exposures unless epidemiological or clinical evidence exists. As a result, toxicological experiments are performed on animals. These animals are exposed to chemicals at much higher levels than found in the environment. The critical doses in animal studies are often extrapolated to "real world" exposures for use in human health risk assessments. In order to be protective of human health, uncertainty factors are used to lower that dose in consideration of variability in sensitivity between animals and humans, and the variability within humans. These uncertainty factors can account for a difference of two to three orders of magnitude when calculating risk. Furthermore, there are hundreds of chemicals for which little toxicological information is known in animals or humans. These chemicals may in fact be toxic at some level, but risks to humans cannot be quantified due to uncertainty.

The amount of contaminated media (fish, soil, water, air) that people eat, drink, inhale or absorb through their skin is another source of uncertainty. Although recent work has improved our understanding of these exposure factors, they are still a source of uncertainty. In the case of the LDW, uncertainty exists with respect to how much fish people eat from the LDW, how often they are eating it, what species they are eating, how often children use public access areas, or how much sediment or soil children may inadvertently eat. Estimates are made based on best available information or worst-case scenarios.

Finally, the amount and type of chemical in contaminated media is another source of uncertainty. Environmental samples are very costly, so it is not practical or efficient to analyze an adequate number of samples for every existing chemical. Instead, sampling usually focuses on contaminants that are thought to be present based on historic land use or knowledge of specific chemical spills. In the case of the LDW, there are over 1,000 sediment samples which were analyzed for numerous chemicals. Most of the sediment samples were analyzed for PCBs due to knowledge of past industrial use, yet there were several relevant chemicals, such as dioxin, for which very little was known. Furthermore, PCB congener data is also lacking for both fish and sediment, and arsenic species (inorganic vs organic) in fish are unknown.

Table 5.

Completed Exposure Pathways in the Lower Duwamish Waterway
Pathway Time Source Media and Transport Point of Exposure Route of Exposure Exposed Population
Fish Consumption -
Industrial facility discharges and spills, municipal discharges, atmospheric deposition, Salmon River Ingestion Recreational, subsistence, and general consumers
Fish Consumption -
Pelagic Fish
Industrial facility discharges and spills, municipal discharges, atmospheric deposition Pelagic Fish River Ingestion Recreational and subsistence consumers
Fish Consumption -
Industrial facility discharges and spills, municipal discharges, atmospheric deposition, Bottomfish River Ingestion Recreational and subsistence consumers
Contact with Sediments -
Industrial facility discharges and spills, municipal discharges, atmospheric deposition, Intertidal sediments Parks and shoreline access points on the river Incidental ingestion and dermal contact Recreational beach users, habitat restoration, on-site workers, remedial workers
Contact with Sediments -
Tribal netting
Industrial facility discharges and spills, municipal discharges, atmospheric deposition, Subtidal and intertidal sediments River when nets are set for harvest of salmon Incidental ingestion and dermal contact Tribal fisherman
Contact with Sediments -
Crab Fishing
Industrial facility discharges and spills, municipal discharges, atmospheric deposition, Subtidal and intertidal sediments River when pots are set for harvest of crab Incidental ingestion and dermal contact Recreational crab fisher

Table 6.

Potential Exposure Pathways in the Lower Duwamish Waterway
Pathway Time Source Media and Transport Point of Exposure Route of Exposure Exposed Population
Shellfish/Crab -
Industrial facility discharges and spills, municipal discharges, atmospheric deposition, Shellfish River Ingestion Recreational and subsistence consumers
Contact with Sediments -
Industrial facility discharges and spills, municipal discharges, atmospheric deposition, Intertidal sediments River sediments Incidental ingestion and dermal contact Recreational and subsistence Shellfishers
Contact with Surface Water -
Industrial facility discharges and spills, municipal discharges, atmospheric deposition, Duwamish river water column Duwamish river study area Incidental ingestion and dermal contact Recreational river users

B. Completed Exposure Pathways

People who recreate or work along the Duwamish River can be exposed to contaminants in sediment and fish/shellfish. The following pathways analysis estimates exposure that might result from eating fish and contacting sediments in the LDW under various scenarios. Exposure assumptions and estimated doses are given in Appendix C.

  1. Fish Consumption

Average, and high-end exposure doses associated with fish consumption from the LDW were calculated for the contaminants of concern in various fish species. Fish consumption rates for various species commonly found in the LDW were taken from a recent survey of the Suquamish Tribe using data gathered from fish consumers only, and a study of recreational anglers in urban embayments of Puget Sound.[21,22,23] Mean recreational consumption was used to approximate average fish consumption for LDW finfish. These rates were derived from a study of on-shore and boat anglers in urban embayments of Puget Sound. Crab and shellfish consumption were not reported in the recreational study, therefore, the median consumption rate from the Suquamish study was used to approximate average consumption for these species. Use of the recreational ingestion rates and median rate from the Suquamish study to predict exposure for an average fish consumer may be an overestimate. (a) High-end consumption was taken from the 90th percentile values from the Suquamish. While there is no existing study of Muckleshoot fish consumption rates, the assumption that Suquamish tribal members eat a similar amount of fish as do the Muckleshoot is considered reasonable. The Suquamish survey reported the highest average consumption rate to date in Washington State. Exposure assumptions and estimated doses are given in Appendix C.

Exposure doses associated with consumption of groups of fish (anadromous, pelagic, benthic, and shellfish) were also calculated for average and high-end consumers. The recreational study did not present consumption rates for groups of fish. The Suquamish study did; therefore, median consumption rates from the Suquamish study were used to approximate the average consumers exposure, and the 90th percentile consumption rate approximated high-end consumption. In addition to the median and high-end ingestion rates taken from the Suquamish survey, doses were calculated for Asian Pacific Islander (API) consumers using consumption rates from the Asian and Pacific Islander Seafood Consumption Study in King County.[25]

One important aspect to consider when estimating exposure to contaminants in fish from a specific water body relates to the percentage of fish consumed from that water body. If consumption of a particular species caught in the LDW is only a portion of the total amount consumed, then the overall dose for that species must consider the dose contribution from other fishing locations. For the purposes of this health assessment, it was assumed that individuals could rely on the LDW for their entire catch. With respect to salmon, this point is less important as little difference exists between contaminant levels in salmon caught from the LDW versus other area of Puget Sound (see Table 7). This fact indicates that the relatively short residence time of immature salmon in the Duwamish River does not significantly contribute to the overall contaminant burden accrued over the life of an individual salmon.

Anadromous (Chinook and Coho Salmon)

Salmon caught in the LDW are consumed by recreational fishers and are an important resource for the Muckleshoot and Suquamish Tribes. Although salmon are a migratory fish and chemical concentrations in salmon are not thought to be site related, there is a considerable amount of harvest and consumption of salmon from within the LDW study area. Therefore, chinook and coho salmon were evaluated in this public health assessment in order to determine the potential health risk to consumers of these species.

RfDs and MRLsOral reference doses (RfDs) and minimal risk levels (MRLs) are levels of exposure to chemicals below which noncancer effects are not expected. MRLs are set by ATSDR for acute, intermediate and chronic exposure.  EPA sets RfDs based on chronic exposure only. An MRL or RfD is derived by dividing a LOAEL or NOAEL by From 1992 - 1998, chinook and coho salmon were sampled within the LDW site and analyzed for pesticides, PCBs, arsenic, lead, copper and mercury. [25]

As shown in Table C3, doses calculated using average exposure assumptions do not exceed any respective RfDs. This result suggests that people who eat what is considered to be an average amount of coho and chinook salmon would not experience any noncancer adverse health effects. In order to estimate the added effect of each contaminant of concern, a combined dose was compared to a "combined" RfD, called a hazard index. Combining all contaminant doses may overestimate the risk for noncancer health effects, but PCBs, DDE, and methylmercury are all associated with developmental and immune toxicity, therefore it is appropriate to add the hazard quotients for these three contaminants.

The dose estimated for the average consumer of all salmon types (Table C6) exceeds the hazard index. However, the hazard index is only slightly exceeded in this case. Because the estimated doses for each individual contaminant are so far below the actual toxic effect levels upon which the respective RfDs are based, the average consumer of salmon from the LDW is not expected to be at risk for any noncancer adverse health effects.

High-end exposure doses estimated for both chinook and coho salmon consumption exceed the PCB RfD.(b) The high-end consumption dose calculated for all salmon types is 5.4 times higher than the PCB RfD and 1.9 times higher than the RfD for methylmercury. These doses are still well below actual toxic effect levels. The background comparison given below in Section F, Table 7 indicates that salmon caught from the LDW study area do not have higher levels of contamination than salmon caught from more pristine areas of Puget Sound. DOH is currently evaluating the potential human health impacts of PCBs in Puget Sound fish.

The primary health concern associated with PCBs and methylmercury relates to developmental effects in children exposed in the womb. Immune system effects are also of concern for PCB exposure and represent the toxic endpoint upon which the RfD and MRL are currently based. Cancer risks are evaluated below on page 33. Chemical-specific toxicity discussions for each contaminant of concern in fish are provided on page 39.

Bottomfish (English Sole)

Concentrations of contaminants in English sole were selected as representative of bottom dwelling fish. Consumption rates are available for English sole/flounder from the Suquamish Tribe survey. English sole contains relatively high levels of PCBs compared to other species and are one of the most characterized, and abundant species within the LDW. [26] English sole has not been reported as a harvested species in the Duwamish River but outreach efforts indicate that flounder are caught in the river and may be confused with English sole.

The PCB dose estimated for the average consumer of English sole slightly exceeds the RfD. The recreational consumption rate used to calculate this dose may be an over-estimate due to the manner in which the data were presented.(c) Average consumption of grouped bottomfish based on the median Suquamish consumption rate on the other hand does not exceed the hazard index (Table C6). Therefore, the average consumer that eats English Sole or bottomfish from the LDW is not expected to experience adverse health effects.

The high-end exposure dose for English Sole, however, exceeds the PCB RfD by approximately 3-fold, and the dose for grouped bottomfish exceeds the PCB RfD by more than 6 times. The arsenic RfD was also exceeded in the high end consumption of grouped bottomfish scenario.

The average level of PCBs in whole fish samples of English sole (958 ug/kg) is nearly 4-fold higher than skinless fillets (267 ug/kg), while livers contain approximately 22-fold more PCBs (5828 ug/kg) than skinless fillets. Although sampling of whole fish and livers from English sole in the LDW is limited, data from other locations in Puget Sound supports this indication that liver and whole body consumption will result in higher PCB exposure compared to skinless fillets (PSAMP).

Other Finfish (Striped Perch and Quillback Rockfish)

Contaminants found in striped perch were chosen as representative of pelagic fish. There were no other species of pelagic fish that were sampled from the LDW.

Doses estimated for average consumers of perch exceed the hazard index while doses for the high-end consumers do not. This anomaly is due to the fact that the consumption rate used to calculate an average person's exposure to perch was based on shore anglers instead of the median consumption rate from the Suquamish study. The fact that high-end consumers from the Suquamish study eat substantially less perch than anglers may be indicative that the shore angler consumption rate is not representative of an average consumer.

The estimated dose for a high-end consumer of all pelagic fish using contaminant concentrations in perch as a surrogate exceeds the hazard index. PCB exposure contributes to the majority of the hazard index with a calculated dose that is nearly three times higher than the RfD.

Although quillback rockfish are less well characterized, they are an important species to evaluate due to the elevated levels of PCBs and mercury detected in the limited sampling data that is available. It should be noted that the rockfish data assessed here comes from samples taken from Elliot Bay near Harbor Island, which is adjacent to and north of the LDW study area. No rockfish samples are available from the Duwamish River and it is not clear whether rockfish even exist in the river.

As indicated in Table C3 (in Appendix C), rockfish represent the most significant risk for fish consumers. The PCB dose calculated using average exposure assumptions is approximately double the RfD while the high-end dose is 12-fold higher. In addition to PCBs, mercury levels are also elevated in rockfish. The high-end dose for mercury triples the RfD. Data for the other contaminants of concern are not available for rockfish indicating that overall risk could be underestimated. Although the rockfish data come from nearby Harbor Island samples, it is not certain if rockfish are present in the LDW. This assessment has included an evaluation of rockfish due to their high contaminant levels, and their proximity to the LDW site.

While English sole and quillback rockfish represent the highest risk for consumers of finfish, it should be noted the three whole fish samples of shiner perch contained an average of 496 ug/kg PCBs. Information from the WDFW indicate that shiner perch are harvested in the LDW [12]. Data regarding PCB levels in shiner perch fillets was not located.

Total Finfish

The consumption rates used to evaluate exposure to contaminants in individual and groups of fish may underestimate total exposure by not accounting for the fact that people eat a variety of fish across groups. The Suquamish fish consumption study reported consumption rates for total finfish. However, many of the species included in that rate do not exist in the LDW. Therefore, the overall finfish consumption rate was not considered appropriate for exposure estimation. The dose estimates given above for individual species and groups are considered to be sufficiently protective because consumption rates are based on consumers only and assume that 100% of an individuals fish diet could come from the LDW.


Crab were evaluated in this assessment because they are reportedly harvested in the LDW study area. A recent survey conducted by King County noted that while most respondents harvest crabs in the Elliot Bay/Harbor Island area a few did report crabbing in the LDW. Edible meat samples from three individual dungeness and nine composite (45 total) red rock crabs taken from the LDW between 1996 - 1998 were analyzed for various contaminants.

Doses estimated from consumption of red rock crab do not exceed respective RfDs for the average or high-end consumer (Table C3). Average consumption of Dungeness crab also does not exceed RfDs, however, the high-end exposure dose calculated for Dungeness crab exceeds the PCB RfD by almost 4-fold and the arsenic RfD by 2-fold.

The sample size for Dungeness crabs is very small, so there is not much confidence in the contaminant levels for these crabs. Red Rock crabs, however, were sampled in greater number, and they too indicated elevated levels of PCBs and mercury. Therefore, it can be assumed that PCBs and mercury are elevated in both red rock and Dungeness crabs.

Other factors to consider are that arsenic dose calculations assume that ten percent of the total arsenic value is the more toxic inorganic form. Because the percent of inorganic arsenic varies between species and can only be estimated, uncertainty is introduced into the arsenic dose estimates. Second, the Suquamish fish consumption survey indicates that, in general, adults eat more fish than children per body weight. However, consumption of Dungeness crab among children appeared to be higher per body weight than adults although sample size for children in this survey was small. Third, consumption rates for crab could be higher than estimated by the Suquamish survey due to lack of other species (i.e., shellfish) available for harvest. In other words, subsistence consumers may collect more of one species if others are not available. The average consumption rate for all shellfish from the Suquamish survey (consumers only) is 10-fold higher than the rate for Dungeness crab or red rock crab alone.

Finally, a single hepatopancreas sample from a Dungeness crab showed 1647 ug/kg PCBs, which is more than 10-fold higher than average levels found in muscle tissue (130 ug/kg). While Dungeness crabs have not been adequately characterized in the LDW, sampling of hepatopancreas in crabs harvested in Elliot Bay and other areas outside of the LDW also indicates that this organ contains substantially higher levels of PCBs than muscle tissue. Samples from Elliot Bay revealed that PCB levels in the hepatopancreas were on average 80 times higher than average levels found in muscle tissue. [27,28]

Asian and Pacific Islanders

The Asian and Pacific Islander (API) community has been identified as a population that consumes fish from the LDW. In general, grouped fish consumption rates for APIs (90th percentile from API study) exceeded that of average fish consumers (median Suquamish rates), but was less than that of high-end consumers (90th percentile Suquamish rates) for all groups except the pelagic group (Table C6). Estimated exposure doses exceed the hazard index for consumption of anadromous, pelagic, benthic, and shellfish fish groups (hazard indices of 2.4, 6.4, 5.3, and 3.7). Therefore, APIs are potentially exposed to contaminants in fish at levels of concern. In addition, the API Seafood Consumption Study revealed that a large percentage (>40%) of Asian Pacific Islanders surveyed consume the whole crab including the hepatopancreas. This population is therefore potentially exposed to higher levels of PCBs from crab consumption due to the tendency for contaminants to concentrate in the hepatopancreas.

Cancer risk - Fish Consumption

Cancer risks were calculated for fish consumer's exposure to COCs that potentially cause cancer in humans: arsenic, PCBs, chlordane, DDE, and seven different carcinogenic polycyclic aromatic hydrocarbons (cPAHs): benzo(a)pyrene, benzo(a)anthracene, benzo(b)fluoranthene, benzo(k)fluoranthene, dibenz(a,h)anthracene, indeno(1,2,3-cd)pyrene, and chrysene. The PAHs were grouped together by using toxic equivalency factors (TEFs). TEFs were applied to each individual PAH based on their relative toxicity to benzo(a)pyrene. The sum of TEF adjusted values is the toxic equivalent (TEQ). [29] Cancer risk estimates for exposure to individual species using average exposure assumptions range from slight (Red rock crab - 1 cancer estimated per 1,000,000 exposed) to low (English sole - 8 cancers estimated per 100,000 exposed. High-end exposure risk ranges from very low (perch - 5 cancers estimated per 1,000,000 exposed) to low (English sole and rockfish - 4 cancers estimated per 10,000 exposed).

In general, arsenic, cPAHs, and PCBs make up the bulk of the cancer risk from exposure to all species with the highest cancer risks attributable to arsenic and cPAHs. cPAHs, however, were not detected in any finfish, but were detected in some mussel samples. Finfish tend to metabolize PAHs more effectively than do shellfish which may explain why PAHs were found in mussels but not in finfish. [30] The methods used in the analyses of PAHs in finfish were not sensitive enough to make an accurate prediction of the amount of the contaminant in the fish. Although exposure to carcinogenic PAHs is expected to occur, the magnitude is likely to be considerably less than the estimated minimum background exposure from sources in food, water, air, and soil.

Risk associated with consumption of English sole was high compared to that of other finfish. Similarly, consumption of bottomfish, in general, represented the highest cancer risk compared to all groups of fish (Table C7). Consumption of Dungeness crab is associated with highest cancer risk (Table C4). Most of this risk is attributable to arsenic, which is relatively high in Dungeness crab compared to most other fish species that were sampled. Cancer is the primary concern for adverse health effects associated with arsenic exposure. However, this concern is based on human exposure to inorganic arsenic in drinking water. Important differences exists between exposure to arsenic in drinking water versus fish including amount and type of arsenic absorbed.

Only slight risks are associated with exposure to DDE and chlordane in all species. A detailed discussion of chemical specific toxicity is given on page 39.

  1. Contact with Sediments

Humans come into contact with contaminated sediment in the LDW in a variety of ways. Tribal netfishers, crabfishers, and children playing along the shore are exposed to contaminants in the sediment through dermal contact and inadvertent sediment ingestion. These scenarios were used to conservatively calculate the dermal and ingestion exposure doses for each exposed population. Exposure assumptions used in the dose calculations are shown in Appendix C, Table C10.

Tribal Netfishing

The Muckleshoot Tribe harvest salmon from the LDW. In the course of doing so, sediment from the bottom of the LDW adheres to the nets, and tribal fishers that handle them come into contact with the sediment. Doses were calculated for tribal net fishers exposed to sediments through dermal contact and inadvertent ingestion of sediment.

It was originally assumed that tribal netfishing was conducted solely by adults, therefore, an exposure dose was calculated using an adult netfisher scenario (exposure assumptions outlined in Appendix C, Table C10). The Suquamish Tribe indicated, however, that children frequently accompany family members while they are fishing and often grow to be fishers when they reach adulthood. As a result, a "worst case" exposure dose was calculated based on this information. All estimated doses were below RfD/MRLs, and the hazard index calculated for this scenario was well below one (Table C11). Therefore, exposure to contaminants in sediment while netfishing on the LDW is not likely to cause adverse noncarcinogenic health effects.

Crab Fishers

Fish and Wildlife Enforcement officers have witnessed people catching Dungeness, red rock, and graceful crab near Terminal 105 and the old railroad bridge near Harbor Island (see Figure 4). [12] Crab pots rest on the bottom of the LDW and exposure to subtidal sediments is likely to occur when they are retrieved from the LDW. Crab fishing can also be accomplished by wading in intertidal areas and retrieving the crabs by hand or rake. Therefore, exposure to contaminants in intertidal and subtidal sediments is possible.

Crab fishing is thought to occur only at select locations; therefore, it is not appropriate to assume exposure to contaminant concentrations from the entire LDW. Sediment samples used to estimate exposure to crab fishers were selected from within a one-thousand foot radius of the Terminal 105 access point.

Based on exposure assumptions outlined in Appendix C, none of the estimated doses exceed respective RfDs or MRLs.The hazard index is also less than one indicating that adverse noncancer health effects are not likely to occur as a result of direct contact with sediment during crab harvesting.

Children playing at parks/access areas

There are at least 15 public access areas along the LDW. Many of these access areas are boat launches and marinas, but there a few places where children play, or might play. For the purpose of this health assessment, five access points were selected as probable locations where children can contact contaminated intertidal sediments: Duwamish Waterway Park, Gateway Park South, Gateway Park North, Boeing View Trail, and Herring's House Park (see Figure 4). Intertidal sediment samples from within 1000 ft of each individual access area and on the same bank of the Duwamish were used to estimate exposure that occurs at each access point (Figures 5 and 6). Attempts to use samples from smaller radii to better evaluate exposure around each access point were made, but these radii yielded too few intertidal samples.

Estimated doses for children who play once per week at these locations were all below RfDs or MRLs. Furthermore, the highest hazard index was 0.4 associated with the Boeing View Trail site. Therefore, no adverse noncancer health effects are expected to result from children playing at the parks along the LDW.

Cancer Risk - Direct Contact

Cancer risks were calculated for direct contact exposure to COCs that potentially cause cancer in humans: arsenic, PCBs, DDE, chlordane, and seven different polycyclic aromatic hydrocarbons (PAHs): benzo(a)pyrene, benzo(a)anthracene, benzo(b)fluoranthene, benzo(k)fluoranthene, dibenz(a,h)anthracene, indeno(1,2,3-cd)pyrene, and chrysene. The PAHs were grouped together by using toxic equivalency factors (TEFs). TEFs were applied to each individual PAH based on their relative toxicity to benzo(a)pyrene. The sum of TEF adjusted values is the toxic equivalent (TEQ). [29] Chemical specific oral cancer potency factors were used to calculate cancer risks. Cancer risk estimates for each direct contact scenario are given in Appendix C table C12. Cancer risks are given for each carcinogenic chemical and are also summed to give an overall cancer risk.

Risk for netfishers was estimated based on a long-term exposure of an adult netfishing on the LDW. The combined cancer risk for tribal fishers based on this scenario was low (approximately 9 cancers estimated for 1,000,000 persons exposed). The worst-case scenario of a child who accompanies an adult while fishing and then becomes a netfisher as an adult yielded a low cancer risk (approximately 1 cancers estimated for 100,000 persons exposed).

In assessing cancer risks associated with people using LDW access areas, the exposure duration was carried forward from childhood to adolescence and through adulthood for a total of 30 years. Combined cancer risks for five different access locations were calculated. Only very low cancer estimated risks were found. Risks ranged from a low of approximately 2 cancers estimated for 1,000,000 persons exposed (associated with the Duwamish River Park scenario) to a high of approximately 5 cancers estimated for 1,000,000 persons exposed (using the Boeing View Trail scenario).

Finally, cancer risks were estimated for a long-term exposure of an older child that crab fishes on the LDW well into adulthood. The combined cancer risk for crab fishers based on this scenario was very low (approximately 4 cancers estimated for 1,000,000 persons exposed).

Estimated cancer risks for all of the preceding scenarios are very low. As mentioned previously, there are a lot of uncertainties associated with estimating risk. Actual risks can be as high as those that are presented here, or they can be as low as zero (no risk). The estimated cancer risks based on the exposure scenarios evaluated for direct contact to LDW sediments are not at levels of public health concern.

Polychlorinated biphenyl -Toxic Equivalents (PCB-TEQs)

Testing for PCBs: Different methods are used to detect PCBs in fish. The results presented as total PCBs are the sum of three different mixtures of PCBs called Aroclor-1248, -1254 and -1260, which are commonly found in fish. More specific analysis of individual PCB congeners can also be performed to provide a measure of dioxin toxic equivalents (TEQ).PCBs are a large family of similar chemicals called congeners. Some PCB congeners have been shown to cause toxic responses similar to those of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD or dioxin). Because TCDD is a potent carcinogen, the cancer risk associated with these dioxin-like PCB congeners should be evaluated. This is accomplished by adjusting dioxin-like PCB congener concentrations with toxic equivalency factors (TEFs) to account for the fact that they are less potent than TCDD. TEFs have been derived for 12 dioxin-like PCB congeners and range from 0.1 to 0.00001. [31] The concentration of a congener in a sample is adjusted by multiplying the laboratory result by the TEF to give a dioxin toxic equivalent (TEQ) for that congener. The sum of individual TEQs is known as the TCDD equivalent, and can be used with the TCDD cancer potency factor to estimate cancer risk.

Of the 1,200 sediment samples collected from the LDW, roughly half were analyzed for one or more PCB congeners. Less than one-third of all analyses were above detection limits. PCB-126, the most potent PCB congener, was detected in only 16 of more than 600 samples analyzed. The high percentage of estimated and nondetected values, particularly with respect to congener PCB-126, indicates that a PCB-TEQ cancer risk calculation must be viewed with caution. For this reason, dioxin-like PCB risks are not presented with, or added to, cancer risks posed by other substances.

Fish tissue analyses did not include individual PCB congeners. The lack of PCB congener analysis in LDW fish could result in an underestimation of overall health risk to fish consumers and represents a data gap.

C. Multiple Exposure Pathways

People that come into direct contact with LDW sediments are also likely to consume seafood from the LDW. This is especially true of tribal fishers. Combined exposure from the fish ingestion and direct contact pathways was assessed using the hazard index approach. The combined risk from both pathways is little different than the risk associated with fish consumption alone indicating that direct contact with sediments contributes little to overall risk. For example, the combined hazard index based on English sole consumption and sediment exposure for the tribal fisher scenario is 4.2, (4.1 HI for English sole consumption + 0.1 HI for sediment exposure = 4.2). Roughly 98% of the overall health risk in this example is attributable to English sole consumption.

D. Potential Exposure Pathways

Potential exposure pathways associated with the LDW site are discussed below. These pathways are not considered complete because data is lacking for key elements necessary to evaluate exposure.

  1. Shellfish Consumption

The DOH Food Safety and Shellfish Programs advise against consumption of shellfish harvested from the King County shoreline, except for Vashon-Maury Island (Figure 8). In addition, PH-SKC warns about contamination in shellfish, crab, and bottomfish near urban areas along the King County shoreline, including Elliott Bay and the Lower Duwamish Waterway. However, consumption of shellfish from the LDW study area has been reported among recreational and subsistence populations.

In an effort to determine the availability of shellfish from the LDW, two preliminary surveys were conducted, the first in June 2000, and the second in June 2001. According to these surveys, clams were located at each sampling site, but were not thought to exist in high enough numbers to support recreational harvest. In addition, the survey also indicated that clams were relatively abundant between Kellogg Island (Figure 7) and Terminal 105, but that the site is accessible only by boat which would limit recreational harvest. Although the initial survey noted horse clams to be the most frequent species encountered, the following survey did not find any horse clams but did note an abundance of Eastern soft shell clams between Terminal 105 and Kellogg Island. It is likely that the initial survey mistook the soft shell for horse clams as they are similar in appearance. [32] EPA, NOAA, and the Muckleshoot Tribe reviewers found the surveys to be inadequate and, as a result, other surveys will be conducted in the future.

The Suquamish Tribe has expressed interest in the potential for future shellfish harvest in the LDW should there be a time when shellfish exist in adequate numbers. The Suquamish expressed concerns that current assessments and future cleanup efforts will not take into account the potential for a significant shellfish harvest by the tribe in the future.

Most of the contaminants of concern in fish and shellfish have been detected in mussels taken from the LDW. The data indicate that contaminant concentrations in mussels are generally lower than other fish species, however, cPAHs were only detected in mussels. Estimated doses of contaminants from mussel consumption do not exceed any reference dose. Consumption of mussels from the LDW does not pose a significant risk based on exposure to chemical contaminants. As noted previously, DOH advises against consuming shellfish from the LDW due to general pollution concerns that include sewage discharge.

No contaminant data for other shellfish species were located. This can be considered a data gap because different types of shellfish can accumulate varying amounts of contaminants. [33]

  1. Contact with sediments - shellfish harvesting/workers

Individuals involved in recreational or subsistence harvest of shellfish may be exposed to contaminated intertidal sediments within the LDW site. Figure 4 provides an illustration of potential exposure points along the LDW site. Harvesting shellfish can result in exposure through inadvertent ingestion and dermal contact with contaminated intertidal sediments.

The extent of intertidal sediment sampling within the LDW varies considerably. Certain areas have been sampled extensively while others are not well characterized. Few intertidal sediment samples have been collected between Kellogg Island and slip #4, and there are a number of public access sites located within this stretch of the LDW. Intertidal sediment where shellfish harvesters or on-site workers could contact contaminated sediments is limited.

Kellogg Island was chosen as a location where people can potentially catch shellfish because they were reported to be relatively abundant in that area. Sediment samples from the intertidal areas surrounding Kellogg Island were used to approximate the levels of contaminated sediment that a shellfisher would encounter. Estimated doses calculated for a 30-year exposure of an older child harvesting to adulthood did not exceed RfDs or MRLs. The hazard index was also less than one indicating that noncarcinogenic adverse health effects are not likely to occur as a result of direct sediment contact in people that shellfish near Kellogg Island.

Cancer risk estimates for each of the contaminants of concern are given in Table C12. The cancer risk for shellfishers based on this scenario is very low (approximately 3 cancers estimated per 1,000,000 exposed people ).

  1. Contact with surface water - swimming

Individuals engaging in water related recreational activities such as swimming within the LDW may come into contact with contaminated surface water. Swimming in the LDW represents a potential exposure pathway of concern as this activity may result in incidental ingestion of and direct contact with contaminants in surface water. Estimating the amount of chemical exposure from swimming in the river is complicated by the lack of surface water sampling data, and the difficulty in estimating dermal absorption and other exposure parameters. Swimming within the LDW study may also allow for the opportunity to come into contact with potentially contaminated intertidal sediments. However, exposure from swimming or other activities that result in contact with surface water is likely to be far less than that associated with consuming fish/shellfish.

There are over 100 storm drains, a number of combined sewer overflows, and miscellaneous outfalls within the LDW study area. CSOs along the LDW represent a potential concern for recreational swimmers (particularly during and following heavy rain events) as the CSOs discharge untreated sewage into the LDW during storm events when capacity is exceeded. Sewage discharged by CSOs can introduce pathogens such as bacteria, viruses, helminthes, and protozoa into the LDW water column. [34] Advisories warning against swimming near CSOs are posted along the King County shoreline and are discussed further on page 47.

The King County Water Quality Assessment evaluated recreational exposure to contaminants in LDW water. The assessment concluded that there was little risk associated with exposure to chemical contaminants in the water column. Pathogen levels, however, were frequently above levels considered acceptable for recreational purposes such as swimming or SCUBA diving. [34]

E. Chemical Specific Toxicity


Arsenic occurs naturally in rock, soil, water, air and plants. It can be distributed and concentrated in the environment through natural processes such as volcanic action, erosion of rock, or by human activities. It is important to distinguish between organic and inorganic arsenic, as the inorganic form is more toxic. Natural mineral deposits in certain areas of Washington State contain large quantities of arsenic that can impact groundwater. Arsenic is used in the production of wood preservatives, and agricultural chemicals including insecticides and herbicides. Additional uses for arsenic include the production of glass, alloys, and use in the electronics industry. Soil arsenic levels in the Puget Sound region have been affected by deposition from the ASARCO smelter that operated for nearly a century in Ruston, WA until it closed in 1985. [35]

Ingestion of inorganic arsenic has been reported to cause more than 30 different adverse health effects in humans, including cardiovascular disease, diabetes mellitus, skin changes, damage to the nervous system, and various forms of cancer. Numerous epidemiologic (human) studies of large numbers of people in several areas of the world have found strong associations between arsenic exposure in drinking water and cancer of the lung, bladder, and skin. The only large scale study of the effects of arsenic-contaminated drinking water on a U.S. population did not demonstrate an association between ingestion of inorganic arsenic in drinking water and cancer, although hypertensive heart disease appeared elevated in the exposed group. [36] The failure to detect an association with cancer in this U.S. population could be explained by differences in exposure, population sensitivity, and statistical power.

EPA has established a chronic oral RfD for arsenic of 0.0003 mg/kg/day based on a NOAEL of 0.0008 mg/kg/day derived from a study in which a Taiwanese population was exposed to arsenic in drinking water. [37] Adverse health effects observed at or near the chronic LOAEL for this study of (0.014 mg/kg/day) include skin cancer, noncancer changes in the skin, vascular disease, and liver enlargement. Less serious effects were also observed in humans near this LOAEL of 0.014 mg/kg/day and include gastrointestinal irritation such as nausea, vomiting, and diarrhea.

EPA has classified arsenic as a known human carcinogen (Group A) and developed an oral cancer slope factor of 1.5 mg/kg/day to estimate the risk of skin cancer resulting from arsenic exposure. Although this number has been questioned, a recent evaluation by EPA suggests that this number may give a good estimate of combined cancer risk (including bladder and lung) from arsenic in drinking water.

All of the toxicological data for arsenic discussed above is considered to be very strong since it is based on human exposure and has undergone significant review. These studies, in fact, form the basis for a reduction in the federal drinking water standard. They are, however, based on drinking water exposure as opposed to direct contact with sediment and consumption of fish associated with the LDW. Estimating an arsenic dose from fish consumption is particularly problematic since results are reported as total arsenic with no distinction between inorganic or organic forms of arsenic. Inorganic arsenic is thought to be the most toxic while organic forms are less toxic. Some forms of organic arsenic, however, may be more toxic than others, or converted to inorganic arsenic in the body. Available data indicate that inorganic arsenic levels in fish/shellfish vary widely, between 0.1 - 41%. Recent shellfish sampling conducted by ATSDR on Marrowstone Island indicated a ten-fold difference in inorganic arsenic content between horse and native littleneck clams. [38,39] This assessment assumes that of the total arsenic reported in fish samples, ten percent consists of inorganic arsenic, which is consistent with current EPA guidance. [40]


Mercury is a naturally occurring element in several different forms. The most important form of mercury related to exposure at the LDW site is methylmercury found in fish. Methylmercury is formed from inorganic mercury by microorganisms that are present in the environment. It is methylmercury that accumulates in the food chain and represents a potential health concern for consumers of fish. Mercury analyses evaluated in this assessment represent total mercury as opposed to methylmercury. Dose calculations, however, do not attempt to fractionate the concentrations as nearly all of the total mercury found in fish is expected to be in the organic, methylmercury form.

Developmental effects are the primary concern regarding methylmercury exposure and have been demonstrated in both animal and human studies. Recent evidence from two separate studies shows impaired development of children whose mothers were exposed to methylmercury by eating fish and whale meat. Mercury levels measured in the hair of these mothers was correlated with decreased performance in motor and learning skills. A third study showed no impact on childhood development in children whose mothers were exposed to mercury in fish while pregnant. ATSDR used this latter study to derive a NOAEL of 0.0013 mg/kg/day upon which a chronic oral MRL of 0.0003 mg/kg/day is based. EPA derived an oral RfD of 0.0001 mg/kg/day based on one of the former studies in which developmental effects were found. [41]

DOH recently derived a tolerable daily intake (TDI) range for methylmercury of 0.000035 to 0.00008 mg/kg/day based on impaired neurological development in children exposed in utero. [42] The upper-bound of this range is consistent with EPA's oral RfD. DOH also recently evaluated methylmercury exposure in fish-consuming populations. The report concludes that some Native American fish consumers are likely to exceed the TDI for methylmercury based on a detailed analysis of fish consumption rates. The report also states that such over exposure to methylmercury needs to be reduced below the TDI by consuming a variety of salmon species in order to limit the amount of chinook salmon consumed. Chinook contain the highest levels of methylmercury of all the salmon species analyzed. [43,44]

Methylmercury is considered to be a Group C possible human carcinogen by EPA based on limited evidence in animals and inadequate evidence in humans. No cancer potency factor is available from EPA with which to estimate cancer risk. The evidence of developmental toxicity following in utero exposure is, however, of primary concern based on the substantial human evidence that forms the basis for a very low RfD.

Polychlorinated biphenyls (PCBs)

PCBs are a group of human-made chlorinated organic chemicals that were first introduced into commercial use in 1929 as insulating fluids for electric transformers and capacitors. Other applications were soon developed that included their use in hydraulic fluids, paint additives, plasticizers, adhesives, and fire retardants. Production of PCBs in the United States stopped in 1977 following concerns over toxicity and persistence in the environment. [45,46]

There are 209 structural variations of PCBs, called congeners that vary by the number and location of chlorine atoms on the base structure. PCBs are often identified by one of their trade names, Aroclor. Aroclors are various mixtures of congeners defined by a four-digit number. The first two digits represent the number of carbon atoms while the second two digits give the percent by weight of chlorination for the congeners in that mixture.(d) In general, PCB persistence and toxicity increases with the degree of chlorination in the mixture.

Liver toxicity has been demonstrated in animals given high doses of PCBs. [47] Liver toxicity and developmental effects are also well documented in residents of Taiwan and Japan exposed to relatively high levels of PCBs through ingestion of contaminated rice oil. However, the association of these effects with PCB exposure is complicated by concurrent exposure to chlorinated dibenzofurans. [46]

While the "rice oil" incidents in Taiwan and Japan provide good evidence of PCB toxicity in humans, recent studies demonstrate that developmental effects can occur at lower levels of PCB exposure. Deficits in neurobehavioral function in children exposed in utero represent the most compelling evidence that environmental exposure to PCBs have caused adverse health effects in humans. Studies of various human populations exposed to PCBs, primarily through the ingestion of fish, have demonstrated deficits in neurobehavioral function. Learning deficits were maintained in the children of one Lake Michigan fish-eating cohort through 11 years of age. Animal studies have also shown adverse effects on development following prenatal exposure of the fetus. [42,48]

Thyroid dysfunction has also been associated with PCB exposure. Several in vitro and animal studies have shown a reduction in thyroid hormone (thyroxine) levels in response to PCB exposure. [49,50,51] A study in rats exposed in utero to PCBs found hearing deficits concurrent with decreasing thyroxine levels. [52] This finding suggests that interference with thyroxine levels could be a mechanism for the developmental effects associated with children exposed to PCBs prior to birth. The potential for PCBs to disrupt hormone function, including the endocrine system, has been suggested as a mechanism for the reproductive effects of PCBs seen in animals. Some human epidemiological studies provide support for the reproductive toxicity of PCBs including effects on menstrual cycles in women and male fertility. [46]

ATSDR has recently reviewed its MRL considering the more recent human developmental studies discussed above. This review concluded that immune system effects seen in monkeys still represent the most sensitive toxic endpoint of PCB exposure. Further, ATSDR concluded that the existing MRL based on this endpoint should not change and would be protective of the developmental effects found in the more recent human epidemiological studies discussed above. [46] DOH is currently evaluating the available literature to determine the most appropriate health comparison value for PCB exposure.

While high dose animal studies demonstrate that PCBs can cause liver tumors in rats, evidence that PCBs can cause cancer in humans is conflicting. Some studies have linked human exposure to organochlorines with breast cancer while other studies have found no association. Other studies suggest a link between PCB exposure in humans and non-Hodgkin's lymphoma (NHL) based on higher PCB blood serum levels in NHL patients versus controls. One recent analysis of a large cohort of workers exposed while manufacturing PCB containing transformers showed no increase in mortality despite high PCB blood serum levels. The previously mentioned rice oil-poisoning incident in Taiwan did not reveal elevations in cancer mortality. However, an examination of residents similarly exposed in Japan did show an increase in mortality from liver cancer.

As noted previously, some PCBs are thought to exert toxicity via a dioxin-like mechanism. Current evidence indicates that 12 PCB congeners act through this mechanism by virtue of their planar structure that allows for binding to the Ah receptor. For each of these, a toxic equivalency factor (TEF) has been established based on enzyme activity triggered through the binding of this receptor. The amount of enzyme activity induced by each congener is compared with that of 2,3,7,8-tetachorodibenzo-p-dioxin (TCDD) in order to generate each TEF. Congener concentrations are multiplied by their respective TEF to generate the dioxin toxic equivalent value. This value can then be used in conjunction with the cancer potency factor for TCDD to estimate a PCB-TEQ cancer risk.

Considerable uncertainty exists with this approach but it does provide an important estimate of PCB toxicity that may be distinct in both mechanism and toxic endpoint. Some evidence suggests that the dioxin-like congeners correlate with immune system and fetal growth effects but not neurobehavioral impairment. [53,54,55]

Polycyclic Aromatic Hydrocarbons (PAHs)

Polycyclic aromatic hydrocarbons (PAHs) are generated by the incomplete combustion of organic matter including oil, wood and coal. They are found in materials such as creosote, coal, coal tar and used motor oil. Fifteen PAHs of similar structure and physical/chemical properties have been identified in significant quantities in the environment. Based on these similarities along with similarities in metabolism and toxicity, PAHs are often grouped together when evaluating their potential for adverse health effects. Some of this group of PAHs have been classified as probable human carcinogens (Group B2) by EPA due to sufficient evidence of carcinogenicity in animals and inadequate evidence in humans. [56]

Benzo(a)pyrene is the only PAH for which EPA has derived a cancer slope factor and was, therefore, used as surrogate to estimate the total cancer risk of PAHs in sediment. It should be noted that benzo(a)pyrene is considered the most carcinogenic of the PAHs and use of its cancer slope factor as a surrogate for total PAH carcinogenicity may overestimate risk. In order to address this issue, an adjustment was made for each cancer causing PAH based on the relative potency of that PAH as compared to benzo(a)pyrene. Evidence of PAH carcinogenicity in humans and animals indicates that tumor location is relevant to exposure route with dermal and inhalation doses yielding skin and lung tumors, respectively.

Fish and shellfish can accumulate PAHs, therefore uncooked fish have some PAHs in them. Benzo(a)pyrene levels in shellfish from uncontaminated waters is estimated to be around 3 ppb. Detection limits used in the analyses of LDW fish and shellfish, however, were not adequate to distinguish the level of PAH contamination. Cooked meats further add to PAH exposure in the diet. Dietary sources make up a large percentage of PAH exposure in the U.S. population, and smoked or barbecued meats and fish contain relatively high levels of PAHs, but the majority of dietary exposure to PAHs for the average person comes from ingestion of vegetables and grains (cereals). [57]


Cadmium is a metal that occurs naturally in the environment. It is used mainly in batteries, pigments, metal coatings, and metal alloys.

The RfD for cadmium that is ingested with food is 0.001 mg/kg/day, and is based on adverse effects in the kidney. EPA classifies cadmium as a Group B1 probable human carcinogen based on limited evidence in human occupational settings, and sufficient evidence based on animal studies. Occupational exposure to inhaled cadmium is suspected to be a cause of lung cancer in workers while animal studies have confirmed the ability of cadmium to cause lung tumors via the inhalation route. Studies of workers exposed to airborne cadmium also suggest a link with prostate cancer. The ability of cadmium to cause cancer via the oral route is disputed. Several human population and laboratory animal studies have failed to show cadmium to be carcinogenic by the oral route, but other studies indicate an increase in prostate tumors, testicular tumors and leukemia in rats following high dietary doses of cadmium. [58,59]

Cadmium is found in most foods at low levels with the lowest levels found in fruits, and the highest found in leafy vegetables and potatoes. Shellfish have higher cadmium levels (up to 1 ppm) than other types of fish or meat.

p,p'-Dichlorodiphenyldichloroethylene (DDE)

DDE is a compound that is formed when the pesticide DDT breaks down in the environment. DDT was banned for use in the United States in 1972, but due to its persistent nature, DDT and its degradation products are often found in fish and other food products.

Neither EPA nor ATSDR have established an oral RfD or MRL for DDE, but EPA gives an RfD of 0.0005 mg/kg/day for DDT that was based on increased liver size in rats exposed to commercial DDT in the diet. EPA classified DDE as a Group B2 probable human carcinogen based on sufficient evidence of carcinogenicity in animal studies and no data in humans. Dietary exposure to DDE caused liver tumors in two strains of mice and in hamsters from which EPA derived an oral slope factor of 0.34 per mg/kg/day. Dietary exposure to DDE also caused thyroid tumors in female rats.

ATSDR's draft interaction profile recommends adding similar noncancer health effects associated with DDE and other contaminants in fish.


Chlordane is a pesticide that was banned in the United States in 1988. It too is very persistent in nature, and is lipophylic which results in the accumulation in animal fat.

The RfD for chlordane is 0.0005 mg/kg/day and is based on liver toxicity in male mice. EPA has classified chlordane as a Group B2 probable human carcinogen based on sufficient evidence of cancer in animals and inadequate data in humans. Exposure to chlordane produced liver tumors in five strains of mice from which EPA derived an oral slope factor of 0.35 per mg/kg/day.

F. Comparison with Background

Table 7 below presents a comparison of contaminants found in Duwamish River fish versus those found in fish from nonurban areas of Puget Sound.

Table 7.

Comparison of contaminants in fish from the Duwamish River versus nonurban areas of Puget Sound a, b, c, d, [23]
Contaminant Chinook Coho English Sole Rockfish
Duwamish nonurban c
Puget Sound
Duwamish nonurban Puget Sound Duwamish nonurban Puget Sound Elliot Bay
(near Harbor Island)
All other locations in Puget Sound
Aroclor-1254 34 30 28 19 143 7.4 73 6.9
Aroclor-1260 20 21 11 11 68 7.6 219 4.6
Mercury 99 87 42 51 48 51 408 288
Arsenic 1.0 1.0 0.8 0.7 11 7.4 NA 2.4

a = Mean concentrations were used for this comparison with nondetects included as whole value
b = Values are given as parts per billion (ppb) except for arsenic values which are in parts per million (ppm)
c = Nonurban locations are Deschutes, Nisqually, Skagit, and Nooksack rivers
d = All data are from the Puget Sound Ambient Monitoring Program

This comparison clearly indicates that English sole in the Duwamish River have been impacted by a source of PCBs that is not universally affecting Puget Sound. Arsenic levels in English sole from the LDW may be elevated in comparison with background. Also of note are the relatively high PCB levels in quillback rockfish samples from Elliot Bay near Harbor Island compared to background and to other species. Though it appears that Harbor Island rockfish have higher mercury levels than the rest of Puget Sound, this comparison cannot be made unless the data are age adjusted. At any rate, these data indicate that limiting consumption of English sole from the LDW and quillback rockfish from Elliot Bay near Harbor Island will reduce overall exposure even if consumption of other species is increased.

Chinook and coho salmon PCB levels in nonurban areas of Puget Sound are similar to those found in the LDW. The Washington State Department of Fish and Wildlife estimated that about 99% of PCBs in adult chinook salmon returning to spawn in the Duwamish/Green River watershed were accumulated in marine waters of Puget Sound or the Pacific Ocean. Furthermore, PCB levels in Coho are slightly lower in northern Puget Sound, and gradually increase in southern areas of Puget Sound. [60] The reason for this trend is thought to be related to the residence time for Coho in Puget Sound. The longer a fish resides in Puget Sound, the more time it has to accumulate PCBs. Salmon returning to watersheds in southern Puget Sound must spend a longer time in the Sound where exposure to PCBs is greater than in the open Pacific Ocean.

Fish from even the most pristine water bodies will accumulate some chemicals from either natural or wide-spreading anthropogenic sources. Reported average mercury levels for the top 10 types of fish consumed in the U.S. range from 20 - 206 ppb. PCB levels detected in Washington freshwater fish fillets (excluding Spokane River) range from 3.4 - 300 ppb (mean = 49).(e),[61] Therefore, subsistence consumers or other people who eat a lot of fish are potentially at risk of adverse health effects even if they eat fish that are relatively low in contaminants.

G. Benefits of Fish Consumption

It is important to consider the very real benefits of eating fish. Fish is an excellent source of protein and has been associated with reduced risk of coronary heart disease. The health benefits of eating fish have been associated with low levels of saturated versus unsaturated fats. Saturated fats are linked with increased cholesterol levels and risk of heart disease while unsaturated fats (e.g., omega-3 polyunsaturated fatty acids) are an essential nutrient. Fish also provide a good source of some vitamins and minerals. [62,63] The American Heart Association recommends two servings of fish per week as part of a healthy diet. [64]

The health benefits of eating fish deserve particular consideration when dealing with subsistence consuming populations. Removal of fish from the diet of subsistence consumers can have serious health, social and economic consequences that must be considered when issuing fish advisories. The Muckleshoot rely on salmon harvested from the Duwamish River as part of a healthy diet, valuable income source, and as an important part of a rich cultural heritage. Other communities living near the Duwamish River may also be impacted by advisories that recommend limits on fish consumption. Outreach efforts indicate that some residents among API communities may eat higher quantities than estimated in this assessment. Consumption advisories for these high-end consumers could, therefore, significantly impact diet.

Any advice given to fish consumers to reduce the amount of fish they eat based on chemical contamination should attempt to balance the health benefits versus the health risks. In general, people should eat fish that are low in contaminants and high in omega-3 fatty acid. Table 8 below shows published levels of omega-3 fatty acids in fish species compared to the average levels of PCBs in LDW fish. Salmon (chinook and coho) have the highest levels of fatty acids, and the lowest levels of PCBs, and therefore, should be the preferred fish to eat from the LDW.

Table 8.

Published levels of omega-3 fatty acid compared to PCB levels in LDW fish Seattle, Washington [65]
Fish Type omega 3 fatty acid (mg/g) a PCB levels in LDW fish (ug/g)
Chinook 14 51
Coho 8 36
Sole / Flounder 2 267
Perch 3 111
Crab 3 110

a = sum of Eicosapentanoic acid (EPA) and Docosohexaenoic acid (DHA)

Fish consumption advice should also take into account that eating alternative sources of protein also has risks. Increasing the consumption rate of beef or pork at the expense of eating fish can increase the risk of heart disease. Some contaminants that are common in fish, such as dioxin, might also be present in other meats.

Exposure to contaminants in fish can be significantly reduced through simple preparation measures. Simply removing the skin of the fish has been shown to reduce PCB exposure. [66] Samples of LDW striped perch with and without skin supports the notion that removal of skin reduces contaminant levels. Skinless striped perch fillets from the LDW contained levels of PCBs that were nearly 30% lower than fillets with skin (Table 9). Furthermore, cooking fish also reduces PCB levels in the fillets by more than 20%, and in some cases, PCBs were nearly entirely removed through cooking. [67,68] Boiling seafood such as shellfish or crabs, can reduce exposure to some contaminants provided that the water is discarded and not incorporated into a broth.

Table 9.

Comparison of PCB levels in striped perch fillets (skinless vs with skin fillets) from the Lower Duwamish Waterway located in Seattle, King County, Washington
Mean PCB levels in Striped Perch Samples from the Duwamish River (ppb)
With Skin Without Skin Decrease
160 113 29%

H. Existing Advisories

PH-SKC has an existing fish consumption advisory for urban areas along the King County shoreline, including Elliott Bay and the Lower Duwamish Waterway. The advisory warns of contaminants in shellfish, crab and bottomfish. DOH Food Safety and Shellfish Programs advise against eating shellfish from the King County shoreline, except for Vashon-Maury Island, based on biological and chemical contamination associated with urban environments. Community outreach efforts, discussed above, have indicated that the communities surrounding the LDW site are mostly unaware of these advisories. The advisory from PH-SKC does not provide consumption limits nor does it give specific advice against eating any of these species.

PH-SKC also advises against swimming near CSOs of which there are many in the LDW. This advisory warns people of the "dangers of swimming or fishing in water that might be polluted because of a sewer pipe overflowing in the area during and after heavy rain. Bacteria and chemicals from CSOs can increase the risk of getting sick from swallowing the water or eating the fish. Public Health recommends that people not go in the water near these signs for 48 hours after a heavy rain." More health related information about CSOs is available from PH-SKC at .

DOH has issued a statewide fish advisory recommending that women of child bearing age and children under six years of age limit their consumption of canned tuna fish and not eat swordfish, shark, tilefish, king mackerel or fresh caught or frozen tuna steak. More information regarding this advisory is available at or by calling toll-free 1-877-485-7316. DOH is currently evaluating PCB contamination in Puget Sound fish. If more stringent consumption limits are derived from this evaluation, they will be applicable to the Duwamish River.

I. Fish Meal Limits

The following meal limits in Table 10 were derived from average and high-end mercury and PCB levels in LDW fish/shellfish. Limits were calculated using average concentration estimates of mercury and PCBs for various fish species with a target hazard index of 1. Exposure parameters are provided in Appendix C, Table C8. These limits represent consumption rates that would be protective of people who consume fish from the LDW. While it is not likely for a person to eat fish solely from the LDW, the limits in Table 10 are for individual species harvested from the LDW. Depending on the source of the fish, people may be able to safely eat more fish meals than shown in Table 10.

The limits are calculated based on fillets or muscle tissue without skin. Consumption of whole fish at these meal limits may result in exposure above safe levels. On the contrary, proper preparation and cooking will reduce PCB exposure further below safe levels.

Table 10.

Meal limits based on PCB, mercury, and DDE contamination in Lower Duwamish Waterway fish, Seattle, Washington.
Fish Species Recommended 8 ounce meals per month
Developmental b Immune c
English Sole 0.9 0.7
Perch 2.1 1.7
Chinook 3.0 3.7
Coho 5.0 5.2
Red Rock Crab 1.9 1.7
Rockfish a 0.6 0.6

a = Rockfish were sampled from Elliot Bay near Harbor Island
b = Based on developmental endpoint of PCBs, mercury, and DDE assuming a female body weight of 60 kg
c = Based on the Immune endpoint of PCBs assuming an adult body weight of 70 kg

Applying the Table 10 meal limits across the general population assumes that meal size will decrease proportionately with body weight. Such an assumption could result in an underestimate of exposure for consumers who eat proportionately more fish per unit of body weight. Table 11 demonstrates how an eight-ounce meal for a 70-kilogram adult would change to remain proportional with body weight.

Table 11.

Adjustment of fish meal size based on the body weight of a fish consumer
Body Weight (lbs) Adjusted Meal Size (oz)
200 10.4
150 7.8
100 5.2
50 2.6
25 1.3
20 1.0

It is important to consider that commercially purchased fish also have contaminants in them. People who abide by meal recommendations for LDW fish based on Table 10, but also eat commercially bought fish may increase their risk for adverse health effects.

J. Child Health Considerations

ATSDR recognizes that infants and children may be more vulnerable to exposures than adults when faced with contamination of air, water, soil, or food. [69] This vulnerability is a result of the following factors:

  • Children are more likely to play outdoors and bring food into contaminated areas.

  • Children are shorter and their breathing zone is closer to the ground, resulting in a greater likelihood to breathe dust, soil, and heavy vapors.

  • Children are smaller and receive higher doses of chemical exposure per body weight.

  • Children's developing body systems are more vulnerable to toxic exposures, especially during critical growth stages in which permanent damage may be incurred.

In this health assessment, exposure scenarios took into account the factors listed above. With regard to fish consumption, ingestion rates from the Suquamish study were normalized based on body weight. The use of adult consumption rates from the Suquamish study was considered to be protective of children due to the finding that Suquamish adults eat more fish per body weight than do children (with the exception of Dungeness crab). The sediment exposure scenarios at public access areas recognized children as the most sensitive receptor and most likely to be exposed to contaminated sediments.

New draft guidance from EPA recognizes that early life exposures associated with some chemicals requires special consideration with regard to cancer risk. [70] Mutagenic chemicals in particular have been identified as causing higher cancer risks when exposure occurs early in life when compared with the same amount of exposure during adulthood. cPAHs have been identified as a mutagenic contaminant of concern in the LDW. Arsenic, DDE, and Chlordane have also tested positive on some assays used to determine a chemicals mutagenic potential. Adjustment factors have been established to compensate for higher risks from early life exposures to these chemicals. A factor of 10 is used to adjust early life exposures before age two, and a factor of 3 is used to adjust exposures between the age of 2 and 15. The following example shows how the lifetime increased cancer risk from exposure cPAHs from consumption of English sole would be adjusted to account for early life exposure.

The cancer risk attributed to cPAH exposure in high-end (subsistence) consumers of English sole is 5 x 10-5 resulting from a lifetime average daily cPAH dose of 7x10-6 mg/kg/day (Table C4). The adjusted risk is as follows:

7x10-6 mg/kg/day x 7.3 kg-day/mg {(2yr/70yr x10) + (13yr/70yr x 3) + (55yr/70yr)}

= 7x10-6 mg/kg/day x 7.3 x(114/70) = 8 x 10-5

The adjustment increases the overall risk associated with cPAH exposure from 5 x 10-5 to 8 x 10-5, or a factor of less than 2. While it may be appropriate to adjust each of the cancer risks attributable to cPAHs by this factor, it should also be remembered that cPAHs were not detected in any finfish, and therefore, the theoretical cancer risks are both very low and highly uncertain.

a consumption rates from the recreational study (Landolt et al) were reported as g/day during the fishing season. This consumption rate was converted to g/kg/day assuming a body weight of 72 kg and the presence of fish in the fishery for 183 days per year for resident fish, and 120 days per year for salmon. The resulting consumption rate may be biased high. Furthermore, the median ingestion rate from Suquamish fish consumers is likely to overestimate average consumption because Suquamish tribe ingestion rates are among the highest in Washington State.
b EPA provides an oral reference dose (RfD) for PCBs that is equivalent to and based on the same human exposure study as the MRL. RfDs have essentially the same definition as MRLs but the two are not always equivalent. ATSDR recently completed an update of the PCB chronic MRL and did not change it. The agency did, however provide a new intermediate MRL for exposure occurring during pregnancy. The intermedia MRL (0.00003 mg/kg/day) is only slightly higher than the chronic MRL of 0.00002 mg/kg/day.
c Ingestion rates were reported as grams per day per season. Since seasons vary, it was unclear what an ingestion rate was over an entire year. For English sole in the LDW, it was assumed that a fishing season was 6 months due to the fact that sole seasonally migrate to deeper water.
d Aroclor-1016 does not follow this naming convention.
e Based on composite samples of freshwater species from several lakes/rivers (excluding Spokane River). Nondetects eliminated from analysis due to high detection limits in some analyses.

Next Section      Table of Contents The U.S. Government's Official Web PortalDepartment of Health and Human Services
Agency for Toxic Substances and Disease Registry, 4770 Buford Hwy NE, Atlanta, GA 30341
Contact CDC: 800-232-4636 / TTY: 888-232-6348

A-Z Index

  1. A
  2. B
  3. C
  4. D
  5. E
  6. F
  7. G
  8. H
  9. I
  10. J
  11. K
  12. L
  13. M
  14. N
  15. O
  16. P
  17. Q
  18. R
  19. S
  20. T
  21. U
  22. V
  23. W
  24. X
  25. Y
  26. Z
  27. #