Report Summary

Report Contents

Risk Assessment
Review and Analysis

Feasibility Study
Review and Analysis

Proposed Cleanup (Action) Levels for
the Fox River and
Green Bay

Literature Cited

Complete Report
for easy printing

Report PDF File

About Dr. Foran

Dr. Foran's CV (PDF)

COMMENTS OF

Jeffery A. Foran, Ph.D.

on

The Draft Baseline Human Health and Ecological Risk Assessment
The Draft Feasibility Study
The Proposed Remedial Action Plan

for the 

Lower Fox River and Green Bay

10 December 2001

1.0 Risk Assessment Review and Analysis

1.1. Human Health Risk Assessment - Overview

PCBs, as well as other contaminants in the Fox River and Green Bay pose significant risks to human health. PCB exposure occurs from a variety of sources and routes including consumption of contaminated fish, consumption of contaminated wildlife, consumption of water, inhalation of aerosols containing PCBs and other contaminants, and dermal contact with contaminated water via recreational and occupational activities. 

The health effects of exposure to PCBs have been well documented (ATSDR 2000) and occur in two fundamental categories – cancer and non-cancer effects. PCBs are suspected of causing a variety of cancers in humans who have been exposed occupationally. There are also indications that some subgroups of women exposed to PCBs in the environment may be at increased risk for breast cancer. There is conclusive evidence that PCBs cause cancer in a variety of non-human animals and in a variety of organ systems. Combination of information from cancer studies of humans and non-human animals has led the U.S. EPA to classify PCBs as probable human carcinogens.

Non-cancer effects of exposure to PCBs include an array of adverse health outcomes including effects on various human organs and systems such as eyes (ocular effects), kidneys (renal effects), and hormones (endocrine effects), developmental effects, neurological effects, immunological effects, reproductive effects, and genotoxic effects. A recent body of scientific evidence has also raised significant concern with a variety of non-cancer effects in the fetus and in young children exposed to PCBs. A thorough discussion of the adverse health effects of exposure to PCBs can be found in the ATSDR’s Toxicological Profile for Polychlorinated Biphenyls (ATSDR 2000), and at www.foxriverwatch.com.

The Fox River/Green Bay Baseline Human Health Risk Assessment (FR/GB BLHHRA) indicates that many residents of the region are subjected to significant cancer and non-cancer risks associated with exposure to PCBs. Of particular concern are the extremely high cancer risks that occur in individuals who consume fish from the system. Cancer risks in recreational and high-intake fish consumers are greater than 1X10^-3 (one excess cancer attributable to PCB exposure in every 1000 exposed individuals). This cancer rate is approximately equal to the cancer rate experienced by individuals who smoke two or three packs of cigarettes a day. It is also 100 to 1000 times greater than cancer risks commonly determined to be "acceptable" by state and federal agencies, and more than one 10 times greater than cancer risks that have triggered stringent regulation of pollutant sources by the U.S. EPA. 

These risk levels, while disturbing in their own right, should be of additional concern because of the large number individuals who face these risks. The BLHHRA estimates that there are upwards of 14,000 highly exposed recreational anglers in the Fox River/Green Bay region. Many family members of these anglers are also likely exposed to PCBs at similar levels. Additionally, 12,000 individuals are considered high-intake consumers from under-represented groups including low-income minority anglers, Hmong anglers, and members of the Oneida band living in the region. Therefore, assuming that for every highly exposed recreational angler there is one highly exposed family member, the total population facing cancer risks in excess of 1X10^-3 approaches 40,000 individuals.

Fish consumers are not the only individuals who face elevated cancer risks from PCB exposure. Hunters, through consumption of contaminated wildlife, face cancer risks approaching 1X10^-4, a level deemed "unacceptable" by both state and federal regulatory agencies. Other exposures (drinking water, recreational exposure, occupational – marine workers) are also associated with cancer risks, although these risks are relatively low compared with risks associated with consumption of contaminated fish and wildlife.

The risk of adverse, non-cancer effects in individuals exposed to PCBs through fish consumption or other routes is also significant. Risks for non-cancer effects are typically assessed quantitatively via a Hazard Index (HI) approach. Simply stated, Hazard Indices of less than 1 indicate that there is a relatively low risk of experiencing non-cancer effects from PCB exposure, while Hazard Indices greater than 1 indicate an increasing risk of non-cancer effects. Stated another way – the lower the HI, the lower the likelihood of an adverse effect, the greater the HI, the greater the likelihood of an adverse effect. 

As with cancer risks, high intake and recreational fish consumers face the highest Hazard Indices, ranging from 10 to nearly 100. Hunters who consume their catch also face hazard indices greater than 10, while other exposure routes (drinking water, etc.) result in hazard indices of less than 1 but greater than 0.1.

1.2 Human Health Risk Assessment – Analysis

Human health risk assessments require information on human exposure to one or more contaminants, and information on the toxicity of specific contaminants such as PCBs. The FR/GB BLHHRA relies heavily on published information on human exposure to and toxicity of PCBs. This approach is both appropriate and necessary, since conducting new exposure and toxicity evaluations is immensely time consuming and cost prohibitive. At the same time, the approach requires selection of data and information from numerous studies, many of which provide different quantitative evaluations of exposure and toxicity. In these cases, the assessment must evaluate the quality of toxicity and exposure data, and choose which data to rely on for the analysis.

The FR/GB BLHHRA cites many of the pertinent studies of PCBs that provide information on toxicity and exposure. It also attempts to choose data that reflect the range of toxicity and exposure estimates. 

1.2.1 Exposure

The BLHHRA evaluates a variety of studies of fish consumption behaviors in recreational anglers and subsistence (high-intake) fish consumers. These studies provide a range of fish consumption rates, and the BLHHRA has chosen a subset of rates from these studies. Unfortunately, the BLHHRA does not rely on a full range of fish consumption rates, nor does it incorporate the "high-end" consumption rates into the risk assessment for cancer and non-cancer effects. 

The BLHHRA uses fish consumption estimates to calculate cancer and non-cancer risks for recreational anglers that range from 4,086 g/yr (central tendency estimate - CTE) to 13,393 g/yr (reasonable maximum estimate - RME), and consumption estimates for high-intake fish consumers that range from 6,129 g/yr (CTE) to 20,203 g/yr (Sections 5.4.3, pages 5-22 to 5-24). However, West et al. (1993), as cited in the BLHHRA, reports fish consumption rates in recreational anglers as high as 28,490 g/yr (RME), and rates in high-intake fish consumers that range from 15,695 g/yr (CTE) to 40,150 g/yr (RME). West’s high-end fish consumption rates for recreational and high-intake fish consumers are more two times greater than "high-end" rates chosen for the BLHHRA. 

Similarly, the BLHHRA incorporates assumptions about how a fish are cleaned and cooked. Cleaning and cooking may reduce the quantities of PCBs in fish; thus, human exposure to PCBs may be reduced with appropriate cleaning and cooking practices. However, as cited in the BLRA (Volume 1, Section 5.2, page 5-6), many individuals do not clean and cook their fish in ways that reduce PCBs in fish tissue; thus, these individuals will be exposed to higher PCB concentrations. Despite this information, the BLHHRA uses a PCB "reduction factor" of 50% based on the assumption that individuals practice appropriate cleaning and cooking procedures.

Finally, the BLHHRA acknowledges (BLRA, Volume 1, Section 5.2, page 5-8) but then disregards that fact that at least some individuals who fish and consume their catch also hunt and consume their game, drink contaminated water, and are exposed to PCBs via recreational and occupational activities. These would clearly be the highest exposed individuals, but are not accounted for in the development of quantitative estimates of exposure and health risks. 

1.2.2. Toxicity

The Baseline Human Health Risk Assessment uses a standard approach to the assessment of PCB toxicity. It follows U.S. EPA guidance for the evaluation, selection, and use of toxicity studies in the risk assessment. It also appropriately considers concurrent exposure to PCBs and other toxicants such as dioxin, through use of toxicity equivalency factors (TEF) and additivity assumptions (for toxicity) in risk calculations. 

The BLHHRA appropriately acknowledges the extensive animal evidence as well as less extensive human epidemiological data that support conclusions of human cancer risk. However, the BLHHRA improperly states (BLRA, Volume 1, Section 5.6.2, page 5-60): "It is also important to note that some studies have concluded that PCBs are not carcinogenic in humans based upon negative epidemiological studies." Negative human epidemiological studies are incapable of proving that PCBs are not human carcinogens. In fact, most scientific studies in general are incapable of proving a negative assertion. The BLHHRA acknowledges appropriately the significant limitations of the Kimbrough study. But on a broader scale, negative epidemiological studies, which are quite common, are virtually never used to "prove" that a substance is not a human carcinogen. Human epidemiological studies typically lack adequate power to identify adverse effects occurring in populations at relatively low levels or frequencies. This is due to the very large sample or population size required to provide adequate statistical power to detect rare or uncommon events. Therefore, it is completely inappropriate to declare that a negative epidemiological study proves or concludes that PCBs don’t cause human cancer, and the language on page 5-60 of the BLHHRA should be modified accordingly.

The BLHHRA describes the use of the reference dose (RfD) in Section 5.7.1, page 5-80. This description is not consistent with the U.S. EPA definition of the RfD (upon which the BLHHRA relies heavily), and as such is misleading. The description of the RfD should state:

Finally, the BLHHRA is lacking in that it does not address the potential for, and in some cases evidence of synergism when assessing the risks of concurrent exposure to multiple toxicants. For example, research indicates that PCBs and mercury, and PCBs and DDE act synergistically in affecting a variety of biological functions such as neurological development and survival (Bemis and Seegal 1999, Wren et al. 1987). PCB co-occurs with mercury and DDE throughout the Fox River/Green Bay system, but the BLHHRA does not address the potential for synergistic toxicity among these contaminants.

1.3. Ecological Risk Assessment – Overview

The goal of the Baseline Ecological Risk Assessment (BLERA) is to determine whether there are ecological risks in the Fox River/Green Bay system at levels that warrant remedial action of stressors causing those risks. A standard, although somewhat simplistic, approach (U.S. EPA 1997) was used to conduct the BLERA for the Fox River/Green Bay system. This approach is driven by the questions posed in the problem formulation stage of the RA. These questions take the following form:

For each measurement and/or assessment endpoint (e.g., fish survival and 
reproduction), are levels of site contaminants sufficient to cause adverse effects? 

To assess risk for each assessment endpoint, the concentration of a contaminant in the ambient environment or in a biological medium (e.g., fish tissue) was compared with safe levels of the contaminant in the medium or in tissue (determined, for example, by the NOAEC TRV). The ratio of the two concentrations is the Hazard Quotient. There is no risk where the HQ based on the NOAEC is less than one. Potential risk for an assessment/measurement endpoint exists where the LOAEC HQ is less than and the NOAEC is greater than 1. Risk increases as the LOAEC HQ increases above 1.

Significant ecological risks occur at virtually all levels of biological organization and for all assessment/measurement endpoints throughout the Fox River/Green Bay system. Hazard Quotients are elevated above 1 for: 

The magnitude of ecological risks is considerable, ranging higher than 350 for mink exposed to PCB in the lower Fox River and Green Bay. Other highly elevated risks include:

Mink exposed to PCBs in Little Lake Butte des Morts (HQ > 170)

Mink exposed to PCBs in the reach from Appleton to Little Rapids (HQ >190)

Mink exposed to PCBs in the reach from Little Rapids to DePere (HQ > 290)

Mink exposed to PCBs in the reach from DePere to Green Bay (HQ > 350)

Mink exposed to PCBs in Green Bay Zone 2 (HQ > 350)

Mink exposed to PCBs in Green Bay Zone 3 (HQ > 190)

Mink exposed to PCBs in Green Bay Zone 4 (HQ > 219) 

1.4. Ecological Risk Assessment – Analysis

The Baseline Ecological Risk Assessment (BLERA) provides an overview of the risks to ecological receptors in the Fox River/Green Bay system associated with exposure to selected chemicals. The report acknowledges that the Hazard Quotient approach used to assess risk can be overly simplistic and potentially unrepresentative of actual risks experienced in the field. Therefore, field validation of HQ-based risk estimates was conducted, and confirmed that adverse ecological effects have indeed occurred in the Fox River and Green Bay.

The HQ-based approach to risk assessment for chemicals of concern is relatively thorough and generally follows U.S. EPA guidelines for the conduct of ecological risk assessments. However, the BLERA falls far short of a comprehensive assessment and, as such, is likely not conservative in that it ignores the issue of concurrent impacts of multiple chemical and non-chemical stressors. The U.S. EPA as well as many non-EPA scientists have acknowledged the importance of considering multiple stressors in ecological risk assessment (Foran and Ferenc 1999, Ferenc and Foran 2000). There is ample evidence that non-chemical stressors (e.g., disease, thermal stress, habitat alteration, and many others) influence an organism’s response to chemical stressors (and vice versa) in ways that may render organisms significantly more susceptible to chemical stressors than predicted from chemical-specific, single-stressor assessments. The consultant’s report itself acknowledges this issue (Draft Baseline Human Health Ecological Risk Assessment, Volume 1, Section 6.3.1, Page 6-39) by stating: "tests using field-collected species are confounded by the presence of multiple contaminants and potentially non-optimal health…and may be affected by both contaminant and non-contaminant stressors." But rather than acknowledge the important influence of non-chemical stressors ("non-optimal health") in assessing risks for chemicals of concern in the Fox River/Green Bay system, the report dismisses this critically important concept by relying on laboratory-based assessment of chemical risk. The result of this simplification is to underestimate risk associated with exposure to Chemicals of Concern in the Fox River/Green Bay system. It is likely then that risk-based cleanup levels will not be protective of ecological receptors, where cleanup levels are derived from assessments that fail to account for multiple chemical and non-chemical stressors.

1.5. Sediment Quality Thresholds

Sediment Quality Thresholds (SQT) are levels (concentrations) of PCBs in the sediment that will not pose a risk to exposed humans or ecological receptors. SQTs were developed only for PCBs, based on human health and ecological risk assessments, as exposure to PCBs poses the greatest risks to human health and ecological receptors. While the report acknowledges that SQTs are not necessarily cleanup levels, they do provide a benchmark to assess the magnitude or extent of risk remaining once cleanup levels have been identified, and they provide goals or benchmarks that indicate PCB sediment concentrations below which risks to humans and ecological receptors are acceptable and/or negligible. 

SQTs are established via modeling that assesses the propensity of PCBs to accumulate in tissues of important ecological receptors (e.g., fish), the degree to which humans or fish-consuming wildlife who consume fish will be exposed to PCBs, and the risks faced by humans and fish-consuming wildlife. SQTs for PCBs are as low as 1 ug/kg (parts per billion or ppb in sediments) for high-intake fish consumers based on an acceptable cancer risk level of 1X10^-6, 28 ug/kg for high-intake fish consumers based on a hazard index of 1, and as low as 24 ug/kg (ppb) for mink based on a NOAEC HQ of 1. 

1.6. Conclusions 

1.6.1 Human Health

The Baseline Human Health Risk Assessment (BLHHRA) identifies extremely elevated cancer and non-cancer risks in large populations of high-intake and recreational fish consumers and in hunters. However, the assessment does not incorporate high-end assumptions associated with fish consumption, it does not address cumulative exposure to contaminants via different activities (fishing, hunting, recreational/occupational activities, etc.), and it assumes, incorrectly, that all fish consumers practice cleaning and cooking procedures that reduce contaminant levels by up to 50%. Incorporating high-end consumption estimates, removing assumptions of contaminant loss from cooking and cleaning, and incorporating concurrent exposure from non-fish sources (e.g., hunting) results in estimates of health risks that are approximately three to four times (3-4X) greater than risk estimates provided in the BLHHRA for high-intake and recreational fish consumers. Further, the assessment of PCB toxicity used in the BLHHRA, while progressive in its incorporation of additivity assumptions, is also not conservative. The BLHHRA fails to address synergism associated with concurrent exposure to PCBs and DDE, and PCBs and mercury. Therefore, despite the highly elevated risk estimates presented in the BLHHRA, the assessment is not conservative and may underestimate total human health risks associated with exposure to contaminants in the Fox River/Green Bay system. 

1.6.2. Ecological Risk Assessment

The Baseline Ecological Risk Assessment also identifies elevated risks to all levels of biological organization in the Fox River/Green Bay system. Highest risks occur in fish-eating mammals (mink) in the Lower Fox River and Zone 2 of Green Bay. 

Despite the highly elevated ecological risk estimates throughout the system, the BLERA is not conservative, and likely underestimates risk to components of the Fox River/Green Bay ecosystem. Underestimation of risk is primarily due to a lack of consideration of the influence of non-chemical stressors (e.g., disease, thermal stress, habitat modification, etc.) on an organism’s susceptibility to chemical toxicity. It has been well documented that non-chemical stressors increase susceptibility to chemical stressors (and vice versa); therefore, adverse impacts likely occur in the Fox River/Green Bay system at contaminant levels below those deemed safe or acceptable in the BLERA.

The implications of lack of conservatism in both the human health and ecological risk assessments are discussed in the following section on Sediment Quality Thresholds.

1.6.3. Sediment Quality Thresholds

Sediment Quality Thresholds (SQTs) are used to predict concentrations of PCBs in sediments that will not pose an unacceptable risk to humans and ecological receptors. SQTs are risk-based numbers and as such, are only protective when based on human health or ecological risk estimates that are adequately protective themselves. As the human health and ecological risk assessments developed for the Fox River/Green Bay BLRA lack appropriate conservatism (they underestimate risk), SQTs based on these assessments are themselves under protective. As acknowledged in the BLRA, SQTs are not cleanup levels themselves. However, they serve as benchmarks for cleanup activities and as assessment tools for the efficacy of remedial actions. In this capacity, SQTs must be based on fully protective assessments of human health and ecological risk.

back to top