CHAPTER TWO Contents

2.1 INTRODUCTION

2.2 HEALTH

2.2.1 Methodologies for Community Health Assessment: Measuring Injury to Health

Scientific, Diplomatic and Political Responses to Great Lakes Pollution

Community Health and Areas of Concern

Health Effects in Canadian Areas of Concern: 17 Health Canada Reports

Mapping

Extended Narrative: Focus on Windsor

Further Statistical Analysis

Responses from the Medical Officers of Health

Availability of Health Data in the United States

Biochemical Epidemiology

Issues of Economic Costs of Disease and of Equity

Contrasting the Scientific Approach and the Public Health Approach

Conclusions

Research Needs

2.2.2 Protecting Human Health from Exposure to Contaminants in Great Lakes Fish

Introduction

Risk Assessment and Limitations

Risk Management

Risk Communication

2.3 WATER QUALITY ASSESSMENT AND REPORTING

2.3.1 Review of Annex 1 of the Great Lakes Water Quality Agreement

Introduction

A Proposal to the Commission

Background Report

Workshop

Findings and Recommendations

2.3.2 Review of Agreement

2.3.3 Nonpoint Sources of Pollution from Land-use Activities

Introduction

Insufficient Persuasive Evidence of the Effectiveness of Best Management Practices

Lack of Performance Standards

Economic Considerations

Inadequate Institutional Arrangements

2.3.4 Evaluation of SOLEC Indicators Relative to the GLWQA

Criteria

Review of SOLEC Indicators

Results

Discussion

2.3.5 Remedial Action Plan Assessment: Site Visits to the Niagara River
and St. Lawrence Areas of Concern

2.3.6 The Use of Atmospheric Modeling in Policy Development
and Using Models to Develop Air Toxics Reduction Strategies

Conclusions and Recommendations

Core Findings

Sources and Loadings

Modeling

Ambient Sampling

Source Control Initiatives

2.4 EMERGING ISSUES IN GREAT LAKES SCIENCE, RESEARCH AND POLICY

2.4.1 Application of a Methodological Framework and a Proposed Process
for Agreement Institutions in Addressing Emerging Issues in Great Lakes

Science, Research and Policy

2.4.2 Green Chemistry

2.4.3 Integrated Observation and Monitoring Network

2.5 ACTIVITIES AND MEETINGS OF THE SCIENCE ADVISORY BOARD
FOR THE 1999-2001 BIENNIAL CYCLE

2.6 BOARD AND WORK GROUP MEMBERSHIP FOR 1999-2001

2.7 REFERENCES

Appendices

I. List of Health Outcomes Selected by Health Canada on the Basis that They Might Be Linked to Pollution

II Mortality, Morbidity and Congenital Anomalies Rates (1986-1992) in the Windsor Area of Concern

Tables

1. SOLEC Evaluative Framework Developed by Clemson University Researchers

2. Ranking of Indicators Deemed as Excellent, Good and Moderate in Fulfilling Reporting Requirements
under the Agreement

3. 12 Principles of Green Chemistry

DISCLAIMER

Every attempt has been made to attribute the expert opinions and comments in this report and to clearly distinguish them from the advice of the Science Advisory Board to the International Joint Commission. While these views are important and relevant for the report, they do not necessarily represent the views of the board or the International Joint Commission.

2.1 INTRODUCTION

_U nder the Great Lakes Water Quality Agreement, the
Science Advisory Board (SAB) was established to
provide scientific advice to the International Joint Commission (IJC) and the Great Lakes Water Quality Board (WQB) and is responsible for developing recommendations on all matters related to research and the development of scientific knowledge pertinent to Great Lakes water quality. Such a broad mandate requires a multi-disciplinary approach and accordingly, members are appointed from each country by the IJC with expertise in the natural, physical and social sciences. Within the broad mandate of the board, there are several processes that determine what is to be done and which of the myriad problems or scientific issues are most salient in terms of water quality, and for consensus building to support SAB recommendations.

Major activities of the SAB are largely tied to the IJC's biennial cycle. Development of board priorities by the IJC is based on consultation with its boards at the beginning of each biennial cycle and with the public during the biennial Public Forum. The Commission assigns a lead role to each of the advisory groups, according to whether the question is one of policy, science or research. A sole priority topic was initially assigned to the SAB for the 1999-2001 biennial cycle titled Methodology for Assessing Community Health in Areas of Concern , building on the previous work of the board in human and community health (1992 workshop in Ann Arbor, Michigan titled Our Community, Our Health , IJC 1994) . Throughout the past decade, as a result of

significant investments by the United States and Canadian governments in Great Lakes environmental health research, there has been a new appreciation of the injury to human health from exposures to persistent toxic substances, particularly from the consumption of Great Lakes fish. Of particular concern are the irreversible changes in fetal developmental processes as a result of maternal consumption of Great Lakes fish prior to and during pregnancy. The challenge for researchers, managers and policy makers is to determine the relevance of the effects identified by the scientific studies and the extent of injury occurring in critical subpopulations in communities, particularly in Areas of Concern (AOCs).

During the preparation of its Tenth Biennial Report on Great Lakes Water Quality, and in assessing the threat to human health from persistent toxic substances during its review of Annex 12, the IJC noted its serious concerns about this injury to human health from exposures to contaminants in Great Lakes fish. Generally, the policy response of governments is to issue fish consumption advisories. As stated in the report, this can only be considered an interim solution pending restoration of the chemical integrity of the ecosystem. Further, Commissioners requested additional advice on the adequacy of the advisories during the 1999-2001 biennial cycle and thus an additional human health related assignment was referred to the board.

A third major activity of the SAB, the review of Annex 1 of the Great Lakes Water Quality Agreement, evolved from the preliminary assessment of Lake Michigan Mass Balance data, by the SAB's Work Group on Parties Implementation. The original intent of this activity, to compare Lake Michigan Mass Balance data with Annex 1 Specific Objectives, led the work group to address broader issues related to the current relevance of Annex 1 to the Agreement, and its usefulness as a framework for reporting the status of Great Lakes water quality. The Commission accepted the advice of the board to fund further study, and in September 2000, this initiative began.

Several other important scientific issues were identified by the board and addressed with the assistance of student teams from Clemson University on the review of the State of the Lakes Ecosystem Conference ( SOLEC ), and from the University of Guelph of the review of Nonpoint Sources of Pollution from Land-use Activities .

The SAB also collaborated with the International Air Quality Advisory Board (IAQAB) on Atmospheric Deposition Modeling to Develop Control Strategies ; with the Water Quality Board on the Review of the Great Lakes Binational Toxics Strategy; and with the Council of Great Lakes Research Managers on the Application of a Methodological Framework and a Proposed Process for Agreement Institutions in Addressing Emerging Issues in Great Lakes.

The board completed a thorough review of its own internal processes for fulfilling its role under the Agreement, and developed Procedural Guidelines encompassing its administrative processes and work planning, the conduct of board workshops and the preparation of the Biennial Report on the Priorities and Progress under the Great Lakes Water Quality Agreement.

During this biennial reporting cycle, the SAB held two public meetings: at the IJC Biennial Meeting in Milwaukee, Wisconsin, held September 1999 (#115) and during the IJC status assessment of the Niagara River AOC, held in Niagara Falls in November 2000 (#120). Both meetings reflect the ongoing practice of the board to encourage public involvement and interaction with the board.

Given the ease of electronic communication with the Internet, there are greater opportunities than ever for interested citizens to be aware of their Great Lakes science heritage and to benefit from its discoveries in terms of societal change. Scientific knowledge, however exciting, is necessary but not sufficient as a sole basis for actions, as it must be applied wisely by decision makers, and their actions must receive public support, in order for progress under the Agreement to be sustained. Great Lakes citizens are encouraged to follow the board activities and advice by visiting their web site at the IJC home page at http://www.ijc.org/rel/boards/sab/index.html or to contact the board secretary Mr. Peter Boyer at boyerp@ijc.windsor.org for further information.

Finally, the board would like to acknowledge the efforts of all of the non-board contributors to the report, many of whom participated as invited experts at board meetings, industry tours, workshops and work group meetings. They include: Annette Ashizawa, Seth Ausubel, Thomas Baldini, Thomas Barnard, Michael Basile, Judy Beck, Mary Lynn Becker, Matthew Becker, Ryan Bodanyi, Ken Bondy, Barry Boyer, Jim Brophy, Jean Burton, Tanya Cabala, Michael Campana, Richard Carrier, Alice Chamberlin, Mathew Child, Theo Colborn, George Costaris, James Cowden, Joseph DePinto, Krista Devine, Jim Drummond, Abdel El-Shaarawi, Rose Ellison, Thomas Emery, Rick Esterline, John Eyles, Sharon Fedman, Ralph Ferguson, Joel Fisher, Irene Gauthier, Matt Hare, Lucy Harrison, Jim Hartnett, Maureen Healey, John Heatley, Allen Heimann, Henry Henderson, Adam Hess, Paul Horvatin, Robert Huggett, Irene Ilia, John Jackson, Lin Kaatz Chary, Neil Kagan, Rimas Kalinauskas, Wilfried

Karmaus, Margaret Keith, Elaine Kennedy, Daniel King, David Kohoko, Horace Krever, George Kuper, Wendy Larson, Kathleen Law, Mike Lawson, Michael Leffler, Matthew Longnecker, Kim Lund, Kevin Lynch, Yang Mao, John Marsden, Barbara McElgunn, Errol Meddinger, Russell Moll, Penelope Moskus, Tom Muir, Michael Murray, William Muszynski, Susan Nameth, Melanie Neilson, Scott Painter, Dale Phenicie, Christian Pupp, Lisa Richman, Kristina Riggle, Amy Roe, Rick Sherrard, Ted Schettler, Gary Silverman, Saulius Simoliunas, George Spira, Doug Spry, Helen Tryphonas, Michael Twiss, Raymond Vaughan, Marcia Valiante, John Vena, Anita Walker, Reuben Warren, Bernard Weiss, Brenda Wheat, Jim Whitaker, Marj Williams, Don Williams, Autumn Workman, Michael Zegarec and Donald Zelazny.

2.2 HEALTH

2.2.1 Methodologies for
Community Health Assessment: Measuring Injury to Health

Scientific, Diplomatic and Political Responses to Great Lakes Pollution

For more than 90 years, the International Joint Commission has been assisting the United States and Canadian governments in preventing and resolving potential disputes concerning the use of the boundary waters between the two countries, from coast to coast, under the Boundary Waters Treaty of 1909. Because the International Joint Commission is a diplomatic organization that undertakes its work by examining issues that are under dispute, it is not only a political, but also a scientific organization. Many of its responsibilities concern water use, diversions, allocations and other aspects of water quantity. But in the past 40 years, there has been a growing emphasis on water quality, particularly through the studies leading up to and subsequent to the signing of the 1972 Great Lakes Water Quality Agreement by the United States and Canada. The Agreement itself was both a political and scientific response to the widespread reports of deteriorating water quality in the Great Lakes basin.

At the core of the water quality aspects of the Boundary Waters Treaty is an agreement that the two governments would not pollute the boundary waters to the injury of health or property on the other side. However, in the preamble to the Great Lakes Water Quality Agreement there is an acknowledgment by the governments that this has happened and that the boundary waters in the Great Lakes basin have been polluted to the injury of health and property on the other side.

Initial priorities and programs were oriented to documenting and controlling eutrophication and developing and implementing common water quality objectives. But in the 1960s and through the 1970s, Great Lakes scientists began reporting observations of gross effects of persistent toxic substances on the reproduction and development of avian wildlife and ranch mink. These observations influenced the renegotiation of the 1978 Agreement. In the 29 years since the signing of the original Agreement, there have been enormous advances in documenting this injury to many species and taxa of wildlife and in formally demonstrating the causal links to specific pollutants. It has, however, proven more elusive to document the injury to human health from exposures to pollutants. Initial cohorts were established in the 1970s to investigate exposures from consumption of Great Lakes fish and the first cohort to examine possible effects on infant development from maternal consumption of Great Lakes fish contaminated with persistent toxic substances (Fein et al. 1983) was established in 1980. By the end of the 1980s, there was such a remarkable concordance between the developmental effects that had been observed in wildlife and those seen in studies of infants that it lead to the successful formulation of the unifying hypothesis of endocrine disruptors as a mechanistic explanatory principle linking effects on reproductive, developmental, neurological, endocrine and immune processes (Colborn and Clement, 1992).

These scientific findings were politically influential in both countries. In the late 1980s, Health Canada instituted its Great Lakes Health Effects Program, and in the early 1990s, the United States, through the amendments to the Clean Water Act, mandated the Agency for Toxic Substances and Disease Registry (ATSDR) to fund epidemiological research through the Critical Programs Act. The ATSDR studies confirmed that fish consumption is the major pathway of exposure to persistent toxic substances, such as dioxin, polychlorinated biphenyls and mercury, and identified at-risk populations, including Native Americans and other minorities, sport anglers, the elderly, males and females of reproductive age, and fetuses and infants of mothers consuming contaminated Great Lakes fish (Johnson et al. 1999). In human studies, increasing levels of sport fish consumption have been associated with difficulties in conception for Michigan sport fish anglers (Courval 1999). ATSDR has funded research and other community-based studies have the ability to influence policy and public health practice, thereby

directly enhancing the health status of vulnerable communities through identifying at risk groups consuming Great Lakes sport fish and by disseminating outreach materials alerting the public about safe fish consumption.

Community Health and Areas of Concern

In the mid-1980s, the IJC expressed concern about the recurrent reports from the governments of locations around the Great Lakes where water quality was out of compliance with water quality objectives, and particularly in relation to pollution by persistent toxic substances. These locations became known as Areas of Concern (AOC) and were intended for special programs to restore water quality through the development and implementation of Remedial Action Plans.

Health Canada, through its Great Lakes Health Effects Program, undertook a research project to investigate whether the incidence rates of diseases were different in the 17 Canadian AOCs compared with the rest of Ontario and to generate hypotheses about whether these differences might be related to exposures from local sources of pollutants. The data and statistics in the reports were compiled from a national databases kept by Statistics Canada for selected health endpoints that might be related to pollution and included mortality, morbidity as hospitalization, congenital anomalies and birth weights. There was considerable nervousness within Health Canada about releasing the 17 reports after comments were received from the local medical officers of health and from officials in Environment Canada and in the Ontario Ministry of Environment. But in November 1999, the reports were released to the public, just before the Great Lakes Health Effects Program was terminated on March 31, 2000.

Partly as a result of the availability of the Health Canada reports, the International Joint Commission directed the SAB to examine methodologies for assessing whether human health effects of pollution are occurring in communities in the Great Lakes basin. The SAB's Work Group on Ecosystem Health developed a workplan for a Workshop on Methodologies for Community Health Assessment for Areas of Concern , which was held on October 4 and 5, 2000 in Windsor, Ontario. The primary concerns to the members of the work group were reliable interpretations of the health data and statistics for the Health Canada reports, and the apparent absence of any comparable data for the AOCs in the United States.

The Workplan

• Review the 17 Great Lakes Health Effects reports to prepare a synthesis of the health status within the communities in the Canadian AOCs, with particular regard to the health status in Windsor.

• Examine the processes being employed in these communities to respond to this new information, with a focus on the response in Windsor.

• Inquire whether comparable U.S. data can be obtained.

• Examine how to obtain data on more subtle quality-of-life indicators within the AOC communities, including such health endpoints as thyroid dysfunction, diabetes and immune dysfunction.

• Obtain indirect evidence from monitoring the incidence of endocrine effects on wildlife populations that are geographically distributed close to AOCs.

• Investigate the possible links between the incidence of disease and exposures to pollutants.

• Investigate the associated societal costs.

Several past activities had been undertaken by the IJC and the governments to gain understanding of the Great Lakes health issues at the community level. In 1992, the Work Group on Ecosystem Health held a workshop titled Our Community, Our Health: Dialogue Between Science and Community . In 1995, Health Canada published two documents titled Investigating Human Community Exposure to Contaminants in the Environment and A Community Handbook. In 1999, the Great Lakes Science Advisory Board held a Meeting to Assess Scientific Issues in Relation to the Effects of Persistent Toxic Substances in Relation to Lakewide Management Plans , and identified the need to assemble epidemiological evidence related to the effects on human communities. Much of the work undertaken has traditionally related to the incidence of cancer. With the recent publication of several papers documenting the experimental induction of a variety of effects from prenatal exposures to low doses of endocrine disruptors, there is now a priority need for communities to investigate whether these subtle effects are occurring among individuals within their communities.

Community health and injury to health from exposures to pollutants connote environmental epidemiology. Epidemiology has been undergoing a period of intellectual crisis and reassessment (Susser and Susser, 1996a, b; Pearce 1996). The crisis has been precipitated by the limits of the methodology for detecting significant relationships between the incidences of disease and exposures to putative risk factors. The reassessment has been prompted by a recognition of the need to reconnect the science of

epidemiology with the practice of public health in its ongoing attempts to intervene effectively with large scale increases in the incidences of certain diseases of uncertain etiology. In the quest for making definitive causal statements relating disease incidences to specific factors, to determine whether preventive interventions are feasible, there has been a recognition of the need to develop an eco-epidemiological approach by integrating molecular epidemiology, with its orientation toward mechanistic hypotheses, with individual and population epidemiology, and to cast these in a social, economic and political context (Susser and Susser, 1996b). There are particular challenges of applying this new approach to environmental epidemiology, with all the attendant difficulties of measuring, estimating or inferring exposures (Pekkanen and Pearce, 2001). The International Joint Commission, comprised of the dual attributes of politics and science, first applied an eco-epidemiological methodology to several Great Lakes case histories in 1989, at its first Cause-Effect Linkages Workshop (Fox 1991).

Within the Great Lakes basin, the political context of the implementation of the Great Lakes Water Quality Agreement has not changed significantly since the early 1980s when the shift to the political right resulted in a general withdrawal of environmental issues from the political agenda of both countries. The consilience of results of so many epidemiological studies, together with the extensive surveys of the concentrations of persistent toxic substances in the Great Lakes environment, and the research on the mechanistic processes, has created a strong dynamic tension with the maintenance of these existing policies formulated more than 20 years ago. The workshop was structured to explore this consilience and lay out the diversity of evidence and implications of not implementing the restoration provisions of the Agreement. The papers from the Workshop on Methodologies for Community Health Assessment for Areas of Concern are being prepared for publication in the journal of the National Institute of Environmental Health Sciences, Environmental Health Perspectives.

One of the critical health endpoints from exposures of communities to toxic substances concerns the effects on neurological development. In his keynote address at the Workshop on Methodologies for Community Health Assessment for Areas of Concern , Ted Schettler, (Schettler et al. 1999) of the Greater Boston Physicians for Social Responsibility, elaborated the evidence of the unique vulnerability of the developing brain to environmental agents, such as organochlorine compounds and metals, at exposure levels that have no lasting effect in the adult. This has been known for a long time, since there are even biblical prohibitions against drinking alcohol because of the long-lasting damage to brain development and function caused by prenatal alcohol exposure. Where registries have been established, the collection of health information on behavioural, learning and developmental effects has been significantly enhanced.

Recommendation

The SAB recommends the following to the IJC.

• Recommend that the Parties establish prospective and retrospective registries of neurological deficits to identify subpopulations at risk from exposures to developmental toxicants.

Health Effects in Canadian Areas of Concern:
17 Health Canada Reports

The 17 Health Canada reports on health data and statistics for diseases and conditions that might be related to pollution in the communities in the Canadian AOCs, represent an enormous database that potentially could contribute to the formulation of statements that could provide a rational basis for remedial action plans to restore environmental quality under the Great Lakes Water Quality Agreement and under the Canada - United States Air Quality Agreement. The reports with the data and statistics can be found at http://www.hc-sc.gc.ca/ehp/ehd/bch/bioregional/healthdata.htm. Before the reports can be utilized for this or any other purpose, such as selecting community health indicators, they needed to be interpreted. The objective of the IJC Workshop on Methodologies for Community Health Assessment for Areas of Concern was to focus on methodologies. The secondary objective was to provide detailed descriptions of the community health status within AOCs. The following is a brief review of the methodology that Health Canada used to compile the reports and of the methodologies used to interpret the database. Health Canada detailed the methods used in the study under the following headings: Assigning Standard Geographic Codes; Selecting Health Outcomes; and Gathering and Analyzing the Data.

Assigning Standard Geographic Codes

Each of the 17 AOCs was described using Standard Geographic Codes that contain provincial, Census Division and Census Sub-Division information, and that coincide with the Canadian process for collection of human health data. Each of the 17 reports contained detailed background information on: the study area and its population; methods used in the study for assigning Standard Geographic Codes; and gathering associated health data. The population within each area in 1991 was determined and compared, as a percentage, with the Ontario provincial population, which was 10,104,317. Data for the incidence rates for diseases and disorders for the populations in all the AOCs were age standardized by gender and comparisons made with the incidence rates for the rest of Ontario.

Selecting Health Outcomes

Health Canada selected approximately 70 categories of health endpoints, using the International Classification of Diseases (ICD 1992). Because this project was undertaken by the former Great Lakes Health Effects Program under the mandate of the Great Lakes Water Quality Agreement, there was an orientation toward selecting diseases and disorders on the basis that they might plausibly be linked to exposures to contaminants in the Great Lakes environment based on references in the published literature. Appendix I. contains the selected health outcomes based on the Ninth Edition of the International Classification of Disease (1992).

Gathering and Analyzing the Data

Population census data for the years 1986 and 1991 were accessed from the Demography Division of Statistics Canada to calculate mortality and morbidity rates on an age-specific and gender-specific basis. Mortality data were provided to Health Canada's Center for Disease Control by Statistics Canada and included information on the cause of death, reported by ICD code, the last location of residence based on the Census Sub-Division, and the sex and age of the deceased. Hospital separations data were supplied by the Canadian Institutes for Health Information, and included data on sex, age and residence, and the ICD code for the diagnosis for the main cause of hospitalization. Health Canada warned about some of the pitfalls of using hospital separations data and these concerns included multiple visits or transfers between or within hospitals. They also exclude visits to clinics, doctor's offices and outpatient departments. Similarly, there may have been difficulties in transforming residence information, based on a postal code or an Ontario Residence Code, into a Census Sub-Division. Further, these data for Ontario do not include Ontario residents who were hospitalized in another province.

These pitfalls were addressed in the Health Canada reports in analyzing the data and statistics. For example, in using the hospital separations data, Health Canada referred to morbidity rates rather than incidence rates. The data were initially analyzed using 19 age groups, but for simplification of reporting were combined into five age categories: all ages; 0-24 years; 25-44 years; 45-74 years; and more than 75 years. The age adjusted mortality and morbidity rates were compared with the rates for the rest of Ontario and ratios calculated comparing the local rates with the provincial rates. In the 17 reports, Health Canada included numerous references to aid in the interpretation of the data and statistics. In addition, several of the reports of the former Great Lakes Health Effects Program of Health Canada (1995, 1997, 1998) proved invaluable in the interpretation of the 17 reports on the health data and statistics particularly in relation to possible causal factors, including occupational exposures to chemicals and lifestyle factors such as alcohol and smoking.

The interpretation of the reports has proved to be a challenge, even for those with advanced degrees in public health. The challenges can be divided into: those associated with the diversity of health outcomes selected; useful ways of aggregating the endpoints into categories; and interpretation of the tables into comprehensible displays and narratives. In the interpretation of the reports, the following three approaches have been taken by the Work Group on Ecosystem Health and its contractors:

a. mapping of the distribution of the statistical significance of the ratio of the incidence rate for a particular health endpoint compared with the rate in the rest of Ontario;

b. extended narrative of the elevated incidences of diseases or conditions within a population in a specific Area of Concern and comparisons with other locations with similar population size and racial profile; and

c. statistical analyses to rank Areas of Concern on the basis of the aggregated severity of incidences of diseases and conditions and to select indices of environmental health.

Mapping

Based on the Health Canada data and statistics, Dr. John Eyles of McMaster University prepared maps of the distribution of incidence rates for some of the diseases and conditions compared with the rest of the Ontario. He concluded that the Health Canada findings present a complex picture of health outcomes in the Canadian AOCs. In some respects, he felt, such a complex picture presents a scientific maze for policy makers in the IJC and in the federal and provincial bodies. There is so much that is potentially significant that it may be difficult to know where to concentrate finite and limited resources. While mapping the Health Canada statistical data provides a useful and user-friendly lens on the issues, there do remain methodological issues connected with the critical data and well documented as a series of caveats in the Health Canada reports themselves. Further, while the rates are usually standardized to the provincial population, this may not be the best comparator for a binational body like the IJC. It is of central importance to Ontario-based policy makers in public health and environmental regulation. And while the identification of statistical significance is important, such an approach may indicate issues of low incidence or even low public health relevance, given current knowledge and priorities.

Dr. Eyles compared the standardized rates, for the selected health outcomes for the various AOCs, with the rates for the rest of the Ontario. Those outcomes in which the ratios of the standardized rates were statistically significantly above one, were identified, based on the argument that any above average deviation from the province is worthy of note. This technique identified, for example, the Detroit River Area of Concern for both men and women, and the Niagara River (women only) as worthy of further investigation, even at the ecological level. Thus, it may be worth doing some analysis of census data to see the associations between some health determinants and the identified health outcomes. In this way, it will be possible to account for, at the ecological level, some of the non-environmental determinants of health. Further work will enable a discussion of potential explanations of the associations between different health outcomes and environmental factors operating in the Canadian AOCs.

Extended Narrative: Focus on Windsor, Ontario

Marcia Valiante of the University of Windsor set the historic context of the bilateral concerns about transboundary air pollution at the Detroit River. Windsor and the outlying areas have had poor air quality as a palpable fact of life for the last 100 years. This region is also subject to greater incidences of certain health problems than other areas in Ontario. Investigating the link between these two facts is like assembling a jigsaw puzzle with many pieces missing and no picture to guide the task of assembling the available information about the state of the air in this region and comment on the relationship between local and transboundary sources.

Marcia Valiante noted that Windsor is uniquely located. It is a city of 200,000 residents (with another 100,000 in neighboring communities in Essex County), yet it has air quality that is generally worse than that of the largest cities in Canada. This is in large part because Windsor's air quality is closely linked to emissions from industry and transportation sources in Detroit and the rest of southeast Michigan, in the larger upwind context of the Great Lakes basin. But Windsor is not simply a faultless victim of another jurisdiction's neglect; local industry and transportation sources also contribute harmful emissions to the atmosphere.

She pointed out that air pollution is usually regulated on the basis of its health impacts due to direct inhalation. Standards in both the U.S. and Canada reflect this traditional concern with air emissions. More recently, other concerns have emerged. Studies have shown that there is a proportionately significant contribution from atmospheric deposition to the Great Lakes for certain persistent substances, such as PCBs and mercury, which may then build up in the environment to harmful levels. As well, many substances no longer emitted locally are still detected in the air and water due to atmospheric transport from distant sources. Finally, the atmosphere allows for the transformation of some substances into other compounds that can then have effects far downwind. This is the case with acid rain and smog. While understanding all of these interrelationships is important to a complete picture of air pollution's impacts, it is difficult to get that complete picture due to major information gaps.

Of the 17 Canadian AOCs, Windsor, Ontario seems to be the municipality ranked among the highest for incidences of diseases that might be related to pollution. Jim Brophy of the Occupational Health Clinics for Ontario Workers and Michael Gilbertson of the International Joint Commission used the Health Canada health data and statistics for the Detroit River Area of Concern to provide an extended narrative of the incidence rates of mortality and of cancer, of morbidity as hospitalizations, and of congenital anomalies. Mortality and morbidity rates from all causes were higher than in the rest of the province. Anomalously high rates of diseases included: various cancers; endocrine, nutritional, metabolic and immunity disorders; and

diseases of the blood and blood-forming organs; nervous system and sense organs; circulatory and respiratory systems; digestive system; genitourinary system; skin and subcutaneous tissue; musculoskeletal system and connective tissues; congenital anomalies and infant mortality. These incidence rates for most diseases were much higher than those in Hamilton, another industrial municipality in southern Ontario, suggesting that, in addition to a variety of local sources of industrial pollution from automobile manufacturing and use, there are transboundary sources of air and water pollution from Detroit that are potentially important causes of these health outcomes in the Windsor AOC. New United States and Canadian websites have been established that contain details of ongoing releases of toxic substances from industries and municipalities in Detroit and Windsor ( http://www.scorecard.org and http://www.scorecard.org/pollutionwatch/ , respectively).

This pilot project by Health Canada would seem to be a useful preliminary method of integrating human health concerns and of priority setting for the administration of the Great Lakes Water Quality Agreement and the Canada - United States Air Quality Agreement. For the detailed community health profile for Windsor and the comparison with Hamilton, Ontario, see Appendix I.

Further Statistical Analyses

The availability of electronic versions of the Health Canada data ( http://www.hc-sc.gc.ca/ehp/ehd/bch/bioregional/healthdata.htm ) and the provision of significant funding by Environment Canada for the development of indicators provides an opportunity to further analyze these databases. For example, ranking, cluster analysis, correlational and principal component techniques could provide better descriptions of the disease profiles in the most polluted areas and selection of particular diseases that might most reliably be used to indicate the environmental health status in a community exposed to pollutants.

Responses from the Medical Officers of Health

After Health Canada had prepared reports on the health data and statistics for each of the 17 Canadian AOCs to generate hypotheses about the possible role of pollution on the incidence of mortality, morbidity and congenital anomalies, the respective report was sent to the local medical officers of health for information. The SAB then contracted with Dr. John Eyles to undertake a brief survey of the medical officers' responses to these reports. In every instance, the reports were referred to staff of the public health units, many of whom found the reports useful and who assisted in interpreting the reports.

There were, however, several concerns raised with Dr. Eyles about the reports. They were described as being too technical and dense and there were concerns expressed that, despite the following caveats, they might raise more questions than answers.

• They would have been more useful if the reports had contained an analysis and summary of the findings.

• Some of the populations in the AOCs were small and therefore provided limited statistical power for assessing the significance of the data.

• The use of hospitalization data may reflect the styles of practice at the facility rather than underlying health problems.

• The geographic boundaries of the AOC may not correspond to those of the public health unit and thus the reports were of limited value for their responsibilities.

The survey indicated that there may be a difference of opinion between the medical officers of health and the members of the public within a community about whether pollution is an important determinant of health. The diseases and risk factors that have been the traditional concerns of the medical officers of health include smoking, nutrition, heart disease, respiratory illness and cancer, and that environmental pollution tended not to be viewed as a factor affecting public health. The public, however, tends to view pollution as more important in determining diseases, especially in industrial settings, but this was viewed by some medical officers of health as a product of the media. Apart from cancer, respiratory disease, especially asthma, and the exacerbation of other disease conditions, such as heart disease and diabetes, the medical officers of health tended to be unclear about any connection between human health and pollution, while waiting for new evidence that could change their minds.

There were no plans on the part of the medical officers of health to release the reports to the general public in their Area of Concern. While most of the medical officers of health recognized that the statistical and epidemiological caveats had been addressed, they believed that the reports could amplify public concern over health problems in the AOCs, given the statistical significance of the presence of so many health endpoints. They believed that the populations in their AOC fared no better or

worse than the other Ontario and Canadian populations, that public trust in all government institutions, including public health, is fragile, and that the reports could, when there is no solution, frighten the public.

Recommendations

The SAB recommends the following to the IJC.

• Recommend that the Parties establish institutional health structures at the local and regional level that can effectively investigate and respond to community health concerns that may be caused by chemical pollutants.

• Link human epidemiology to exposure data on air, water, sediments and biota in the preparation of future reports on Remedial Action Plans and Lakewide Management Plans.

Availability of Health Data in the United States

The Work Group on Ecosystem Health let two contracts to assess the availability of comparable health data for the eight Great Lakes states. One contract was with a student of Dr. David Carpenter at the State University of New York at Albany who collected information on the availability of health data in New York state. The second contract was with a student of Dr. Diane Henshel of the School of Public Health and Policy in Indiana University at Bloomington. The goals for these studies were to assess the availability of health data in the eight Great Lakes states: Illinois, Indiana, Michigan, Minnesota, New York, Ohio, Pennsylvania and Wisconsin. The specific aims were to:

1. identify the status of the development of health databases (i.e. health registries) and the collection practices in each state and in the municipalities designated as an Area of Concern;

2. identify the accessibility of the public health databases relevant to exposures to pollutants;

3. characterize the quality of each database; and

4. identify the barriers to database access and information on database quality.

In order to quantify data accessibility, the state databases were ranked on several parameters that affect the usefulness of the public health databases for research purposes. Overall, the New York public health databases and registries stand out among the eight states for completeness and accessibility to researchers. The Illinois and Wisconsin databases rank just below that of New York, followed by the databases for Michigan and Pennsylvania. What registries exist for public health databases in Indiana are difficult to access. Databases for the Great Lakes region of Ohio are not accessible, although the public health databases for the rest of the state are available. The Minnesota registries do not present information at an individual level, and are therefore of minimal use to researchers trying to correlate data across multiple endpoints. In sum, there is too little consistency in the collection and availability of public health information across the eight Great Lakes states, making it more difficult for researchers to compare and use public health data on a region-wide basis.

In New York state, the Statewide Planning and Research Cooperative System records diagnoses by International Classification of Disease (ICD) code for all patients admitted to state-regulated hospital facilities. There are six AOCs in New York, and three, the Buffalo River, Niagara River and 18 Mile Creek, are contiguous in western New York and have similar contaminants. These records can be utilized to compare the incidences of selected diseases reported in hospitalized patients who reside in zip codes that are within 15 miles of any of these sites to those of residents in the rest of New York state. Dr. Carpenter found significant elevations of disorders of the female genital tract, primarily due to endometriosis, and thyroid disease in women at all ages above 25 years, and significant elevations of diagnosis of ischemic heart disease and diseases of arteries, arterioles and capillaries in both men and women at some but not all age groups. The incidence of diabetes was significantly elevated in both men and women in the age range 25-44 years, but was significantly less in ages 55-74 years. Disease of the ovaries and testes were not different from the rest of New York. While many factors influence incidence of thyroid, genital and heart disease and diabetes, these results are consistent with the hypothesis that exposure to environmental contaminants through residence near polluted sites may be contributory.

There is a paradox between the need to access information on the health of individuals in order to operate a complex health care system with care that is accessible, efficient and of high quality and the need to protect the privacy of personal health care

information. In Canada, there has been an impressive record of protecting the privacy of an individual's health information while making data available to bona fide health researchers. There are, however, bills both in the Federal Parliament of Canada and in the Ontario Provincial Legislature to introduce new legislation that may make access to and use of health data more complicated. Individual consent may be required to use information in new ways that had not been contemplated when the data was originally collected. Stronger requirements for obtaining these kinds of consent have been incorporated into legislation in the United States and in Europe.

Recommendations

In the preparation of evidence of injury to human health from exposures to pollutants at the boundary, in accordance with Article IV of the Boundary Waters Treaty, the SAB recommends the following to the IJC.

• Recommend that the Parties facilitate the access of researchers to health information, while not compromising the rights of individuals to privacy and confidentiality.

• Recommend that the Parties make representations, with respect to pending legislation on the privacy and confidentiality of health information, to ensure that the capacity to monitor long term trends in pollutant-induced diseases and disorders is not jeopardized.

Biochemical Epidemiology

Glen Fox of the Canadian Wildlife Service, working at the biochemical and individual epizootiological levels, cited studies of Great Lakes wildlife species afflicted with thyroid and other endocrine disorders, metabolic diseases, altered immune function, reproductive impairment, developmental toxicity, genotoxicity or cancer, attributable to exposures to persistent organic pollutants, particularly PCBs, PCDDs and PAHs. The frequency and severity of these effects occurred in the most contaminated sites of Green Bay, Saginaw Bay, Lake Ontario, the St. Lawrence estuary, and, more recently, Lake Erie, some of which are Areas of Concern. Because these health impairments in wildlife resemble those observed with increased incidence in human subpopulations in one or more AOCs, wildlife could be used as a sentinel of the likely biochemical effects occurring in exposed human communities. These wildlife data were gathered as a result of academic research or research-based monitoring, rather than a formal effects monitoring program. While there are adequate long-term monitoring programs to document trends in concentrations of persistent toxic substances in the Great Lakes environment, there is no formal existing program for gathering long-term evidence for determining trends in the incidence and severity of their effects in wildlife. Such a program, using one or more sentinel wildlife species, would allow the Parties to the Great Lakes Water Quality Agreement to optimally use such information as a basis for decisions and policies regarding the effects of chemical exposures on human populations.

Recommendations

The SAB recommends the following to the IJC.

• Recommend that the Parties develop a coordinated binational monitoring program to determine the incidence of health effects in wildlife that have been attributed to exposures to persistent toxic substances.

Current human health and wildlife monitoring programs sometimes have levels of chemical analytical detection that are well above the concentrations at which biological effects occur.

• Recommend that the Parties monitor the chemical exposures of human and wildlife populations using limits of detection appropriate to the known toxicology of these substances.

Three papers explored the possibility of using the distribution of endocrine and immune diseases in human communities to identify small populations at high risk and indicate Areas of Concern. The requirements for diseases to serve this purpose include:

• agreement that the disease and markers can be employed as valid health indicators;

• the necessity that the disease have a short latency period in order to facilitate early detection;

• full clinical manifestation of the disease; and

• easy ascertainment from population-based investigations or from registries.

The first of these three papers was by Dr. Wilfried Karmaus concerning the use of thyroid disease. Although there is evidence that alterations in thyroid hormones are related to environmental exposures, thyroid disorders do not seem well suited to identify community exposures to thyroid active substances. While environmental exposures can produce effects within a short latency period, these are predominantly subclinical effects and may likely be undetected except in a special investigation. A few studies, occurring mainly in an occupational setting, have demonstrated a higher incidence of clinical thyroid diseases or related delays in neurological development in childhood, associated with environmental exposure. Thus, to identify potential health risks of thyroid active compounds in communities, epidemiological studies, including effect monitoring and human biomonitoring, are necessary. To overcome these limitations, Dr. Karmaus proposed the development of a network of exposed communities concerned about exposures. In addition to a representative national sampling, a network would provide assessments of exposures and health outcomes with different communities mutually serving as exposed and control groups. Residents of communities that participate in epidemiological studies are all too often subsequently neglected in the dissemination of information about the risks identified. Such a network could foster risk communication and prevention within communities where studies have been conducted.

The second paper, by Dr. Matthew Longnecker, concerned the etiology of diabetes within communities. The rates of both type 1 and type 2 diabetes mellitus have been increasing in the U.S. and elsewhere, and genetic factors account for less than half of the new cases. These observations suggest that environmental factors cause both type 1 and type 2 diabetes. Occupational exposures have been associated with increased risk of diabetes. In addition, recent data suggest that toxic substances in the environment, other than viruses or immunogenic dietary components, are associated with the occurrence of type 1 and type 2 diabetes. For type 1 diabetes, intake of nitrates, nitrites, and N-nitroso compounds have been associated with increased risk. Overall, however, the data were limited and inconsistent. With respect to type 2 diabetes, data on arsenic and 2,3,7,8-tetrachlorodibenzo- p -dioxin (TCDD) in relation to risk were suggestive of an association but were inconclusive. The occupational data suggested that more data on exposure to N-nitroso compounds, arsenic, dioxins, talc, and straight oil machining fluids in relation to diabetes would be useful. Dr. Longnecker concluded that although environmental factors other than contaminants may account for the majority of type 1 and type 2 diabetes, the etiologic role of several contaminants and occupational exposures deserves further study.

The third paper, by Dr. Helen Tryphonas, Health Canada, concerned diseases of the immune system. Experimental animal studies indicate that environmental contaminants can have adverse effects on several organs and tissues of the immune system. Such effects are known to lead to increased host susceptibility to microbial infections and compromised immunosurveillance mechanisms that are normally instrumental in the elimination of neoplastic cells and the prevention of autoimmune diseases. Evaluation of the potential risks that environmental contaminants pose to the human immune system is currently accomplished via extrapolation of experimentally derived animal data to human. This process requires that uncertainty factors, such as interspecies differences and genetic variability, be taken into consideration. This process would be more manageable if data similar to that derived from experimental animals were available for human populations exposed to environmental contaminants. In view of the continuous exposure of humans to environmental contaminants there is, presently, a pressing need to improve the process of risk assessment by enlarging the human data base. To generate the much needed data, one must first identify a set of clinically relevant endpoints. Such endpoints, when adequately standardized, should be incorporated into the design of prospective epidemiologic studies.

While all three endpoints have potential to be used in these ways, there is still a need for developmental work and, if they are validated, make them operational.

Issues of Economic Costs of Disease and of Equity

The Work Group on Ecosystem Health had an interest in the direct and indirect economic costs of diseases that might be associated with exposures to pollution. In the field of economics there is extensive work being undertaken on this topic. Tom Muir of the Ontario Region of Environment Canada chose the following four candidate health outcomes for analysis: diabetes; Parkinson's disease; hypothyroidism; and deficits in IQ.

The methods used involve the examination of several lines of evidence, including empirical, methodological and theoretical.

First, the literature was reviewed to determine the evidence leading to the published concerns that exposures to environmental agents, particularly persistent toxic substances, are plausible risk factors to children, and the chosen effects or outcomes. Second, literature was reviewed to gauge and assess the extent to which approaches and methodology to measure such financial and economic costs and impacts, in general, are developed, and the extent to which case studies on our chosen outcomes have been undertaken. Where such methodologies and case studies existed, they were either cited, adapted, updated or expanded. Third, where no existing studies were found that evaluate or cost any of the selected effects, primary data sources were searched for, and where possible, estimates were made using the existing methodologies and/or economic theory. Throughout this exercise, efforts were made to develop estimates not only for Canada or the Canadian jurisdictions, but for the United States. In some cases, there are cost estimates for other countries, in the literature, and these are cited for comparison.

Based on these several lines of evidence, there are indications of an enormous cost burden for these diseases or conditions in the United States, Canada and elsewhere. Tom Muir reviewed actual social and economic costs, constructed estimates of some costs from pertinent sources, and provided several hypothetical examples that are consistent with published evidence. Many detailed costs are estimated, but these are fragmented, and missing in coverage and jurisdiction. Nonetheless, the cumulative costs identified are potentially very large and undoubtedly extend into the billions of dollars.

In the United States, environmental justice, concerned with equitable representation of all affected persons, particularly low-income people and persons of color, in environmental and environmental health decision-making, has become a prominent political issue. Dr. Rueben Warren of the Center for Disease Control and Prevention drew attention to the Presidential Executive Order 12898, on environmental justice, which encourages all federal agencies to ensure environmental justice in their policies, regulations and programmatic activities. People in communities across the country experience adverse living and health conditions due, in part, to the limited implementation of the Executive Order. Adverse health conditions associated with environmental toxins are a major problem expressed by environmental justice advocates and people living in close proximity to toxic and hazardous waste sites. They perceive a disproportionate burden of disease, dysfunction and premature death due to chemical contamination. They also believe that public health officials have not adequately responded to their health concerns. Dr. Warren discussed the history and current activities of the environmental justice movement and described human health threats posed by toxic and hazardous wastes. He reviewed strategies for collaboration between environmental justice advocates and public health officials that, if implemented, should enhance the missions of both groups. More importantly, it is expected that through increased basic and applied research and public health advocacy and practice, the well-being and quality of life of low-income and people of color will improve.

Dr. Bernard Weiss of the School of Medicine and Dentistry at the University of Rochester, contrasted the ethical requirements for conducting experimental studies on drugs in humans with those for populations exposed to environmental contaminants. Drug testing in humans is governed by a body of principles whose main tenets have evolved over the past few decades. Three of these tenets provide the foundations for judging the ethical adequacy of such an experiment. One addresses the question of who receives the benefits of the research and who bears its burdens. A second requires that the research maximize the potential benefits to the subjects and minimize the risk of harm. The third concerns the source of guidelines for informed consent and requires that subjects enter into the research voluntarily and with adequate information. By contrast, unlike research conducted to evaluate drugs, those responsible for environmental exposures to potentially toxic chemicals neither survey those exposed for their consent, nor provide an appropriate calculus for measuring risks and benefits. Dr. Weiss concluded that a process of ethics assessment may need to be incorporated, as a new element, into risk characterization, especially for exposure to developmental neurotoxicants, where the risk-benefit incompatibility between different populations can be so striking.

Contrasting the Scientific Approach and the Public Health Approach

The Great Lakes Science Advisory Board noted that there has been a significant debate recently about decision making in relation to the protection of public health. Mr. Horace Krever headed a commission into a major public health disaster in Canada from the presence of HIV and hepatitis-C in the blood supply. The findings from the judicial investigation of the blood supply disaster may have important lessons for those involved in policy making under the Agreement. Specifically, pollution of the Great Lakes with persistent toxic substances may present several analogous situations.

Mr. Krever pointed out that a chief factor leading to the infection of so many people with a deadly virus was the influence of the traditional way of thinking of the scientists, when what was more appropriate was the method of the public health practitioners.

He noted that there are essential differences between the scientific approach and the public health approach. The former was characterized by a refusal to accept that the illness could be spread by blood until Koch's Postulates had been satisfied and this lead to lengthy and undue delays in introducing the screening of blood donors and the subsequent testing of blood donations. The absence of definitive proof of a link between AIDS and blood transfusion was consistently used as a justification for maintaining the status quo . Strong action to reduce the risk of AIDS should not have required conclusive evidence. If there was even a possibility of transmission of the virus via blood, there was, above all, a moral and legal obligation to protect the blood recipient. Where there is reasonable evidence of an impending threat of public health it is inappropriate to require proof of causation beyond a reasonable doubt before taking steps to avert the threat.

Mr. Krever noted that, in environmental matters, this precautionary principle has become part of a number of international treaties and declarations, including the Second and Third International Conference of the Protection of the North Sea concerning ocean dumping. The application of the precautionary principle is not problem free. First, on some occasions, it will turn out after the fact that precaution was, with the benefit of hindsight, not necessary, and moreover, was costly. The second problem relates to the application of cost-benefit analysis that should not be a deterrent in the application of the precautionary principle. Risk management is defective if it protects only the risk creators and permitters and not also the person suffering the harm when inevitably the risk accrues. For these circumstances, there should be no fault compensation for victims of the harm created by the risk. The compassion of the society can be judged by the measure it takes to reduce the impact of the tragedy on its members.

Mr. Krever concluded by posing several questions. What should we as a society do about polluters, and about pollution? Should we shut down the polluter's business and so create unemployment by making them undertake preventive measures that are so cost prohibitive that their businesses become unprofitable? Or do we simply warn consumers of the products poisoned by the pollution? More philosophically, what right do we have, as the current and very temporary trustees of the environment, to decide, even for apparently sound reasons, to permit the destruction of land and water?

Conclusions

Epidemiological data on incidences of mortality, morbidity as hospitalization, congenital anomalies and infant mortality are useful to detect gross differences in rates between communities that might be related to pollutants. On the basis of epidemiological data and statistics on health endpoints that might be related to exposures to pollutants, there are several Areas of Concern that have elevated incidences of diseases. Similarly, research on wildlife populations for exposure and disease has provided exposure and effects information that is directly relevant to research on human health.

Local health authorities have well developed institutional structures for preventing diseases caused by microbiological agents and for investigating and responding to disease outbreaks. In contrast, there is a generalized reluctance at all levels of government to detect and publicly acknowledge anomalous incidences of diseases and disorders within communities that may have been caused by chemical pollutants, and there seems to be an apparent aversion to establishing the necessary institutional health structures at the local and regional levels that can effectively investigate and respond to community health crises that may be caused by chemical pollutants.

There is growing evidence of an increase in learning, behavior and developmental problems in the North American population. Exposure to persistent pollutants may contribute to this situation by their actions on physiological functions, and in particular the neuro-endocrine axis. While these may not be readily apparent in the individual, they can have important effects on a population level. There is a priority to address these questions through:

• tracking, monitoring and surveillance using registries, hospitalization, school performance records, and use of pharmaceutical utilization data; and

• integration of data bases on exposure assessment, health, and academic and behavioral performance indicators.

Research Needs

The SAB has identified the following specific research needs concerning community health:

• to examine the subtle decrements in neurofunctional capacities, immunological, hormonal and reproductive functions in

relation to exposures in adult and child populations;

• to determine whether subtle functional decrements in wildlife reduce survival, growth and population numbers of wildlife;

• to use the new information on the human genome to evaluate how genetic susceptibility explains human responses to environmental pollutants and help identify the more sensitive sub-populations that may need additional consideration in the development of regulations and regulatory actions;

• to investigate interrelations between exposure and the socio-economic environment with a view toward better policy making;

• to develop methodologies that will allow for better integration of information from exposure, health and social data sets, and to develop more effective means of applying animal experimentation and wildlife studies to humans; and

• to assess the impact of the biological effects resulting from chemical exposure on community well-being.

2.2.2 Protecting Human Health from Exposure to Contaminants in Great Lakes Fish

Introduction

Fish in the Great Lakes basin are contaminated with a variety of persistent toxic substances and this raises concerns about the suitability of fish as food. In the past 30 years, agencies responsible for public health and for fisheries have had to respond to the challenge of maintaining the recreational use of the Great Lakes fisheries resources while protecting public health from exposures to persistent toxic substances. Governments have responded to this challenge by publishing advisories on the consumption of Great Lakes fish. For example, in Ontario, there is a Guide to Eating Ontario Sport Fish , published and distributed annually to anglers, particularly through the Liquor Control Board of Ontario and outfitters selling fishing licenses. For the past decade, the Council of Great Lakes Governors has convened a Great Lakes Fish Advisory Task Force to develop a common protocol for an advisory for contaminants in fish across all the Great Lakes states. Fish consumption advisories have been based on the estimation of `safe' concentrations of each contaminant based on toxicological data. Toxicology experiments have traditionally been undertaken at high dosage concentrations, and the safe concentrations have been estimated by applying safety factors to the dosage at which there was no effect observed. It has been assumed that the public would be protected as long as the concentrations in the Great Lakes fish that were being consumed were below these concentrations. As the IJC pointed out in its Tenth Biennial Report on Great Lakes Water Quality, fish consumption advisories are not a guarantee of safety and for more than the past 30 years, fish advisories have generally become more restrictive as knowledge increased and more sensitive endpoints were reported from human health research.

Risk Assessment and Limitations

In 1983, the U.S. Environmental Protection Service introduced a formal risk assessment process that has been widely applied to the setting of `safe' concentrations for consumption of Great Lakes fish. The current advisories are based on a series of assumptions, including:

• the risk can be defined;

• the critical variables to be managed can be identified; and

• specific actions and techniques are available and effective for achieving management goals.

Most Great Lake jurisdictions include consideration of the species consumed, intervals of consumption, quantities consumed, cooking methods, location caught, size of fish and exposure of sensitive populations. There is, however, always a danger that current advisories do not reflect the latest research.

For example, in 1990, the endocrine disruptor hypothesis was proposed (Colborn and Clement, 1992; Colborn et al. 1993). The thesis is that structural and functional developmental processes, such as differentiation of the reproductive anatomy and of neurological and immunological processes, is under the control of a large variety of chemical messengers. These include, not only the traditional hormones from glands, but also various growth factors, interleukins and cellular receptors, collectively known, for the purposes of this hypothesis, as the endocrine system. The mechanism of control action of these chemical

messengers is through interactions at receptor sites and turning specific genes on or off at predetermined periods of development. The products from these genes in turn affect other cellular processes, such as cell division or cellular differentiation. The concentrations at which these chemical messengers operate are at fractions of a trillionth of a gram.

Certain natural and man-made chemicals can interfere with the production, transport and metabolism of these chemical messengers, or they can mimic or block the chemical messengers at the receptor sites. These interferences with the endocrine system can result in irreversible alterations to a wide variety of developmental processes, including the reproductive anatomy (Gray et al. 1999), the neurological (Colborn et al. 1998) and immunological (Voccia et al. 1999; Colborn 1995; Porter et al. 1999) systems. These natural and man-made endocrine disruptors are at concentrations much higher than the concentrations of the chemical messengers that control developmental processes. They are not necessarily bound to carrier proteins that moderate the endocrine activity of the normal hormones. In traditional toxicology, the effect becomes more pronounced as the dose or concentration of the compound increases yielding a monotonic relationship. However, the dose-response relations with endocrine disruptors tend to be non-monotonic.

Though this is a well-established phenomenon in endocrinology, it has not been a general consideration in traditional toxicology. On October 10-12, 2000, at the request of the U.S. Environmental Protection Agency, the National Toxicology Program and the National Institute of Environmental Health Sciences organized and conducted a scientific peer review meeting to evaluate reported low-dose effects and dose-response relationships for endocrine disrupting chemicals. The participants were from a variety of interests including industry and environmental non-government organizations. For this meeting, `low-dose effects' referred to biological changes that occur in the range of human exposures or at doses that are lower than those typically used in the U.S. EPA's standard testing paradigm for evaluating reproductive and developmental toxicity. The Statistics and Dose-Response Modeling Subpanel analyzed the data for 38 studies prior to the meeting and provided its analyses to the four other subpanels on: Bisphenol A; Estradiol and other estrogens; androgens and antiandrogens; biological factors; and study design.

Based upon presentations by the individual subpanels and general discussions during the plenary session, preliminary conclusions from this peer-review meeting include the following.

• Low-dose effects have been clearly demonstrated for estradiol and some estrogenic compounds. For example, low-dose findings for nonylphenol and the phytoestrogen genistein include effects on the immune system and on neurological structure.

• Effects of antiandrogenic compounds have been demonstrated for some endpoints and the dose-response curve appears linear to the lowest dose tested; however, it was noted that the available studies were not designed to evaluate low-dose effects as defined for this review.

Workshop participants identified areas for additional research that would clarify uncertainties about the occurrence of low-dose effects and better characterize those observed effects. These include using pharmacological and genetic approaches to determine mechanisms of action and to characterize dose-response relationships, characterizing response longevity from gestation through adulthood, evaluating long-term health outcomes, investigating the basis for immune system effects, and determining the impact of variations in endogenous hormone levels. The implications of these irreversible low-dose phenomena, particularly for community health, are far reaching and the main consideration is that very low concentrations of these substances cause endocrine disruption, and humans should not be exposed to them. The final report of the National Institute of Environmental Health Sciences' Endocrine Disruptor Low-Dose Peer Review has been released for public comment and is posted at

http://www.ntp-server.niehs.nih.gov/htdocs/liason/LowDosePeerFinalRpt.pdf.

There is a wide array of documented effects in humans, wildlife and in experimental animals in the laboratory. In several populations, there have been increased rates of testicular and prostate cancer, cryptorchidism and hypospadias, and decreases in sperm quality (Toppari et al. 1996; Swan et al. 1997). Recent epidemiological studies have shown that women who ate Lake Ontario fish for seven years prior to pregnancy had shorter menstrual cycles (Mendola et al. 1997) and maternal consumption of Lake Ontario fish for three to six years was associated with a reduced biological capacity to reproduce (Buck et al. 2000). Similarly, an epidemiological study of anglers and their families found an association between the amount of sport-caught fish consumed by males and a delay in conception (Courval et al. 1999). Consumption of contaminated fish from the St. Lawrence River has been associated with a decline in short-term memory, attention and fine motor skills in adult fisherfolk (Mergler et al. 1997; 1999). Similar findings have recently been published that show the impact of PCBs and other fish-borne contaminants on intellectual functioning in older adults who ate more than 24 lbs. per year of sport fish caught in Lake Michigan (Schantz et al.

2001). There are many chemicals in Great Lakes fish that have been shown to be endocrine disruptors, including DDT and metabolites, PCBs and dioxins. More particularly, there are increasing Great Lakes concentrations of brominated organic flame retardants that are endocrine disruptors, such as polybrominated biphenyl ethers (Darnerud et al. 2001). In addition, there are many modern pesticides in use that are endocrine disruptors and to which the population in the Great Lakes basin is exposed.

Of particular concern are the effects on the developing brain. For example, in a cohort of infants established in 1980, maternal consumption of Lake Michigan fish prior to and during pregnancy was associated with the following effects associated with exposure in utero to PCBs: poorer performance in tests of visual recognition memory in infants (Jacobson et al. 1985); poorer verbal and numerical memory in four year old children (Jacobson et al. 1990); and a loss of more than six IQ points by the time the children had become 11 years old (Jacobson and Jacobson, 1996). Ten years later, the results were replicated in another cohort of infants established in 1990, some of whose mothers had eaten Lake Ontario fish prior to pregnancy (Lonky et al. 1996). The effects were most pronounced in infants who were most exposed to the higher chlorinated PCBs, and included an inability to habituate to unpleasant events (Stewart et al. 2000). Subsequent studies at six and at 12 months showed poorer performance in intelligence tests in those children with the highest PCB levels (Darvill et al. 2000). Though these neurological effects are statistically robust, there have been questions about whether they are biologically significant since the shift downwards in scores is within two standard deviations of the mean and thus not outside the lower 95 percent confidence limit for the population. Should these statistically significant losses be regarded as adverse biological effects? It would seem that no community or civilization, and particularly one that is so highly dependent on the collection and processing of information, can afford the systematic loss of neurological functioning (Weiss 1997).

Studies of several different human populations exposed to PCBs before birth, suggest that prenatal exposure results in a reduction of overall IQ of six to eight points, and other studies show motor and sensory decrements upon pesticide exposures. While this may not be a sufficiently large decrement so as to greatly influence any single individual's lifetime productivity, on a population basis this may exert very important societal effects. Most human populations are never exposed to a single toxic substance, and indeed Great Lakes fish contain many different chemicals. However animal studies of single chemicals, such as PCBs, methyl mercury and persistent pesticides, alone or in combination, or animals being fed contaminated fish, all confirm alterations in brain function consistent with reduced intelligence, a shortened attention span and increased frustration.

The new information about endocrine disruptors poses several questions in relation to traditional risk assessment. Can `safe' concentrations of these compounds be determined from traditional toxicology testing at high doses? Did the application of traditional toxicological approaches to derive `safe' levels for compounds, that are now known to be endocrine disruptors, lead to injury to human health even in those who followed the fish consumption advisories? Where there is a possibility that injury may occur, is it essential, in the interest of public health and safety, that responsible parties be identified and empowered to take appropriate direct action to avoid, prevent or mitigate future harm? Might abrogation of this responsibility, by responsible parties in matters related to public health protection, result in findings of liability, negligence and even criminality by the courts or in official inquiries, such as the Krever Commission? Where the consequences are serious, would decision makers be held accountable for adopting policies, programs and practices that were not risk averse, and that promoted safety, instead of precaution? Are the scientific uncertainties of such a degree that the general public should be warned not to eat any fish from the Great Lakes or the St. Lawrence River, until remedial actions have reduced the concentrations to levels at which effects on the endocrine system and on developmental processes do not occur?

Risk Management

Fish consumption advisories have been described (Knuth 1995) as a "risk management tool designed to inform fish consumers about how to minimize exposures to chemical contaminants." Other risk management tools have been implemented in the past, including closure of commercial fisheries and the mid-1970s prohibition by New York state of the possession of certain sport fish from Lake Ontario. This last means was met with such flagrant violation of the law that the regulation was rescinded within months of proclamation. The objective in developing fish consumption advisories has been to give people enough information to make their own decisions about how much Great Lakes fish to eat and their likely exposure to persistent toxic substances through consumption of fish.

The development and implementation of the risk management program for consumption of fish from the Great Lakes is based on the following premises:

• the principal route of human exposure to persistent toxic substances in the Great Lakes region is through the food chain;

• there is a significant body of scientific research related to understanding human health effects from eating Great Lakes fish;

• implementation of remedial work will lead to declines in the concentrations of persistent toxic substances in fish; and

• fish consumption advisories must be based on assessing risk based on the most sensitive end points and to the most sensitive subpopulations.

There is a significant body of evidence that, for example, neurological effects from prenatal exposure to PCBs may be the most sensitive endpoint, and women of child-bearing age and children may be the most sensitive subpopulations. For these reasons, risk management, in the form of fish consumption advisories, is not necessarily uniform for the entire population. Based on scientific research, risk managers might communicate different advice to different age groups ranging from unlimited fish consumption, to complete avoidance.

Risk Communication

The success of any risk management approach depends on effective risk communication. This, in turn, requires that there is not only agreement on a clear message to be communicated, but also appropriate methods for communicating that message. In the past, the states and provinces each issued fish consumption advisories applicable to the fish caught in their jurisdictions. Because of differences in the approaches used by the jurisdictions to calculate acceptable exposures, the resulting advice might differ between jurisdictions. In the mid-1980s, there was a growing concern that advice in some jurisdictions was less protective of the public and sensitive subpopulations and this lead to confusion for the general public. In an effort to address this inconsistency, the Council of Great Lakes Governors convened a Great Lakes Fish Advisory Task Force to develop a common protocol for an advisory for PCBs across all the Great Lakes states. However, this effort took nearly 10 years of scientific review and negotiations before completion in 1995, and no other common advisories have been developed for any other persistent toxic substance.

There has been extensive social science research undertaken on the effectiveness of communication of the fish consumption advisories and much of this has been reviewed by Grondin and LaRue (http://www.hc-sc.gc.ca/ehp/ehd/catalogue/bch_pubs/great_lakes_part2.pdf). For example, early work among New York state anglers indicated that those who were older, more educated, higher income, male and white had a higher awareness of the advisories (Connelly et al. 1993). Overall knowledge on the adverse health effects of fish consumption was high, though knowledge of the recommendations in the advisories tended to be inaccurate. While dissemination of the information through brochures distributed with fishing licences is an effective means of communicating the message to certain groups, there are many subpopulations, such as women and certain ethnic groups, who were not being reached (Knuth 1995). Programs have been developed to communicate more effectively with these groups. A more recent survey has shown the scale of Great Lakes fish consumption (Tilden et al. 1997). In the eight Great Lakes states, about 4.7 million people eat Great Lakes fish, and women accounted for 44 percent of these Great Lakes fish consumers. This study confirmed the results from the previous work, that women tended to have a poor awareness of the advisories, suggesting the need for special risk communication strategies.

There are difficulties in obtaining similar statistics for Canadian angling in the Great Lakes, because no comparable random survey has been undertaken. There is one estimate from the mid-1980s that 37 percent of the general population of Ontario (about 10 million people) participated in sportfishing an average of five times per person per year (Usher et al. 1987). A survey by the Department of Fisheries and Oceans (1997) showed that there were 1,928,568 licenses issued in 1995, of which 1,342,567 were to Ontario residents. Instead of a random sampling approach to estimate the number of anglers in the province, Health Canada (1997), through its former Great Lakes Health Effects Program, undertook a Fish and Wildlife Nutrition Project as a series of surveys of anglers at the following five Great Lakes Areas of Concern: Metro Toronto; St. Clair River; Detroit River; Hamilton Harbour; and the Niagara River (http://www.hc-sc.gc.ca/ehp/ehd/catalogue/bch_pubs/monograph.pdf ).

There is no specific requirement that the fish consumption advisories reflect current research results. While public and private efforts throughout the basin promote the consumption of Great Lakes fish, research undertaken in the past decade continues to raise serious public health questions about the harmful effects of exposures to persistent toxic substances. While these forces tend to minimize the risk detailed under current fish advisories, new research suggests that many of the current advisories are set at much too high a level, and do not adequately protect public health. If these questions escalate among the public into serious

concerns about health, there is the potential for widespread dissatisfaction with the fish consumption advisory approach. As understanding of the toxic effects of exposures to persistent toxic substances increases over time, a transparent and accountable process for updating the advisories with current scientific understanding of health effects is needed.

In the case of other highly exposed groups, such as ethnic and native subsistence consumers who choose to exceed or ignore the recommendations of the advisories, they assume a risk that is accordingly higher than would otherwise be the case if the advisory were to be followed. In order to reduce the risk to this group to within a no observable effects level assumed by the advisories, it is necessary to improve communication and understanding, sufficient to support a modification of behaviour that respects the advisory. Recent social research is only beginning to identify the important elements of a risk communication strategy to better convey information to support improved personal decision making.

Because of the serious public health implications already acknowledged from consuming Great Lakes fish, and the complex scientific issues related to the significance of new findings from the latest scientific research on human health effects, the SAB recommends that the task of reviewing and developing a uniform, protective advisory be undertaken on a scientific basis by a third party, with adequate human and financial resources provided by the governments to achieve a rigorous and credible result.

Recommendation

Accordingly, the SAB recommends the following to the IJC.

• Recommend to the Parties that the U.S. National Academy of Sciences and the Royal Society of Canada convene a binational committee to develop a uniform and consistent protocol to protect human health from contaminants in Great Lakes fish.

Given insufficient progress on this matter to date, the conduct of such a study should give consideration, but not necessarily be restricted to the following issues:

• the application of the latest health effects research in risk assessment calculations and the development of a process to include new findings;

• the use of a safety factor above and beyond the risk assessment calculation to ensure protection for the range of human responses to toxins, for potential synergistic effects of multiple toxic exposures, for effects not yet fully quantified, and protection of subpopulations at greater risk than the target population, particularly children;

• the improvement of risk communication efforts, particularly to protect sensitive populations, for example, subsistence anglers; and

• the identification of risk communication challenges, including the key factors related to perception and awareness that affect changes in attitudes and fish consumption behaviour.

2.3 WATER QUALITY ASSESSMENT AND REPORTING

2.3.1 Review of Annex 1 of the Great Lakes Water Quality Agreement

Introduction

In fall 1999, the Work Group on Parties Implementation began work to obtain newly released data from U.S. EPA's Lake Michigan Mass Balance (LMMB) project and compare these data to Specific Objectives in Annex 1 of the Agreement.

Review of the available data showed that only two of the compounds included in the LMMB project, mercury and trans-nonachlor (a component of chlordane), are comparable to Specific Objectives in Annex 1. More importantly, the data revealed that concentrations of these two compounds in the open lake water column are approximately three orders of magnitude below the Specific Objectives.