On 16 October 1996, a malfunction at the Swan Hills Special Waste Treatment Center (SHSWTC) in Alberta, Canada, released an undetermined quantity of persistent organic pollutants to the atmosphere, including PCBs, PCDDs and PCDFs. The circumstances of exposure are detailed in Part 1, Background and Policy Issues. An ecologically based, staged health risk assessment was conducted in two parts with two levels of government as sponsors. The first, called the Swan Hills Study, is described in this part, which was conducted by the Government of Alberta to evaluate the human health risks, primarily by determining contaminant levels in wild game and fish and in serum of residents of the area to reflect body burden. A diet and activity survey was conducted by telephone to determine and to inform an initial advisory on consumption of country foods. A subsequent evaluation, called the Lesser Slave Lake Study, focused exclusively on Aboriginal residents in the area and is presented in Part 3 of this case study. Because this is a case study and not a research report, the findings are presented as they became available at the time of the study.

## KEY MESSAGES

• Ecosystem fate and mobility studies (ecotoxicology) can be used to anticipate future human health risks.

• In the real world, studies in which decisions are necessarily based are not always complete or performed as they might be for research studies.

• Persons living traditional lifestyles, such as Aboriginal communities in northern Canada, might be at higher risk for certain types of pollution than urban residents; the risk is specific to the situation.

• Food consumption advisories are difficult public health decisions with many tradeoffs.

## INTRODUCTION

Human health risks associated with environmental release of persistent organic pollutants (POPs) depend on the ecological behavior of the compounds and pathways by which the POPs reach human populations. Assessing the risk to human or animal populations requires knowledge of the release, transport, and fate of POPs in the environment called “ecotoxicology”. Human risk can be inferred from the delivery to the human receptor along ecological pathways that represent routes of exposure.

In 1996, an incident at the Swan Hills Special Waste Treatment Center (SHSWTC, now called the “Swan Hills Treatment Facility”) in the town of Swan Hills, province of Alberta, resulted in uncontrolled emissions of polychlorinated biphenyls (PCBs), polychlorinated dibenzo-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and other organochlorines [1]. The background and history are provided in Part 1 of this case study [2]. We report here a human health hazard assessment that was based on contaminant levels in which ecological pathways of bioaccumulation and bioconcentration leading to human consumption were studied. By tracking the behavior of POPs through the ecosystem, educated judgments could be derived regarding potential future risk. Health advisories based on these judgments reflected a risk management strategy that took into account the movement of POPs through the ecosystem prior to widespread exposure of the resident population.

In this paper, we summarize the results of individual studies published in detail elsewhere [1, 38] and describe the application of the information derived from the data in each study into the health advisory process. We discuss problems associated with modifying the risk management strategy for the population presumed to be at greatest risk. Names of provincial agencies are given as they were in 1997; most have been changed since but their areas of responsibility are obvious from their names.

## BACKGROUND

The SHSWTC is a purpose-built facility owned by the provincial Government of Alberta which began operation in 1987 to receive and dispose potentially toxic wastes safely. At the time of the incident, SHSWTC was owned by the Province of Alberta and operated by Chem-Security, Ltd. The facility was sited on a relatively remote plateau in northern central Alberta in part to isolate it from population centers and in part because the community of Swan Hills invited it, as a means of economic diversification. Further details on the history of the plant are provided at the end of this case study, when policy is considered.

Central to the waste treatment process for organochlorines is a high-temperature incinerator complex designed to destroy organic materials contained in the liquid, solid and sludge waste received at the facility. This consisted, at the time, of a rotary kiln and two rocking kiln incinerators. Combustion by-products are scrubbed to remove particulate matter and acidic gases prior to being discharged to the ambient air through the stack. The incident (actually two related incidents occurring close together) involved a breach in containment in the pyrolysis operation that vented through a defect in a heat exchange manifold and bypassed the scrubbing battery [3].

An environmental monitoring program was undertaken to assess the impact of the release on local vegetation, wildlife and human food sources [5, 9]. Concern that the release of these potentially toxic compounds might affect human health in the area arose immediately.

On 16 October 1996, a malfunction in the incinerator at the SHSWTC resulted in a leak. The leak vented to the outside and released quantities of organochlorine compounds, including dioxins, furans and PCBs. The plant had by then treated 27 million kilo grams of hazardous waste, much of it contaminated with PCBs, during 1996 alone.

Polychlorinated dioxins and furans and PCBs, together with a number of discontinued pesticides not present in quantity at the SHSWTC, are environmentally “persistent organic pollutants” (POPs), as defined by the Stockholm Convention on Persistent Organic Pollutants (2001). Because these compounds resist degradation in the environment (either by physical causes or metabolism by organisms) and tend to biomagnify with ascending trophic levels, PCDDs, PCDFs and PCBs became the principal focus of concern for ecotoxicology and human health.

### Ecosystem Movement

Concern over the distribution and extent of contamination resulting from this release led to preliminary testing of wildlife harvested near the facility by the facility operator in December 1996. Levels of PCBs, PCDDs and PCDFs were elevated in various environmental samples including spruce needles [3], snow pack [4], surface water sediment [4], deer and moose [6] and brook trout [7].

Studies conducted by the colleagues at the University of Alberta and Environment Canada (see Acknowledgments, at end) demonstrated increased levels and congener enrichment of PCBs in snow and lake sediments within 3 km in a southeast direction from the plant, in the direction of the prevailing winds documented on the day of the incident. (These are reported as mass per unit area rather than volume, because the depth of the snowpack and sediment column vary with location.) PCB concentrations in the snow peaked at 2,200 ng/m2 at a point just under 2 km east of the plant when compared to 75 ng/m2 to the west of the plant. Sediment from Chrystina Lake, 1 km east of the site, demonstrated a very high peak accumulation rate of 82 μg/m2/y, occurring in sediment deposited in 1997, following the incident. The PCBs present in the sediment were enriched in heavier penta-, hexa- and heptachlorobiphenyls, which are known to have short atmospheric residence times. Similarly, measurements in early 1997 demonstrated that PCB levels were increased in conifer needles downwind of the plant, peaking at 285 ng/g at 0.7 km to the east, compared to generally 10 ng/g to the west or east-southeast. These studies confirmed that POPs had been released from the plant and had entered ecosystem pathways.

Evaluation of the risk to human health therefore required knowledge of the stage and magnitude of migration of POPs along pathways specific to the boreal forest ecosystem. Traditional means of evaluating the hazard by assessing distribution by air and water would not provide a reliable exposure projection (see Figure 1 for ecosystem pathways).

FIGURE 1.

Schema of the ecotoxicology assumptions of the Swan Hills study, showing probable pathways. Anticipated ecosystem migration of POPs from the Swan Hills Special Waste Treatment Center. It was recognized early that contamination of groundwater was not a principal concern for human health. Potential exposure of residents was driven by consumption of fish and meat from deer and moose in the area, because of bioconcentration.

FIGURE 1.

Schema of the ecotoxicology assumptions of the Swan Hills study, showing probable pathways. Anticipated ecosystem migration of POPs from the Swan Hills Special Waste Treatment Center. It was recognized early that contamination of groundwater was not a principal concern for human health. Potential exposure of residents was driven by consumption of fish and meat from deer and moose in the area, because of bioconcentration.

### The Anomalous Ungulate and First Advisory

Then an unexpected anomaly was discovered. These concerns were reinforced when the carcass of a deer harvested nearby was found to have grossly elevated levels of PCBs in tissue. The animal came to be known as the “anomalous ungulate”.

An elevated level of total PCBs was found in muscle (13 ppb) and fat (calculated to be 382 ppb on the basis of 3.4% fat in muscle) of one deer taken near the fence of the facility several weeks after the release. Further analysis showed total equivalency (a summary measure) for dioxin and furan content of 120 pg/g in fat and 3.9 pg/g in muscle. By comparison, no detectable PCBs were found in a deer obtained as road kill away from the Swan Hills area and toxic equivalencies (TEQs) for that deer was only approximately 3.6 pg/g in fat and 0.69 in muscle. The calculated concentration in fat exceeded the guideline of 200 ppb for commercial beef fat promulgated by Health Canada in the Canadian Guidelines for Chemical Contaminants in Food. Venison is much leaner than beef, however, making it unlikely that a person would consume enough fat to acquire a significant body burden from that source alone.

The surprising observation of a level that was high in a local deer was unexpected and has never been satisfactorily explained. At the time, it raised the possibility of much more severe and widespread contamination than had been expected but had not yet been found. More comprehensive investigations were then undertaken.

The so-called “anomalous ungulate” prompted an urgent, initial food consumption advisory issued by the health officials (see Table 5 for current guidelines). No example of comparable elevation in contamination was found in deer after this incident. However, the anomalous elevation found in this one animal raised concerns that drove the stringent health advisory for several years.

An initial food consumption advisory was issued in the winter of 1996 based on a preliminary assessment. This notice advised against eating wild game taken from the Swan Hills area, defined as game taken within 30 km of the plant. Meat taken from the area that had already been frozen for future consumption was to be stored pending further notices. Notwithstanding these precautionary measures, the public was advised that no health effects were likely to have occurred. Diet and activity survey and estimating exposure ratio were initiated in early 1997 to obtain accurate information on the dietary habits and activities of individuals living in the study area and to assess exposure levels resulting from consuming contaminated wild game and fish. The resulting estimates were compared to existing guidelines.

Food consumption advisories were then revised based on the new information. Alberta Health announced public health advisories that recommended that children, pregnant women and nursing mothers refrain from eating meat from deer or moose harvested within a 30-kilometer radius of the SHSWTC and advised all other residents to limit their intake to 13 ounces per month and to avoid organ meats. Women who were pregnant or breastfeeding and young children were advised to avoid eating meat and fish from that area.

The advisory essentially shut down the hunting and fishing guide activities, disproportionately affecting First Nations people for several years thereafter.

## CONTAMINANT TRACING THROUGH ECOLOGICAL PATHWAYS (THE SWAN HILLS STUDY)

The Swan Hills Study was an investigation of the exposure potential for human populations in the months and few years following the event. It was sponsored by the Government of Alberta, led jointly by Alberta Health and Alberta Environment.

Ecologically based investigations were conducted to determine the contaminant levels in the vicinity of the facility and the spatial and temporal distribution in air, vegetation, snow, water and sediment [3, 4]. Several options were considered for design of a study to assess POPs residues as indicators along pathways that might reach human populations or parallel patterns of human exposure. These proposed designs included modeling airborne movement, sampling from specified distances from the plant, and sampling on top carnivores as indicator species. Each of these strategies was dismissed for one or more of the following reasons. The airborne distribution of POPs may not predict human exposure mediated through migratory species (especially moose) or in fish collected downstream. Geographic displacement has little bearing on the movement of POPs through ecological pathways. The movement of POPs in the months following the incident was slowed by winter conditions, and contamination levels in indicator species would not have reached steady state by spring. It was decided instead to evaluate movement and bioaccumulation of POPs through ecological pathways and trophic levels.

One body of water was of particular concern, Chrystina Lake (sometimes called “Windy Lake”). This is a small lake in the boreal forest 1.5 km east by northeast of SHSWTC. It is the center of Chrystina Lake Provincial Recreational Area, which features a campground and has been popular for fishing, hiking, canoeing and kayaking and snowmobiling. Like many rivers and lakes in northern Alberta, Chrystina Lake is known to have trace water levels of mercury from natural sources.

The assessment included wild game and fish monitoring programs and human blood sampling [6, 7]. The primary goals of the assessment were 1) to evaluate the potential for human exposure to PCBs and PCDDs/PCDFs through consumption of wild game and fish, 2) to estimate the existing levels of these contaminants in the human population, and 3) to assess probable movement and mobilization that might suggest a future risk. This assessment was also designed to support public health intervention in the short term. The company operating the special waste treatment plant continued to carry out its own annual monitoring program on air, water, soil, vegetation, fish, voles and wild game [5, 9].

### Exposure Assessment: Wild Game

Fresh deer samples were taken directly from the Swan Hills area, within the 30-km radius designated for this study. Three whitetail deer were collected at distances of 10 km, 20 km, 30 km to the east of the facility. Eleven road-kill adult deer carcasses were collected from other locations of Alberta as a control group. Frozen deer and moose samples were also taken from animals preserved in home freezers and donated by local licensed hunters and First Nations people. Approximately 40 people donated 60 frozen deer and moose meat samples collected between October 1996 and February 1997 from within a 30-km radius of the facility. Samples tested consisted of muscle, liver and kidney. All specimens were kept frozen prior to laboratory analysis.

### Exposure Assessment: Fish

Fish samples were taken from Chrystina Lake, which was immediately downwind of the plant at the time of the incident, and two reference lakes, Roche Lake, about 33 km east of the facility and downwind, and Chip Lake, a large lake and bird sanctuary 179 km south of Swan Hills and not directly downwind. The lakes therefore provided a gradient of probability of deposition from high (Chrystina), medium (Roche), to low (Chip), which served as the reference.

Field collection was carried out by Alberta Environmental Protection during June and July, 1997. A total of 16 brook trout were collected from Chrystina Lake, about 1.5 km east by northeast of the facility and downwind, with an average age of 2.0 years (a range of 1 to 3) and average weight of 112 g (a range of 60 to 229). Seventeen northern pike were collected from Roche Lake with an average age of 4.4 years (a range of 3 to 6) and average weight of 1.2 kg (a range of 0.8 to 2.3). A total of 32 northern pike were collected from Chip Lake (a reference lake) with an average age of 5.4 years (a range of 4 to 7) and average weight of 1.1 kg (a range of 0.6 to 1.5). Both fillet and liver were analyzed. For Chrystina and Roche Lake samples, each composite sample was formed from 4 (or 5) fish from a single species from the same lake with approximately the same length and weight. For Chip Lake (reference) samples, each composite sample was formed from 6 or 7 fish. A total of 26 composite samples were formed. All specimens were kept frozen prior to laboratory analysis.

### Exposure Assessment: Human Populations

The target population for human exposure assessment consisted of individuals older than 18 years who resided in communities of the Swan Hills area. Two sub-populations were chosen for study: residents of the Town of Swan Hills and persons living in all communities within a 100 km radius of the facility. During March and April 1997, a telephone interview was conducted to identify potential participants and their consumption patterns. A total of 65 participants donated blood. About 50 ml of venous blood was collected from each non-fasting participant. Six composite samples were collected in Edmonton, Alberta, for comparison. These samples were formed from six age- and gender-specific groups: males aged 17–35, males aged 35–55, males older than 55, females aged 17–35, females aged 35–55, and females older than 55 years. All specimens were kept frozen prior to shipping to the laboratory. Chemical analysis was as described with the addition that total lipid concentration was calculated by summation of the amount of triglycerides and total cholesterol [8].

### Health Hazard Assessment

Contaminant concentrations in environmental media and potentially exposed food sources were assessed to track the distribution of contaminants through its ecological pathways to the human receptor. (This approach is various known as “fate and disposition studies” or “ecotoxicology”.)

Survey measurements were used to assess risks associated with the demographic characteristics, dietary habits and activity patterns of potentially exposed populations. The results were compared to the literature on human exposure in other settings emphasizing Canadian populations [1013], recognizing that there is no exactly comparable situation on record of such exposure on the edge of wilderness.

POPs body burden studies were also carried out using blood (serum) determinations on a sample of cooperating residents of Swan Hills. Sixty-five residents of Swan Hills participated, divided into a group that consumed some country foods (43, 22 men and 21 women) and those who did not (22, 6 men and 16 women), with a mean age of 40 years. Only six residents self-identified as Aboriginal, however, so this part of the study should be considered to be primarily a study of townspeople. For reference, 12 pooled samples were taken from 150 subjects who lived out of the area, elsewhere in the province. (The use of pooled samples means that blood taken from many individuals were combined, mixed and determined in one analysis reported as an average, rather than individually determined. This was done to reduce costs.)

The information from each study component was integrated into a comprehensive risk assessment in order to inform the public health interpretation and revise the initial food consumption advisory.

Concentrations of PCBs, PCDDs, and PCDFs were converted to the sum of toxicologically comparable activities calculated as equivalencies to 2,3,7,8- tetrachlorodibenzodioxin (TCDD), a quantity called “total dioxin-like toxic equivalencies” (Σdioxin-like TEQs). Equivalency factors followed recommendations of the World Health Organization European Center for Environment and Health and the International Programme for Chemical Safety [11, 13]. The coefficients used to weigh the mass measurements for each dioxin and furan (the “toxic equivalence factor”) were those current as of 1996 [12, 13].

### Diet and Activity Survey

Advisories for the consumption of foods requires knowledge of contaminant levels, food preparation (when the contaminant is modified by food processing), serving portions, frequency of meals, and the relationship to human body burden [1417]. Ideally, they also take into consideration health benefits of consumption (a particular consideration with respect to fish).

A diet and activity survey was conducted through telephone interviews during March and April 1997. The survey was divided into two phases. Of 370 subjects contacted in the town of Swan Hills, 327 (88%) including 12 Aboriginal people participated in the first telephone interview. A second, follow-up telephone interview was then conducted with 100 participants randomly selected from among those who had participated in the first telephone interview.

Participants were asked to recall their consumption of wild game and fish and their outdoor recreational activities within 100 km radius of the facility for the previous 12 months. Specifically, the initial survey was used to determine types of outdoor activities within the study area; frequency, duration and amount of wild game and fish consumption; and the respondents’ awareness of an adherence to the existing food consumption advisory. The second survey requested demographic characteristics (age, gender, ethnic group, weight and height, occupation, duration of residency and number of persons in the household); proportion of activity time spent indoors and outdoors and detailed information about outdoor activities; detailed information about daily food consumption including consumption of wild game, fish and wild fruit, vegetables and herbs, and cooking preparation techniques for wild game and fish; and more detailed information about lifestyle (use of alcohol and cigarettes, health conditions and perception of the current health advisory).

Estimated daily intake (EDI) was calculated as follows:

$EDI=C×IR×BF/BW$

C is measured concentrations of contaminants, IR is food consumption rate, BF is bioavailability factor (assuming 100%), and BW is average body weight (73 kg for Albertans). Exposure ratios (ER) were calculated by using the following equation:

$ER=EDI/TDI$

Tolerable daily intakes (TDIs) are 1 µg/kg/d for PCBs and 10 pg/kg/d for TCDD as determined by Health Canada. Note that this is not a toxicity threshold. It is a level of consumption that is presumed to be on a par with current habits and diet that are not associated with known harm. In theory, a TDI set at this level would not increase any risk to health, even for an outcome that may exist but not be detected or recognized, and so this TDI is intended to be precautionary.

### Management and Communication

A process was established to manage the investigations. The work group undertook data collection, laboratory analysis, data analysis, data interpretation and data presentation. A specially constituted Science Advisory Committee, consisting of local, national and international experts, the author included, provided advice on direction of the study and assistance in interpreting results. A Public Health Advisory Committee advised on issues of the human health risk and management options. The risk communication team at Alberta Health delivered the information to targeted audience and stakeholders through various channels including public media, local community meetings, direct mail and telephone.

## RESULTS

The results of the study were shared in detail with community residents, including First Nation peoples, by Alberta Health and then were presented in detail at an international conference on dioxin in 1998 and published in the conference proceedings [38].

### Ecotoxicology

Table 1 presents a summary of data on residues in wild game, simplified from the original report [6]. Deer (whitetail or mule) were either harvested by local hunters or found as recent road kill.

TABLE 1.

Summary of PCB, PCDD and PCDF Levels in Fresh Deer Samples

ContaminantSwan HillsAlberta Reference
LiverMuscleLiverMuscle
Sample size 11 11
Detectable PCBs found
Detectable PCDD+PCDFs found 10
Lipid content (%) 3.20 1.87 3.42 1.64
Mean of ∑PCBcongener(µg/kg) (range) 1178
(103–2799)
509
(ND–1527)
194
(ND–1177)
158
(ND–821)
Mean of ∑PCDD/Fcongener(ng/kg) 2349 153 17
∑ PCDD/F 4698 32 100
∑ Dioxin-like compounds (range) 5989
(74–9198)
1031
(14–3038)
136
(15–819)
13
(0.98–92)
% of ∑PCB/∑Dioxin-like compounds 22 97 26 35
% of ∑PCDD+PCDF/∑Dioxin-like compounds 78 74 65
% of ∑non-ortho-PCB/∑Dioxin-like compounds 21 96 0.77 21
ContaminantSwan HillsAlberta Reference
LiverMuscleLiverMuscle
Sample size 11 11
Detectable PCBs found
Detectable PCDD+PCDFs found 10
Lipid content (%) 3.20 1.87 3.42 1.64
Mean of ∑PCBcongener(µg/kg) (range) 1178
(103–2799)
509
(ND–1527)
194
(ND–1177)
158
(ND–821)
Mean of ∑PCDD/Fcongener(ng/kg) 2349 153 17
∑ PCDD/F 4698 32 100
∑ Dioxin-like compounds (range) 5989
(74–9198)
1031
(14–3038)
136
(15–819)
13
(0.98–92)
% of ∑PCB/∑Dioxin-like compounds 22 97 26 35
% of ∑PCDD+PCDF/∑Dioxin-like compounds 78 74 65
% of ∑non-ortho-PCB/∑Dioxin-like compounds 21 96 0.77 21

Note: ∑ = total. Mass is wet weight. 1 µg/kg = 1 part per billion (ppb). 1 ng/kg = 1 part per trillion (ppt).

The levels of PCBs, dioxins and furans were significantly elevated in both liver and muscle samples from the study area as compared to the reference areas. The patterns of PCB congener distribution were different for deer samples from study and control areas. The majority of dioxin-like TEQ value was due to dioxin and furan concentrations in liver. One PCB congener, 126, accounted for 97% of total dioxin-like TEQ in muscle from deer in Swan Hills.

Table 2 presents a summary of data on residues in fish, simplified from the original report [7]. The levels of PCBs, dioxins and furans were significantly elevated in the liver and muscle samples of fish taken from the lake nearest and downwind of the SHSWTC, specifically Chrystina Lake (immediately downwind). Under normal circumstances, northern pike, a predator, would be expected to have higher contaminant concentrations than brook trout which feed on planktonic invertebrates. The lower contaminant values in pike from Roche Lake (933 km east of Swan Hills) and Chip (reference) Lake indicate very low contaminant background.

TABLE 2.

Summary of PCB, and PCDD/F Levels in Fish Samples

ParameterMuscle
Liver
ChrystinaRocheChipChrystinaRocheChip
Sample size (composite)
Detects of PCBs
Detects of PCDD/Fs
Lipid content (%) 0.4 0.6 0.3 N/A 4.3 3.6
Mean of ∑PCBcongener(µg/kg) (range) 18
(9.7–27)
1.0
(0.3–2.8)
0.25
(0.04–0.7)
70
(41–117)
7.8
(1.2–14)
6.4
(3–18)
Mean of ∑PCDD/Fcongener(ng/kg) (range) 22
(12–30)
0.93
(0.7–1.1)
0.68
(ND–1.2)
227
(55–351)
1.2
(ND–2)
7.5
(ND–19)
∑ PCDD/F 2.7 0.003 0.002 16 0.2 0.1
∑ Dioxin-like compounds (range) 12.4
(6–19)
0.24
(0.01–1)
0.004
(ND–0.007)
61
(24–107)
3.2
(0.7–5.5)
2.4
(1–7)
% of ∑PCB/∑Dioxin-like 78 98 56 73 93 95
% of ∑PCDD-F/∑Dioxin-like 22 44 27
ParameterMuscle
Liver
ChrystinaRocheChipChrystinaRocheChip
Sample size (composite)
Detects of PCBs
Detects of PCDD/Fs
Lipid content (%) 0.4 0.6 0.3 N/A 4.3 3.6
Mean of ∑PCBcongener(µg/kg) (range) 18
(9.7–27)
1.0
(0.3–2.8)
0.25
(0.04–0.7)
70
(41–117)
7.8
(1.2–14)
6.4
(3–18)
Mean of ∑PCDD/Fcongener(ng/kg) (range) 22
(12–30)
0.93
(0.7–1.1)
0.68
(ND–1.2)
227
(55–351)
1.2
(ND–2)
7.5
(ND–19)
∑ PCDD/F 2.7 0.003 0.002 16 0.2 0.1
∑ Dioxin-like compounds (range) 12.4
(6–19)
0.24
(0.01–1)
0.004
(ND–0.007)
61
(24–107)
3.2
(0.7–5.5)
2.4
(1–7)
% of ∑PCB/∑Dioxin-like 78 98 56 73 93 95
% of ∑PCDD-F/∑Dioxin-like 22 44 27

### Human Populations

No statistically significant difference in human serum levels of PCBs, PCDDs, and PCDFs based on current detection limits was observed for the 65 residents living within a 100 km radius of the SHSWTC compared to subjects living in an urban reference area. There was no statistically significant difference between local wild game and fish consumers and non consumers. Different PCB congener patterns were observed in a small number of the workers at the SHSWTC.

The PCB, PCDD and PCDF congener patterns were also different in the two human sample groups. More PCB congeners were detected in the community samples than in the pooled urban reference samples. The majority of ∑PCDD and PCDF contributions to TEQ was due to PCDD, which contributed 83% to ∑PCDD/F TEQ in the community samples and 94% in the pooled samples. Thus, dioxins rather than furans dominated among the contaminants.

The POPs data for Swan Hills residents will not be presented here, for three reasons. First, they are low. Most congener levels were at or near the level of detection, both for residents and, especially, reference. Related to that, the results are known to be imprecise and probably inaccurate but very low. The laboratory that did the initial work discovered on doing replicate samples that there was a non-systematic error in measurement, such that the discrepancy could not be corrected by simple adjustment. Error, of course, increases when levels are close to the limit of detection and that appears to be part of the reason for this problem. Thus, they will not be reproduced here although values were in fact published before the error was discovered [8]. Notwithstanding the problem, generalizations can be made.

Preliminary analysis indicated that there was no observable difference in levels of PCBs, PCDD and PCDFs based on current detection limits was observed between individuals who consumed wild game and fish taken from within 30 km radius of the SHSWTC and those who did not.

### Daily Intake and Exposure Ratio

Estimated daily intake and exposure ratios (see formula above) are presented in Table 3. The daily intake of PCBs, PCDD and PCDFs for the average Canadian comes mainly from the diet through commercial food sources, and to a lesser extent, from breathing contaminants in air and drinking water. The daily intake from background exposure for adult Canadians is estimated to be 2–4 pg/kg/d of PCDD+PCDF [10]. Health Canada Guidelines set a limit of 10 pg/kg (body weight)/day. For a 73 kg (160 pound) man, that equated to 3 oz/week or 13 oz/mo. of deer and moose meat from within 30 km around SHSWTC, or 6 oz/week or 26 oz/mo. of brook trout from Chrystina Lake. (Amounts are expressed in ounces because that was the measure most often used in food preparation in the community: 1 oz = 0.028 kg.)

TABLE 3.

Estimated Daily Intake (EDI) and Exposure Ratio (ER)

Consumption GroupHigh IntakeMedium IntakeLow IntakeVery Low Intake
Percentile Concentration*50th90th50th90th50th90th50th90th
Wild game EDI ∑PCB(µg/kg/d) 0.02 0.2 0.005 0.06 0.001 0.01 0.0002 0.002
ER ∑PCB 0.02 0.2 0.005 0.06 0.001 0.01 0.0002 0.002
EDI ∑TEQ (pg/kg/d) 6.2 175 2.0 53 0.4 12 0.07 1.8
ER ∑TEQ** 0.6 17.5 0.2 5.3 0.04 1.2 0.007 0.18

Brook trout EDI ∑PCB(µg/kg/d) 0.04 0.08 0.01 0.02 0.003 0.007 0.0005 0.001
ER ∑PCB 0.04 0.08 0.01 0.02 0.003 0.007 0.0005 0.001
EDI ∑TEQ (pg/kg/d) 28 66 18 2.2 5.1 0.3 0.8
ER ∑TEQ 2.8 6.6 0.8 1.8 0.2 0.5 0.03 0.08
Consumption GroupHigh IntakeMedium IntakeLow IntakeVery Low Intake
Percentile Concentration*50th90th50th90th50th90th50th90th
Wild game EDI ∑PCB(µg/kg/d) 0.02 0.2 0.005 0.06 0.001 0.01 0.0002 0.002
ER ∑PCB 0.02 0.2 0.005 0.06 0.001 0.01 0.0002 0.002
EDI ∑TEQ (pg/kg/d) 6.2 175 2.0 53 0.4 12 0.07 1.8
ER ∑TEQ** 0.6 17.5 0.2 5.3 0.04 1.2 0.007 0.18

Brook trout EDI ∑PCB(µg/kg/d) 0.04 0.08 0.01 0.02 0.003 0.007 0.0005 0.001
ER ∑PCB 0.04 0.08 0.01 0.02 0.003 0.007 0.0005 0.001
EDI ∑TEQ (pg/kg/d) 28 66 18 2.2 5.1 0.3 0.8
ER ∑TEQ 2.8 6.6 0.8 1.8 0.2 0.5 0.03 0.08

*Concentrations at 50th percentile were 6.5 µg/kg, wet weight, for ∑PCBs and 2.4 ng/kg for ∑TEQ in deer muscle, and 18 µg/kg for PCBs and 12 ng/kg for ∑TEQ in brook trout muscle; concentrations at 90th percentile were 73 µg/kg for PCBs and 67 ng/kg for ∑TEQ in deer muscle, and 36 µg/kg for PCBs and 29 ng/kg for ∑TEQ in brook trout muscle.

**∑TEQ = ∑dioxin-like TEQ.

The exposure ratio (ER) reflects the ratio between the actual level of exposure (external dose) in a particular circumstance and a reference standard associated with observed toxicity in humans or animals. In the current assessment, the estimated daily intake and exposure ratios provide insight into additional exposure that might be expected from consuming local wild game and fish. Not unexpectedly, this analysis predicted that body burden elevations due to higher-than average intake were most likely to occur in the highest consumers in the high and medium consumption groups. In the event, elevated serum POPs were not seen to be elevated in these groups. However, the analysis informed and gave guidance for reformulating the public health advisory.

## DISCUSSION

The results indicate that residents in these communities of northern Alberta have lower PCB and dioxin-like TEQ serum levels than do residents of other jurisdictions. The comparison of our results with other jurisdictions is difficult because of the different background exposures, timing of exposure, laboratory analytical methods and quantitative methods. Trends revealed that the levels of PCBs, dioxins and furans declined in human blood over past 25 years. As reported by Health Canada, the mean serum levels of ∑PCBcongener in Canadians ranged from 1.48 to 1.84 μg/L, whole weight [10, 11]. In Germany, the mean TEQ values in human blood decreased to 16.5 pg/g blood lipid in 1996 from 43 pg/g reported for 1989 [12]. The downward trends may reflect the decreased exposure to chlorinated organic compounds for the general population since 1980s and the gradual elimination of these contaminants from human body due to half-time elimination rate of 5 to 10 years for dioxins, furans and PCBs.

Around the world, various regulatory guidelines have been developed for the most potent dioxin, TCDD, the most toxic dioxin, based on cancer risk. The guidelines are expressed as a reference dose (RfD) or a TDI, that is, a lifetime daily dose for TCDD which is believed to be without potential health effects to humans. In the past, TCDD has been treated as a threshold carcinogen by some regulatory agencies. Based on this assumption, a value of 10 pg/kg body weight/day has been adopted by Health Canada [11]. Some PCB and PCDD/F congeners produce similar toxic effects to humans and animals as TCDD. The similarity of toxicity between these congeners and TCDD was assessed using toxic equivalency factors (TEFs). ∑dioxin-like TEQ combines ∑PCB TEQ and ∑PCDD and ∑PCDF TEQ.

Based on concentrations at the 90th percentile of ∑dioxin-like TEQ equivalency to TCDD, a set of consumption limits was recommended (Table 4). These consumption limits provided guidance on the evaluation of the potential risk associated with exposure to PCBs and PCDD/Fs for individuals who consume deer or moose meat taken within a 30-km radius of the facility and/or brook trout from Chrystina Lake, near the facility. The estimated values represent the amount of meat from deer or moose and from edible portions of brook trout expected to generate a risk no greater than the TDI proposed by Health Canada, based on a lifetime of daily consumption (see Table 5 for current guidelines). The toxicity of TCDD and related congeners also includes reproductive, developmental and immunotoxic effects [12]. Children and pregnant women or women who are breastfeeding are susceptible groups and as a precautionary measure were advised to avoid consuming wild game and fish from the area. Because contaminants tend to accumulate in the internal organs in various animals and the measured levels of the contaminants were very high in the liver samples, it was recommended that people avoid consumption of viscera from wild game and fish.

TABLE 4.

Estimated Daily Intake* (EDI) and Exposure Ratio** (ER)

Consumption group →High ConsumptionMedium ConsumptionLow ConsumptionVery Low Consumption
Consumption
(muscle only) →
>100 g/d30–99 g/d5–29 g/d<4 g/d
Source Parameter Percentile → 50 90 50 90 50 90 50 90
Wild game*** EDI ∑PCB µg 0.02 0.2 0.005 0.06 0.001 0.01 0.0002 0.002
ER ∑PCB 0.02 0.2 0.005 0.06 0.001 0.01 0.0002 0.002
EDI ∑TEQ pg 6.2 175 2.0 53 0.4 12 0.07 1.8
ER ∑TEQ 0.6 17.5 0.2 5.3 0.04 1.2 0.007 0.18
Brook trout EDI ∑PCB µg 0.04 0.08 0.01 0.02 0.003 0.007 0.0005 0.001
ER ∑PCB 0.04 0.08 0.01 0.02 0.003 0.007 0.0005 0.001
EDI ∑TEQ pg 28 66 18 2.2 5.1 0.3 0.8
ER ∑TEQ 2.8 6.6 0.8 1.8 0.2 0.5 0.03 0.08
Consumption group →High ConsumptionMedium ConsumptionLow ConsumptionVery Low Consumption
Consumption
(muscle only) →
>100 g/d30–99 g/d5–29 g/d<4 g/d
Source Parameter Percentile → 50 90 50 90 50 90 50 90
Wild game*** EDI ∑PCB µg 0.02 0.2 0.005 0.06 0.001 0.01 0.0002 0.002
ER ∑PCB 0.02 0.2 0.005 0.06 0.001 0.01 0.0002 0.002
EDI ∑TEQ pg 6.2 175 2.0 53 0.4 12 0.07 1.8
ER ∑TEQ 0.6 17.5 0.2 5.3 0.04 1.2 0.007 0.18
Brook trout EDI ∑PCB µg 0.04 0.08 0.01 0.02 0.003 0.007 0.0005 0.001
ER ∑PCB 0.04 0.08 0.01 0.02 0.003 0.007 0.0005 0.001
EDI ∑TEQ pg 28 66 18 2.2 5.1 0.3 0.8
ER ∑TEQ 2.8 6.6 0.8 1.8 0.2 0.5 0.03 0.08

*EDI = EEstimated daily intake. All values are per kg per day (kg/d)-1.

**ER = Exposure ratio, with EDI in numerator and Health Canada guidelines in denominator (dimensionless). Ratios >1.0 indicate consumption higher than Health Canada guidelines and are highlighted in bold.

***Deer or moose.

TABLE 5.

Risk-Based Specific Consumption Limits

FoodMeat: deer and mooseFish: brook trout
Location ≤30 km radius of SHSWTC Chrystina lake
Tissue Muscle Muscle
Health Canada TDI for TCDD(pg/kg bw/day) 10 10
Body weight (kg) based on Alberta average (73 kg man) 73 73
Consumption limit (oz/week)
Consumption limit (oz/month) 13 26
FoodMeat: deer and mooseFish: brook trout
Location ≤30 km radius of SHSWTC Chrystina lake
Tissue Muscle Muscle
Health Canada TDI for TCDD(pg/kg bw/day) 10 10
Body weight (kg) based on Alberta average (73 kg man) 73 73
Consumption limit (oz/week)
Consumption limit (oz/month) 13 26

## FAST-FORWARD TO 2010

Because of on-going concern and a desire to reassure residents that contaminant levels had remained low, the Government of Alberta initiated a long-term follow-up program following the Swan Hills study. It measured POPs levels in fish to 2012 and in wild game to 2010 [18].

### New Information on Contaminant Levels

Information available since 2000 [18] is not as detailed as in the original Swan Hills study but documents the following trends for POPs in deer:

• Over time, POP levels (measured as total TEQs) declined in muscle and liver of deer over the 13 years. This was true for PCDDs, PCDFs and PCBs.

• POPs levels (all three types) decreased, as a trend, with increasing distance from the plant in deer.

• PCB levels in both muscle and liver, and PCDDs and PCDFs in liver only, remained higher close to the plant (within 10 km) than in a reference area much further away.

• PCDD levels in muscle were the same near the plant and in a reference area but in liver the levels were higher close to the plant.

• The three types of POPs maintained the same congener pattern (throughout the region, which is an indication that they came from the same source: SHSWTC.

Similarly, for brook trout in Chrystina Lake, the levels of PCDDs and PCDFs declined over 15 years [19, 20]. However, PCBs remained about the same.

The health risks, relative to a Health Canada guideline, were calculated for chronic disease based on organochlorine POPs toxicity. The advisories are based on the assumption that the benefits of health protection outweigh the risks associated with avoiding country foods. For the majority population, this is not in serious doubt. However, Aboriginal communities may have other needs and concerns.

Wild game and fish supplement the diet of a number of people living in the area surrounding the facility. Concern has been raised by both recreational users and traditional Aboriginal users because these groups consume more wild game and fish than the general population. The balance between nutritional benefits and health risks arising from the consumption of contaminated food is an important consideration in issuing public health advisories [14, 19]. When country foods are not available, residents must purchase foods at a store. Store-bought foods, as a generalization, are highly processed in northern latitudes (therefore high in sodium, saturated fats and sugar), less dense in caloric value, and are associated with increased risk for diabetes, which is prevalent in northern communities. Store-bought foods also do not reflect or reinforce cultural values or traditions. Thus, foregoing country foods is apt to increase costs, reduce nutrition, increase risk of disease, and lead to cultural disruption, especially in raising children. Likewise, some non-nutritional benefits are often considered in developing a risk management strategy. These non-nutritional benefits include immunological defenses provided to nursing infants from mother’s milk, in which POPs tend to be concentrated.

Cultural issues, however, are usually not usually taken into account in food advisories. Such issues include maintaining traditional lifestyle elements that are important for cultural reasons (hunting as initiation into maturity for young men, food gathering as contributing to family for girls), food preferences (such as liver), and community cohesion (hunting, dressing game and distributing meat).

Under other circumstances, an advisory that interfered with traditional activities might have compromised cultural values that were important to the sustainability of the community. However, data from a survey conducted by Health Canada in 1997 and 1998 (unpublished) suggested that this was not a critical factor in this case. The particular area of greatest concern, just east of SHSWTC and Chrystina Lake but west of Roche Lake, was not identified as a critical area of traditional use by indigenous peoples in harvesting wildlife and plants, which limited the potential for harm.

The dietary survey determined the extent of wild game and fish consumption by residents of the study area. Survey results indicate that only a very small number of people consumed large quantities of wild game and fish caught near the facility; this limited the number of individuals to be at risk. Second, the advisories substituted consumption limits developed from risk estimates for an outright ban on consumption. These limits do indicate that wild game and fish may still be safely consumed in moderation. Third, the current advisories are restricted to a 30-km radius of the facility, limiting interference with harvesting to the smallest area possible consistent with public protection. Traditional and recreational users can still safely consume wild game and fish obtained from outside the affected area; traditional hunting and fishing areas previously covered by the ban have now been opened. Finally, the consumption limits provided in the advisories were calculated in reference to uncooked food.

Food preparation methods that reduce POPs intake further were encouraged in response to inquiries. Many studies have shown that appropriate food preparation and cooking techniques can reduce the concentrations of PCBs, PCDDs and PCDFs in fish and meat [16, 21, 22]. These techniques such as removing the skin prior to cooking, broiling and baking are recommended for individuals who continue to consume wild game and fish from the areas immediately surrounding the facility.

However, there may be benefits to continued use of country foods by populations for whom it is culturally important, specifically, the Aboriginal population. This was not reflected in the initial assessment but its importance came to be recognized in the revised advisory. This case study therefore illustrates both an ecological approach to risk assessment and a culturally sensitive approach to risk management.

The original public health advisory was released both to protect against contamination that may have already been reaching country food consumers and in order to “get ahead” of the problem should contaminant levels rise significantly. In the event, they did not and the on-going monitoring program gave reassurance that POPs levels were declining and had not peaked at levels presenting a risk to human health.

The current advisory dates from 2012 and reads in its entirety as follows [20]:

Alberta’s Chief Medical Officer of Health issued a public health notice in 2012 to revise the existing food consumption advisory in the Swan Hills area.

• Limit eating wild game taken from within a 15-km radius of Swan Hills Treatment Centre to 150 g (two servings) per day

• Avoid eating organ meat (liver, kidney) or using fat from wild game harvested within a 15-km radius of the treatment centre

• Pregnant or breastfeeding women should avoid eating wild game taken from within a 15-km radius of the treatment centre

• Young children should avoid eating wild game taken from within a 15-km radius of the treatment centre.

Restrictions on fish consumption in the area are not unlike advisories for other bodies of water in northern Alberta. For northern pike and walleye over eight pounds (3.6 kg) and four pounds (1.8 kg), respectively, there is no limit on consumption by male adolescents and adults (12 years and older), a limit of 2.0 servings per week for women of childbearing age (15–49 years) or who are pregnant, 1.0 serving for older children (5–11 years), and 0.5 serving for younger children (1–5 years). A serving for this purpose is defined as a piece of fish as large as the palm of one’s hand or 75 g. However, these restrictions are based on mercury contamination, not levels of POPs. Mercury contamination is present naturally in many northern waterways; there is no industrial source [18]. Thus, although an advisory continues, it is no longer driven by POPs for fish consumption. Public health advisories for fish consumption are often complicated because they take into account population risk for effects on brain development at various ages and in pregnancy.

## CONCLUSION

The rationale of both the Swan Hills and Lesser Slave Lake studies was that POPs had entered the ecosystem and pathways that would eventually result in human exposure. The magnitude of exposure did not appear to constitute a major risk to health, however. This was gratifying but unexpected, as the public health authorities of the province had geared up for a massive response to an imminent public health emergency that did not materialize.

Food advisories were problematic; they were still in place almost 20 years later, but driven more by the hazard of mercury than by POPs. This placed a constraint mainly on the Aboriginal population, which was disproportionately affected. Because of ecological pathways and routes of exposure in the North, persons living traditional lifestyles, such as Aboriginal communities in northern Canada, might be at higher risk from certain types of pollution than urban residents; the risk is specific to the situation. This was anticipated by the Swan Hills study but not demonstrated in this particular case.

Ecosystem fate and mobility studies (ecotoxicology) can be used to anticipate future human health risk. The Swan Hills Study was elegant and methodologically sophisticated for its time but, as always, not ideal. The determination of organochlorine concentrations in blood was flawed by an intractable methodological problem, and so should not be cited as accurate values. However, in the event, the margin of error was more than sufficient to describe a low public health risk for the population and strict accuracy was not needed. This is typically the case for problems of public health significance: small inaccuracies, however regrettable, are often unimportant in determining the public health risk.

Part 3 continues the case study with a targeted, although limited, study of the population at greatest risk, the Aboriginal community.

Many scientists and public health professionals took part in the several studies that together constituted the evaluation of the Swan Hills incident. The following colleagues performed or substantially contributed to work that was incorporated into this case study: Stephan Gabos, Donald Schopflocher, John Waters, Karen Grimsrud, Derek G. Muir, David Schindler, Ken Froese, Jules M. Blais, Sub Ramamoorthy, Siu Chan, Chris Le, Arnold Schechter, Weiping Zhang (Chen), Pierre Band, Cornelia Kreplin, D. Onderka, Margo Pybus, Kevin McLeod, Erik Ellehoj, and by health professionals and managers of Health Canada. Data cited in this case study are publicly available with the exception of the Lesser Slave Lake study, which have been publicly presented but is published here in the journal literature for the first time. The case study was taught for several years as a case study in PubH 243 Public Health Practice at the George Washington University and Ms. Amy O’Connor developed the narrative further for her MPH research project in 2007.

## FUNDING

Funding for Dr. Guidotti’s participation in this case study in the role of consultant was initially provided by Alberta Environment as the lead agency, and subsequently by Health Canada.

## COMPETING INTERESTS

The author has no competing interest to declare.

## DISCLOSURE

During the events described in this case study, the author served as a principal consultant in environmental health for Alberta Environment, Alberta Health, and Alberta Justice, and, later, Health Canada. (Agency names are given as they were in 1997.)

## REFERENCES

REFERENCES
1.
Sherbaniuk R.
The price of protection
.
Alberta Views
.
1998
;
3
:
26
33
.
2.
Guidotti TL.
Evaluating risk after a hazardous waste treatment plant released persistent organic pollutants: Part 1, Background and policy issues
.
Case Stud Environ
.
2018
. DOI: https://doi.org/10.1525/cse.2017.000711.
3.
Froese KL, Blais JM, Muir DG.
Conifer Forest Vegetation as an Indicator of PCB Exposure in the Region of Swan Hills, Alberta, Canada
. In: Johansson N et al., editors.
Organohalogen Compounds, Ecotoxicology, Environmental Levels, Northern Environments. 18th Symposium on Halogenated Environmental Organic Pollutants
,
Stockholm, Sweden
, August 17–21;
Swedish Environmental Protection Agency
,
1998
. pp.
185
188
.
4.
Blais JM, Froese KL, Schindler DW, Muir DG.
Assessment of PCBs in Snow and Lakes Sediments Following A Major Release from the Alberta SHSWTC Near Swan Hills, Alberta, Canada
. In: Johansson N et al., editors.
Organohalogen Compounds, Ecotoxicology, Environmental Levels, Northern Environments. 18th Symposium on Halogenated Environmental Organic Pollutants, Stockholm, Sweden, August 17–21
;
Swedish Environmental Protection Agency, 1998
. pp.
189
192
.
5.
Chem Security.
Operator of SHSWTC, Waste Treatment Center Environmental Monitoring Results 1996, 95-IND-237, Alberta, Canada
;
1997
.
6.
Gabos S, Schopflocher D, Muir DG et al.
Levels of PCBs, PCDDs and PCDFs in Deer and Moose in Alberta, Canada following Accidental Release from a SHSWTC
. In: Johansson N et al., editors.
Organohalogen Compounds, Ecotoxicology, Environmental Levels, Northern Environments. 18th Symposium on Halogenated Environmental Organic Pollutants, Stockholm, Sweden, August 17–21
;
Swedish Environmental Protection Agency, 1998
. pp.
169
172
.
7.
Gabos S, Schopflocher D, Muir DG et al.
Levels of PCBs, PCDDs and PCDFs in Fish in Alberta, Canada Following Accidental Release from a SHSWTC
. In: Johansson N et al., editors.
Organohalogen Compounds, Ecotoxicology, Environmental Levels, Northern Environments. 18th Symposium on Halogenated Environmental Organic Pollutants, Stockholm, Sweden, August 17–21
;
Swedish Environmental Protection Agency
,
1998
. pp.
173
176
.
8.
Chen W, Schopflocher D, Muir DG, et al.
Preliminary Analysis of Levels of PCBs, PCDDs and PCDFs in Blood in Residents of Swan Hills and Surrounding Communities, Alberta, Canada. (1998). Levels of PCBs, PCDDs and PCDFs in Deer and Moose in Alberta, Canada following Accidental Release from a SHSWTC
. In: Johansson N et al., editors.
Organohalogen Compounds, Ecotoxicology, Environmental Levels, Northern Environments. 18
th
Symposium on Halogenated Environmental Organic Pollutants, Stockholm, Sweden, August 17–21
;
Swedish Environmental Protection Agency, 1998.
pp.
177
180
.
9.
Chem Security.
Operator of SHSWTC, Waste Treatment Center Environmental Monitoring Results 1997, 95-IND-237, Alberta, Canada
;
1998
.
10.
Gilman A, Newhook R, Birmingham B.
An update assessment of the exposure of Canadians to dioxins and furans
.
Chemosphere
.
1991
;
1991
(
23
):
1661
1667
.
11.
Priority Substances List Assessment Report No. 1: Polychlorinated Dibenzodioxins and Polychlorinated Dibenzofurans. KE3619.P74 1990
;
1990
.
12.
DeVito MJ, Birnbaum LS, Farland WH, Gasiewicz TA.
Comparisons of estimated human body burdens of dioxin-like chemicals and TCDD body burdens in experimentally exposed animals
.
Environ Health Perspect
.
1995
;
103
:
820
831
.
13.
Ahlborg UG, Becking GC, Birnbaum LS et al.
Toxic equivalency factors for dioxin-like PCBs
.
Chemosphere
.
1994
;
28
(
8
):
1049
1067
.
14.
Kimbrough RD.
Consumption of fish: benefits and perceived risk
.
J Toxicol Environ Health
.
1991
;
33
:
81
91
.
15.
Burger J, Gochfeld M.
Fish advisories: useful or difficult to interpret?
Risk Safety Environ
.
1996
;
23
(
7
):
23
33
.
16.
Schecter A, Päpke O, Dellarco M.
Dioxin, Dibenzofuran, and PCB Congeners in Cooked and Uncooked Foods
. In: Hites R, editor.
Organohalogen Compounds, Vol. 33. Toxaphene, Transport and Fate, Ecotoxicology and Human Exposure. 17th International Symposium on Chlorinated Dioxins and Related compounds, Indianapolis, Indiana, USA, August 25–29
;
American Society for Testing and Materials
,
1997
. pp.
462
466
.
17.
Swan Hills Waste Treatment Center: Long-Term Follow-Up Human Health Risk Assessment Program.
Wild Game and Fish Monitoring, 1997–2012
.
Government of Alberta;
2013
18.
Chen W, Gabos S.
Mercury in fish in Chrystina Lake, Alberta, Canada. [Abstract, International Society for Environmental Epidemiology.]
.
Epidemiology
.
2003
;
14
(
5
):
S32
.
19.
Guidotti TL, Naidoo K.
Hunting, Trapping, and Wilderness-Related Work
. In: Weissman DN, Huy JM, editors.
Occupational Infectious Diseases. Clinics in Occupational and Environmental Medicine
. Vol.
2
.
Orlando FL
:
Elsevier Science
;
2002
. pp.
651
661
.
20.
Recommended fish consumption limits in Alberta.
Government of Alberta
;
2016
. Available: https://mywildalberta.ca/fishing/advisories-closures/documents/FishConsumptionLimitsAlberta-Feb24-2016.pdf. Accessed 16 August 2017.
21.
Petroske EP, Zaylskie RG, Feil VJ.
The Effect of Cooking on Dioxin and Furan Concentrations in Beef
. In: Hites R, editor.
Organohalogen Compounds, Vol. 33. Toxaphene, Transport and Fate, Ecotoxicology and Human Exposure. 17th International Symposium on Chlorinated Dioxins and Related compounds, Indianapolis, Indiana, USA, August 25-29
;
American Society for Testing and Materials, 1997
. pp.
436
439
.
22.
Zabik ME, Zabik MJ.
Tetrachlorodibenzo-p-dioxin residue reduction by cooking/processing of fish fillets harvested from the Great Lakes
.
Bull Environ Contam Toxicol
.
1995
;
55
:
264
269
.