Human Health Risk Assessment of Mercury in Fish and Health Benefits of Fish Consumption

Authored by scientists of the Bureau of Chemical Safety, Food Directorate, Health Products and Food Branch, Health Canada, with input from:

Bureau of Nutritional Sciences, Food Directorate, Health Products and Food Branch, Health Canada
Office of Nutrition Policy and Promotion, Health Products and Food Branch, Health Canada
Healthy Environments and Consumer Safety Branch, Health Canada
Fish, Seafood and Production Division, Canadian Food Inspection Agency
Environment Canada
Fisheries and Oceans Canada

Table of Contents

• 1.0 Background
  • 1.1 Purpose of this document
  • 1.2 Sources of Human Exposure to Mercury
  • 1.3 The Chemical Forms of Mercury in Fish
  • 1.4 The Ratio of Methylmercury to Total Mercury in Retail Fish
• 2.0 Hazard Characterisation: Health Hazards of Methylmercury
• 3.0 Health Benefits of Fish Consumption
• 4.0 Exposure Assessment
  • 4.1 Relative Popularity of Various Types of Fish in Canada
    • Table 4.1 (a): Fisheries Data (import, export, landing volumes) for Selected Fish Types.
    • Table 4.1 (b): Food consumption survey - Fish Results
  • 4.2 Fish Consumption Values for Canadians
  • 4.3 Meal Portion Sizes for Specific Types of Fish
  • 4.4 Levels of Mercury in Retail Fish
  • 4.5 Exposure Assessment: Estimation of Human Exposure to Methylmercury in Fish
    • Table 4.5: Probable Daily Intake of Methylmercury from Fish
• 5.0 Risk Characterisation
  • 5.1 Comparison of the PDI to the Provisional Tolerable Daily Intakes of Methylmercury
    • Table 5.1: % Probable Tolerable Daily Intake values.
• 6.0 Knowledge Gaps/Uncertainties
• 7.0 References (for both the health risk assessment/benefit document and the risk management strategy document)
• Appendix
  • Appendix I: Summary data for those samples of fish that were found by the Canadian Food Inspection Agency to contain, on average, approximately 0.2 ppm or less total mercury.
  • Appendix II: Summary data for those samples of fish that were found in various surveys to contain, on average, greater than 0.2 ppm total mercury.
  • Appendix III: List of those fish for which at least one individual sample was found by the Canadian Food Inspection Agency to contain greater than 0.5 ppm total mercury.
  • Appendix IV: Fish Consumption: Review of the Current Intake Figures for Canadian Consumers and Further Recommendations.

1.0 Background

1.1 Purpose of This Document

This paper has two principal purposes:

• To describe the reassessment of mercury and retail fish undertaken by the Bureau of Chemical Safety (BCS), Food Directorate (FD), Health Product and Food Branch (HPFB), Health Canada (HC).
• To serve as documentation, describing both risks and benefits, to support the development of appropriate amendments to the risk management strategies and policies.

It is important to note that this document pertains to retail fish only. "Sport fish" usually fall under the jurisdiction of provincial or territorial governments, and occasionally under the jurisdiction of the federal Department of Fisheries and Oceans (for inland waters that have been created or altered under permission of a Fisheries Act Authorization) or Parks Canada (for waters within federal park boundaries).

Risk assessment refers to a process by which the degree and nature of a risk can be characterised. Through risk assessment, it can be determined whether there is a need for risk management. The words "risk management" refer to the prevention and control options that can be employed to reduce risk.

The need for a management strategy to reduce the risk of unacceptable exposures to mercury from fish consumption was first identified by the Health Protection Branch (now the Health Products and Food Branch) of Health Canada in the late 1960s, when it established a standard^{Footnote 1} for mercury in fish. For those who regularly consume fish, fish constitutes the main source of dietary exposure to mercury, which can cause adverse health effects in humans at sufficiently high exposures. More recently, the risk management strategy was further developed by recognising the positive role that fish plays in overall nutrition and the risks to health that can arise from decreasing fish consumption.

The risk management approach at the time of preparing this reassessment paper consisted of two elements:

1. Application of a standard of 0.5 parts per million (ppm) total mercury to all commercially-sold fish except three piscivorous^{Footnote 2} fish, namely, shark, swordfish, and fresh/frozen tuna.
2. A consumer advisory, last re-issued in 2002, for the three fish that are exempted from the standard. The advisory recommends that the general adult population limit consumption of these fish to one meal per week and pregnant women, women of child bearing age and young children limit consumption to no more than one meal per month.

The standard of 0.5 ppm total mercury is enforceable by the Canadian Food Inspection Agency. The advisory, issued by Health Canada, was available on the websites of both Health Canada and the Canadian Food Inspection Agency.

The current reassessment was undertaken in order to consider any data that had been generated since the last assessment. The most up-to-date data available on the levels of mercury in various types of fish available for sale in Canada were considered. A review of the available information on retail fish consumption by Canadians was also undertaken. An initial review of this information suggested that the risk management approach as described above may not have been adequately protective and therefore required more detailed reassessment and revision.

Similar reviews of the risk management of mercury in fish have been conducted in other jurisdictions. For example, health authorities in the United States, Australia and New Zealand, the United Kingdom, and Ireland each released updated consumer advisories on the consumption of these fish in 2004. These updates appeared not long after a review of the health hazards of methylmercury by the Joint World Health Organization/Food and Agriculture Organization Expert Committee on Food Additives (JECFA)^{Footnote 3}.

1.2 Sources of Human Exposure to Mercury

Mercury is a naturally-occurring element that is released from soil and rocks through weathering, from volcanoes and forest fires, and is found in lakes and oceans. Certain human activities, such as combustion of fossil fuel and deforestation leading to soil erosion and lixiviation^{Footnote 4}, can also release mercury to the environment (Roulet et al., 1999; Health Canada, 2004b). Concentrations of mercury in ambient air and water are extremely low and so do not constitute a significant source of mercury exposure in humans (Clarkson et al., 2003; Goyer and Clarkson, 2001). Rather, humans are generally exposed to mercury through the diet.

Traces of mercury are present in all foods. Concentrations are quite low in fruits and vegetables because mercury uptake by plants from soil is low (European Commission, 2003). In contrast, levels are highest in certain types of fish although trace levels can be found in nearly all types of fish, which absorb mercury from the water and from the organisms that they consume. The Canadian Total Diet Studies, carried out by the Food Research Division (FRD), Bureau of Chemical Safety (BCS), in addition to other total diet studies conducted worldwide, have demonstrated that fish are the primary source of dietary mercury intake for the average population (Dabeka et al., 2003).

Other possible sources of exposure to very low levels of mercury include mercury amalgam dental fillings; certain vaccines containing thimerosal, a preservative containing ethylmercury (Clarkson et al., 2003); mercury-containing products (mercury thermometers or fluorescent lighting) that have been accidentally broken; and cigarette smoke. Exposure to mercury can also occur in occupational settings where mercury or mercury-containing compounds are used, for example, manufacturers of electric equipment, medical devices or automotive parts that contain mercury; chemical processing plants that use mercury; metal processing; municipal, medical, and hazardous waste incineration plants; medical facilities where equipment may contain elemental mercury; etc. (Health Canada, 2004).

1.3 The Chemical Forms of Mercury in Fish

Mercury exists in different chemical forms. Metallic mercury or elemental mercury is the silvery, shiny liquid that was once commonly used in, for example, thermometers. Other forms of mercury can be classified as either "inorganic" or "organic." Inorganic mercury includes inorganic mercury salts such as mercuric chloride (chemical symbol HgCl₂). Mercury is classified as organic when it is bound to a chemical species that is largely comprised of carbon. For example, ethylmercury (chemical symbol CH₃CH₂Hg⁺) is the active ingredient in a preservative used in vaccines. Methylmercury (chemical symbol CH₃Hg⁺) is another organic form of mercury that can be found in the aquatic environment, although normally at much lower levels than inorganic mercury (Environment Canada, 2005).

Mercury is not locked permanently in each of its different forms. Rather, various processes result in environmental cycling of mercury among its different chemical forms. For example, microbial activity is one process that can transform inorganic mercury to methylmercury.

With respect to the types of mercury found in fish, both inorganic and organic mercury may be present. However, methylmercury is the predominant form of mercury in fish. Its chemical properties allow it to rapidly diffuse and tightly bind to proteins in aquatic biota, including the proteins in the muscle tissue of fish. This leads to bioaccumulation in the fish, with the mercury level increasing with age of the fish. In turn, biomagnification along the food chain leads to higher mercury levels in piscivorous fish that are higher in the food chain than in fish and other organisms that are low in the food chain. Inorganic mercury can also bioaccumulate but to a far lesser extent than methylmercury.

1.4 The Ratio of Methylmercury to Total Mercury in Retail Fish

In the majority of cases, analyses of fish (and other food) samples involve the measurement of the sum of all mercury (or "total mercury") in the sample, regardless of the chemical form in which it is present. The analysis of individual mercury species, such as methylmercury, is more expensive than the analysis of total mercury.

From a human health perspective, it is the amount of methylmercury, rather than total mercury, that is of most interest, since methylmercury is much more readily absorbed into the human bloodstream. As a result, it is simply assumed, for the purposes of health risk assessments, that 100% of the "total mercury" is in the methylated form as methylmercury.

Several studies have measured the actual portion of total mercury that is present in fish as methylmercury. Levels can be variable, even among fish of the same species. For example, in four samples of sablefish, the percentage of total mercury that was in the organic form ranged from 81% to 95% (CFIA, 2003b). In samples of various species of tuna, the portion of total mercury present as methylmercury ranged from 70 to 77% (Yamashita et al., 2005) and 61 to 94% (Forsyth et al., 2004). In ten samples of swordfish, the percentage varied from 43 to 76% and in three samples of marlin, from 51 to 63% (Forsyth et al., 2004). Yamashita et al. (2005) reported similar results, with an average percentage of 72% in seven samples of swordfish and an average percentage of 43% in seven samples of blue marlin.

A wide range of percentages (30 to 79%) were also found in 37 samples of canned tuna (Forsyth et al., 2004).

In Forsyth et al. (2004), the authors concluded that, based on the data, a fixed conversion factor to estimate methylmercury levels from total mercury concentrations may not provide accurate estimates of actual methylmercury levels for health risk assessment purposes. Based on findings such as these, human health risk assessors at Health Canada continue to make the conservative assumption, for the purposes of assessments, that 100% of total mercury is present as methylmercury.

2.0 Hazard Characterisation: Health Hazards of Methylmercury

When ingested, organic mercury such as methylmercury, unlike elemental or inorganic forms, is almost completely absorbed from the gastrointestinal tract and distributed to all tissues. Methylmercury also readily crosses both the blood-brain barrier and the placenta. Some of the distributed methylmercury can be converted to inorganic mercury, mainly by microflora in the intestines (Clarkson, 2002).

A wide range of adverse health effects has been observed in humans following methylmercury exposure, the severity largely depending upon the magnitude of the dose and the duration of exposure. The central and peripheral nervous systems are generally considered to be the target organs of organic mercury-induced toxicity in humans.

For short- to long-term exposures to very high levels of methylmercury, the earliest neurological effects are non-specific symptoms such as paresthesia^{Footnote 5}, malaise and blurred vision. Subsequently, other signs such as concentric constriction of the visual field, deafness, dysarthria^{Footnote 6} and ataxia^{Footnote 7} appear. At very high exposures, methylmercury poisoning can lead to coma and death. For example, in a widely cited environmental contamination incident which occurred in the Minamata Bay area of Japan beginning in the 1950s, more than 900 people died after eating highly contaminated seafood (up to 40 ppm methylmercury) (National Institute for Minimata Disease). An additional 20,000 individuals were thought to have suffered various other forms of neurological damage in this episode.

Chronic (long-term) exposure among adults to methylmercury at low doses may not result in readily observable symptoms. There have been limited experimental studies suggesting that the nutritional benefits of selenium and omega 3 fatty acids, both commonly found in fish and/or marine mammals, might counteract the adverse effects of methylmercury to some degree (Health Canada, 2004). It should be noted, however, that animal studies on selenium and mercury have employed relatively high doses of mercury and selenium. The significance of these results to human dietary regimes where mercury and selenium exposures are much lower is not clear. With respect to dietary effects on methylmercury toxicity, many studies have investigated possible effects, either beneficial or antagonistic, of certain foods and dietary habits on the toxicity of mercury (Chapman & Chan, 2000). The authors noted that there is little evidence of nutrient effects at the population level, although there are a number of studies demonstrating that nutrients interact with the metabolism of mercury at the physiologic level.

Chronic exposure to mercury compounds may have negative effects on the immune system (Moszczysnki, 1997) and there is emerging evidence of potential cardiovascular effects (Stern, 2005). For example, the results of a recent epidemiological study of men in eastern Finland suggested that high mercury content in hair may be a risk factor for acute coronary events and cardiovascular disease, coronary heart disease, and all cause mortality (Virtanen et al., 2005). While the latter study suggests that effects related to cardiovascular disease are seen at similar hair mercury concentrations as those associated with effects related to neurodevelopment, these results should be regarded as preliminary.

The body of evidence available to date still suggests that the developing fetus is the most sensitive sub-population. Fetal exposure to methylmercury may affect the developing nervous system at substantially lower doses than in adults. Epidemiological studies, including recent studies in fish-eating populations in the Seychelle Islands in the Indian Ocean and the Faroe Islands in the North Atlantic Ocean, have employed very sensitive neurobehavioural tests to observe subtle neurodevelopmental effects in children. These studies have shown that nervous system domains involving fine motor function, attention, verbal learning and memory can be affected.

The World Health Organization (WHO, 1972) has established a provisional tolerable daily intake^{Footnote 8} (pTDI) of total mercury in adults (based on a 60 kg body weight) of 0.71 µg per kg body weight per day, of which no more than two-thirds or 0.47 µg/kg bw/day should be methylmercury (µg = microgram = 10-⁶ g). The values for total mercury and for methylmercury were thought to reflect the average ratio of total mercury to methylmercury in food. Bureau of Chemical Safety concurs with this assessment and has employed the 0.47 µg/kg bw/day value for the general population for many years.

In 2003, based on more recent findings on the effects of methylmercury on fetal/infant neurodevelopment, the Joint WHO/FAO Expert Committee on Food Additives (JECFA) recommended a provisional tolerable weekly intake (pTWI) for methylmercury of 1.6 µg/kg bw/week (equivalent to 0.23 µg methylmercury/kg bw/day) in order to sufficiently protect the developing fetus (WHO, 2003). This pTWI value was based on the observed association between maternal methylmercury exposure (estimated intake of 1.5 µg/kg bw/day) and developmental effects in children. While the neuropsychological effects measured in the children were described by the committee as not appreciably adverse, the ultimate severity or long term consequences involving nervous system domains associated with cognition and learning are unknown.

The JECFA pTWI is consistent with the pTDI for methylmercury of 0.20 µg/kg bw/day, established by Bureau of Chemical Safety in 1997, for women of child-bearing age and young children, in recognition of the increased susceptibility of the developing fetus and young children to the effects of methylmercury. The Bureau of Chemical Safety pTDI was derived based on a 10 ppm maternal hair methylmercury level as the approximate threshold (Grandjean et al., 1997) for neuropsychological dysfunctions. This value was first converted to a corresponding blood methylmercury level and then to a dietary methylmercury intake level using an equation employed by the U.S. EPA (1995). A 5-fold uncertainty factor was applied to this intake level to obtain a pTDI of methylmercury for women of child-bearing age and young children of 0.20 µg/kg bw/day (Feeley and Lo, 1998).

Bureau of Chemical Safety continues to periodically assess the pTDI, taking into consideration any new research findings on the toxicity of methylmercury.

The age at which neurodevelopment would not be as sensitive to the effects of methylmercury and at which sensitivity would be considered equivalent to that of the "general population" is not clearly established. By default, BCS tends to apply the lower pTDI value to young children up to the age of 12 years.

3.0 Health Benefits of Fish Consumption

Fish is considered an excellent source of high quality protein and is one of the best food sources of the long-chain omega-3 fatty acids, DHA and EPA^{Footnote 9}. The omega-3 fatty acids are required in the diet and are considered important to heart health, and brain and eye development. Fish is the most significant source of naturally-occurring Vitamin D which plays an important role in the body's use of calcium, a mineral required for sound teeth and bones. Fish and shellfish also contribute valuable minerals to the diet such as selenium, iodine, magnesium, iron and copper which the body requires for diverse functions such as growth, repair and proper functioning.

The Bureau of Nutritional Sciences (BNS), Food Directorate (FD), Health Products and Food Branch (HPFB), Health Canada (HC) estimates that a 100 g serving of salmon provides slightly more than the Adequate Intake (AI)^{Footnote 10} of Vitamin D for those between 2 and 50 years. A 100 g serving of canned tuna (just less than a drained 170-g can of tuna) furnishes 40% of the AI of Vitamin D, slightly exceeds the Recommended Dietary Allowance^{Footnote 11} for selenium, and significantly contributes to total intake of long-chain omega-fatty acids. Some studies have shown that one or more servings per week of fish can significantly decrease the risk of sudden cardiac death. There is some evidence that regular maternal fish consumption might have beneficial effects on fetal development.

The Office of Nutrition Policy and Promotion (ONPP), Health Products and Food Branch (HPFB), in consultation with Bureau of Nutritional Sciences, examined the evidence surrounding fatty fish and health benefits as part of the revision process for Canada's recently revised Food Guide to Healthy Eating. Reports and publications from well established expert groups and organizations, including the WHO/FAO, the American Dietary Guidelines Advisory Committee, the (UK) Scientific Advisory Committee on Nutrition/Committee on Toxicology, and the Harvard Center for Risk Analysis were considered. These recent publications specifically assessed the relationship between fish and cardiovascular health through systematic reviews or published meta analyses of prospective and intervention studies. For example, among the publications considered was one describing the results of a meta-analysis of cohort studies that examined the association between fish intake and coronary heart disease mortality (He et al., 2004). This study was considered to provide strong evidence of an association between the intake of fatty fish and cardiovascular health. Bureau of Nutritional Sciences also conducted a review of the evidence relating fish consumption and EPA and DHA intakes to cardiovascular health. The evidence of an association between reduced risk of sudden cardiac death and fish consumption with a frequency of at least once a week was consistent.

Office of Nutrition Policy and Promotion's recommendation to consume fish, particularly fatty fish, is based on scientific evidence and is included in the revised Food Guide to Healthy Eating.

One recent study has demonstrated that certain methods of cooking fish may reduce the health benefits of fish. Researchers in the U.S. studied fish intake and the risk of atrial fibrillation, a common arrhythmia (Mozaffarian et al., 2004). They found that among elderly adults, consumption of tuna or other broiled or baked fish, but not "fried fish or fish sandwiches", is associated with lower incidence of atrial fibrillation, even after adjustment for a variety of other factors. In a subsequent study (Mozaffarian et al., 2006), the researchers found that consumption of broiled or baked fish, including tuna, was associated with improved cardiac hemodynamics (blood flow patterns), but fried fish intake was associated with structural abnormalities indicative of systolic dysfunction (where the heart contracts less forcefully and cannot pump out as much of the blood that is returned to it) and potential coronary atherosclerosis (build-up of plaque in the arteries).

It has been postulated that omega-3 fatty acids and micronutrients such as iodine, iron, and choline in fish may modify the toxic action of methylmercury in fish-eating populations (Clarkson and Strain, 2003). A more recent study found that higher fish consumption during pregnancy was associated with better visual recognition memory, which was used as a measure of infant cognition. The benefit appeared to be greatest among infants whose mothers had consumed more fish relative to others in the study group but also had lower hair mercury levels at delivery. The authors suggested that these observations were due to the type of fish that they were consuming. In the same study, higher maternal hair mercury levels relative to other mothers in the study group were associated with lower cognition in infants (Oken et al., 2005). Observations such as these might help to explain why ongoing epidemiological studies of heavy fish consumers in the Seychelle Islands do not clearly reveal adverse effects on neurodevelopment.

In the assessment of risk of methylmercury exposure through fish consumption, Bureau of Chemical Safety did not consider, in a quantitative manner, the nutritional benefits of fish consumption against the risks of methylmercury exposure.

A recent publication considered three fish consumption scenarios as theoretical responses to fish consumption advisories in the U.S. (Cohen, 2005). The author quantified the relative impacts of mercury exposure and omega-3 fatty acids on cognitive development as measured by IQ. The author also considered the net impact of omega-3 fatty acids and mercury on coronary heart disease mortality and the net impact of fish consumption on stroke incidence and mortality. Their analysis of the aggregated health outcomes suggest that overall population well-being would improve if all women of childbearing age shifted to low mercury fish without altering their rate of fish consumption (and with no change in fish consumption among others segments of the population). However, for a consumption scenario in which all segments of the adult population reduced fish consumption by 17%, their analysis suggested that the overall net effect would be negative. This was due to the negative cardiovascular impact such as coronary heart disease among adults, which outweighed the small benefit to infant cognition that remained after the 17% reduction in fish consumption.

It is considered essential that any communications to the public include information on the health benefits of fish consumption alongside information on the risks of methylmercury exposure so that citizens can consider both the benefits and risks in reaching their own decisions about appropriate fish consumption. Studies on the nutritional benefits of fish are supportive of efforts to influence consumers' behaviour by modifying the types of fish regularly chosen rather than by decreasing overall fish consumption.

4.0 Exposure Assesssment

The exposure assessment evaluates the potential exposure to methylmercury from the consumption of retail fish. Potential exposure is a function of (1) the amount of fish that is consumed on a regular basis, and (2) the amount of methylmercury that is present in fish. Information relevant to these two components of the exposure assessment are provided in the following sections.

4.1 Relative Popularity of Various Types of Fish in Canada

The relative popularity of various types of fish in Canada can be evaluated using market share information. Since this information was not readily available to Health Canada, the volume of fish imports to Canada and the volume of domestic fish landings sold at retail in Canada (assumed to be the difference between what is landed and what is exported) were used as indicators of the relative amounts of different fish available in Canada on a yearly basis and, in turn, the relative popularity of different types of fish. This data was obtained from Fisheries and Oceans Canada (DFO, 2005 and 2006).

It has been assumed, for the fish listed in Table 4.1 (a), that all imports to Canada and domestic landings remaining in Canada enter the market for human consumption. A portion of imported fish (approximately 30% by volume but only 4% by value) is used to make fishmeal or fish oil for eventual use in the manufacture of livestock and fish feed (Agriculture and Agri-Food Canada, 2003b). However, most of the fish used for this purpose are small, bony fish such as anchovies, sardines, Atlantic mackerel, herring, and whiting.

Data for various years for selected types of fish are listed in Table 4.1 (a) below. Barracuda, escolar, halibut, orange roughy, marlin, sablefish, sea bass, shark, swordfish, and certain types of tuna are included in the table because each were found, in at least one of the surveys listed in Appendix II, to contain total mercury at average levels approximating or greater than 0.5 ppm (these average values are highlighted in the table of Appendix II). Shrimp and cod are also included in the following table to demonstrate the high end of the range of quantities for all fish in the database.

There are no domestic commercial fisheries for barracuda, escolar, marlin, grouper or orange roughy (DFO, 2006). Although these fish are imported to Canada, as evidenced by their availability to the CFIA during compliance monitoring activities, there was no relevant data in DFO's import database. The absence of import data in DFO's database suggests that imports of these species are negligible (DFO, 2006).

The data in Table 4.1(a) suggest that, among those fish listed, shrimp, cod, skipjack/bonito tuna (which are available canned), are the most popular. Sablefish, and presumably barracuda, escolar, marlin, and orange roughy, are among the least widely-consumed fish. The data suggest that the availability of shark, swordfish, and bluefin tuna (sold as fresh/frozen steaks) is much lower than that of canned tuna. Although data below seem to suggest that dogfish shark is consumed in North America, no information on dogfish consumption in Canada was located. One account states that there is no dogfish market in Canada (Hines, 2005). Rather, it is expected that most domestic dogfish is exported to Europe where it is served deep-fried (e.g. "fish and chips" in the United Kingdom), smoked (in Germany), etc. (Godknecht, 1999).

Table 4.1(a): Fisheries data for selected fish types in metric tonnes (1000 kg). Unless specifically noted, the source of this data is Fisheries and Oceans Canada (including DFO's Statistical Services: DFO, 2005). The remainder was provided directly to Health Canada by DFO (2006).

Piscivorous types of fish such as swordfish and shark have been considered in the past by Health Canada to be delicacies not frequently consumed by the Canadian population. Their higher cost and lower availability have contributed to their lower frequency of consumption relative to other types of fish.

A similar pattern in fish popularity is observed in the results of food surveys, of approximately 18 400 people from across Canada, that were conducted in the 1990's as part of the Health Canada / Canadian Heart Health Initiative (Health Canada and CHHI, 2004). Individuals were asked to recall what they had consumed in a 24-hour period. Information on the frequency of consumption of specified foods, including fish, was not obtained. However, the data allows calculation of the percentage of respondents consuming each particular type of fish, as well as typical adult serving sizes of various fish types (assuming that any fish reported to have been consumed during the 24-hour period was consumed at one meal, rather than during several meals over the day).

Table 4.1(b) includes some of those fish types that contained total mercury at levels, on average, greater than 0.5 ppm in Appendix II. Information on barracuda, escolar, orange roughy, and marlin were not obtained but it is expected that the results would be at least similar to, if not lower than, those for sablefish and sea bass. Some fish types that contained lower than 0.5 ppm on average are also included for comparison.

Of the fish listed in the Table 4.1(b), the most widely consumed on the survey day was canned light tuna (various species). Canned white tuna (albacore), which is more expensive than light tuna, was consumed by a smaller number of respondents. It also ranked, in terms of percentage of survey respondents consuming it, behind canned salmon, farmed salmon and wild salmon, and crab.

The percent values for the remaining fish in Table 4.1(b) suggest that tuna (as steaks) is more widely consumed than shark and that smaller percentages (0.1% or less) of the respondents consumed sablefish, sea bass, and swordfish.

Source:

Data generated in 2004 by the Bureau of Biostatistics and Computer Applications (BBCA) using information gathered by the Health Canada and Canadian Heart Health Initiative between 1990 and 1997.

Notes:

1. "nc" means that none of the survey respondents reported consuming this particular type of fish during the specified 24-hour period.
2. There are uncertainties in these estimates. The uncertainties would be higher in the case of EO values for fish that were consumed by only a small percentage of respondents.
3. The category "tuna" included survey food codes: "bluefin", "skipjack", and "yellowfin (albacore, ahi)."

The popularity of canned tuna is also evident in the data generated as part of a national study that was conducted in 1990 for then Health and Welfare Canada by Market Facts of Canada Limited (1991). The survey examined the consumption of fish and shellfish, regardless of source (restaurant, grocery store, or caught by self). Of all seafood meals reported to have been consumed by respondents over the three-month survey period, 14% were tuna, the majority, 95%, of which was canned. Salmon and shrimp were also popular (17 and 11% respectively of all seafood meals).

Canned tuna may be particularly popular among children. Although recent Canadian data regarding children's consumption of canned tuna was not readily available, for children 14 and under in the United States, the most commonly consumed seafood species is tuna (U.S. EPA, 2002), most of which is likely to be canned.

The other fish varieties that contained, on average, 0.5 ppm or greater total mercury did not appear in the Market Facts list of fish that were consumed in at least 3% of the meals. The number of shark purchases was noted to be too small to allow reporting and the reported swordfish meals were "prepared outside the home" (inferred to mean purchased and consumed in a restaurant) in almost all cases.

The data appearing in both Tables 4.1(a) and (b) are somewhat consistent with observations in the United States, where per-capita consumption data were used as an indicator of fish popularity. The ten most popular fish, based on data from the U.S. National Marine Fisheries Service (National Fisheries Institute, 2005), are shrimp, canned tuna, salmon, pollock, catfish, tilapia, crab, cod, clams, and flatfish.

4.2 Fish Consumption Values for Canadians

The amount of fish consumed over a given period of time is a function of fish meal size and the frequency at which fish meals are eaten; that is, the number of fish meals eaten during that period of time. Such information was gathered from the monthly diaries (covering three months) of 3,815 respondents to the Market Facts survey (1991). It was determined that the average consumption of fish per day of the month among those adults reporting to consume fish ("eaters only") is 22 grams. This "eaters-only" value has been employed by Bureau of Chemical Safety for assessments of most commercial fish and was recently reaffirmed in a review conducted by the Bureau of available information on fish consumption (Appendix IV)

For children, aged 1 to 5 years of age, average consumption of finfish was reported in the Market Facts report to be 10 grams per day of the month and for children 6 to 12 years of age, 14 grams per day of the month. In a review of available information on fish consumption in Canada (Appendix IV). Bureau of Chemical Safety recommended that these intake figures of 10 and 14 grams per day of the month for fish (excluding shellfish) would be appropriate for children aged 1 to 4 years and 5 to 11 years, respectively.

Certain fish varieties, such as shark, swordfish, and tuna steaks, that are not as widely available and are usually available in less abundant quantities, are considered to be consumed less frequently such that the daily intake over a month would be considerably less than 22 grams.

Similarly, it is possible that the mean consumption of canned tuna may be higher than 22 g/day, considering its wide availability, its cost, and convenience in terms of preparation for consumption (cooking is not required). Bureau of Chemical Safety has received inquiries from adult members of the public who report consuming as much as two cans of light tuna every day, which is equivalent to 240g per day (based on the weight of two drained 170-g cans).

The Market Facts survey (1991) found that the 75th percentile of "eaters" of seafood (finfish and shellfish combined) ingested 33 g/day while the 90th percentile of "eaters only" ingested 49g/day. The Bureau currently employs 40 g as an estimate of daily fish intake by adults who are at the high end of fish intakes (Appendix IV)

4.3 Meal Portion Sizes for Specific Types of Fish

A review of information on fish consumption, including portion sizes, was conducted by Bureau of Chemical Safety in 2004 (Appendix IV). The data reviewed included per capita disappearance data and dietary survey data for Canada and other countries as well as the published literature. A weight of 150 g was determined to be the best estimate of average portion size for adults. For children 5-11 and 1-4 years of age, the best estimates of portion sizes, based on the available data, were 125 g and 75 g respectively.

Although some of the data that were reviewed supported a smaller portion size of 100 g for adults, the data supporting 150 g was favoured as a reasonable, and somewhat more conservative (e.g. protective), estimate of average meal size for an adult. Similarly, the portion sizes for children are also considered conservative.

Bureau of Chemical Safety has defined the "reasonable daily intake" of "fish, shellfish" to be 100 g. Reasonable daily intake values for various foods appear in Part D, Schedule K of the Canadian Food and Drug Regulations and are mainly used as a basis upon which to determine the permissible levels at which certain vitamins and minerals may be added to food (various sections in Division 8, Part B and in Divisions 1 and 2 in Part D of the Regulations).

Since the promulgation of regulations requiring a "nutrition facts table" on the label of prepackaged foods, consumers have ready access to nutritional information that is provided on the basis of a "serving of stated size." However, the "serving of stated size" is not standardised. A review of frozen fish products in a grocery store in Ottawa, Ontario demonstrates that the food industry employs various values for the serving size depending on the product. For example, nutrition facts for frozen fillets were often provided on the basis of 142 g (1 fillet) although a 100 g and a 108 g serving size were also observed. Nutrition facts for frozen tuna and salmon burgers were on a 142 g (1 burger) basis and for other frozen fish, such as shrimp, lobster, calamari, octopus, and scallops, the reported serving sizes ranged from 49 g to 113 g, depending on the product.

The recently revised Health Canada Food Guide for Healthy Eating recommends at least 2 servings per week of fish. One serving of fish is defined as 75 g. However, the guide also recommends 2 to 3 food guide servings per day of "meat and alternatives", a category in which fish is included.

Portion sizes for canned tuna, in particular, may differ from the average portion size of finfish in general, due to tuna's availability in discrete containers (either cans or pouches). The size of the discrete package of tuna purchased will influence the portion size consumed. Cans of tuna, including their packing liquid (oil or water) typically weigh 170 g, and 120 g after draining, although other sizes are also available (e.g. 99 g and 198 g undrained). Pouches of tuna do not require draining and tend to contain less tuna than the cans (e.g. 85 g), although pouched tuna fillets in 142 g portions are available in the United States (their availability in Canada was unknown at the time of posting of this document) (Appendix IV)

4.4 Levels of Mercury in Retail Fish

Certain types of fish available at retail tend to contain lower levels of total mercury than others. The table shown in Appendix I lists those fish species that were recently analysed by the Canadian Food Inspection Agency (CFIA) and found to contain total mercury at average levels either approximating or less than 0.2 ppm (this value was arbitrarily chosen but is less than half the current standard of 0.5 ppm). A wide variety of fish appear on this list including shellfish (e.g. oyster, clams, scallops, mussels), shrimp, salmon, trout, herring, cod, flounder, lobster, crab, and lake whitefish.

Fish that are higher on their respective food chains tend to contain higher levels of total mercury. The table in Appendix II lists those fish for which mean levels of total mercury were greater than 0.2 ppm. Barracuda, escolar, marlin, orange roughy, sablefish, sea bass, shark, swordfish, and certain types of tuna were found (in at least one of the surveys listed in the table) to contain total mercury (on average) at levels approximating or greater than 0.5 ppm (these mean values are highlighted in the table in Appendix II). Marlin, shark, swordfish, and fresh tuna were each reported at least once to contain total mercury at average levels greater than 1.0 ppm.

It is important to note that the CFIA data do not necessarily represent the mercury content of species that are available at the retail level. Imported fish are sampled on the premises of the importers. Sampling of domestic fish occurs at the processing plants. Some of these fish may not reach Canadian retail due to compliance activities or, in the case of the domestic catch, due to the fact that a large portion is exported to other markets. In evaluating potential exposure of Canadians to mercury in different types of fish, it has been assumed that the CFIA data is representative of retail-level fish. While this may not be the case, a certain amount of non-compliant fish will be found in the Canadian retail marketplace since it is not possible to achieve 100% compliance.

It should also be noted that industrial processing or domestic cooking techniques for fish do not appreciably reduce the mercury concentrations in fish (Goyer, 2001). Therefore, in evaluations of potential exposure to mercury from fish, the total mercury concentrations in raw fish serve as a reasonable approximation of concentrations in fish that has been prepared (e.g. baked, poached, smoked, cured, etc.) for consumption.

4.5 Exposure Assessment: Estimation of Human Exposure to Methylmercury in Fish

To determine possible exposures to methylmercury from the consumption of various fish, average Probable Daily Intake (PDI) values were calculated for (1) the general adult population; (2) pregnant women/women of reproductive age; (3) children aged 5-11 years of age; and (4) children 1-4 years of age:

PDI (µg/kg bw/day) =

• fish muscle intake (g/day) x [methylmercury concentration (µg/g)] ÷ average body weight (kg)

Based on findings such as these, human health risk assessors at Health Canada continue to make the conservative assumption, for the purposes of assessments, that 100% of total mercury is present as methylmercury (see Section 1.4). Where CFIA data were used, it was assumed that the levels are representative of those found at the retail level, although this may not necessarily be the case (see Section 4.4). The consumption figures described in Section 4.2 were employed. Average body weights are from Nutrition Canada studies of the Canadian population: 60 kg for adults, 26.4 kg for children aged 5-11 years and 14.4 kg for children aged 1-4 years. More recent mean body weight values are higher; for example, 66.8 kg for females 19-30 years of age, approximately 31.8 kg for males aged 5-11 years, and 15.8 kg for females 1-4 years (Canadian Community Health Survey, cycle 2.2, 2004). The lower body weight values were used as default values for the calculations.

Notes to Table 4.5:

• Age groupings: "Gen" refers to adult men and to adult women who are not of reproductive age. "5-11 yrs" and "1-4 yrs" are the two age groupings for young children.
• The daily fish intakes employed in the calculations were 22 g/day (adults, male and female); 14 g/day (5-11 yrs); and 10 g/day (1-4 yrs).
• µg/g or micrograms per gram is equivalent to ppm or parts per million.
• In the case of cusk and "tuna, canned (species not specified)", the methylmercury concentration was available for only one sample (N=1)^{Footnote 13}.
• For "barracuda B not from the U.S.", samples from the 1990's (results not shown in above table) were from Oman, New Zealand, and Portugal.
• For marlin, the concentration used in the exposure calculations was the average of a relatively large data set (N=53) and the value is consistent with values for the other comparatively large data sets (see Appendix II).
• For sablefish, the 0.71 ppm value from the 2002-2003 CFIA survey was not employed in the exposure assessment as it was for samples available prior to the implementation of a fisheries management strategy to eliminate the source of sablefish containing elevated mercury. (CFIA, 2003b)
• For sea bass, although the value employed in the exposure calculations is not as recent as the 0.31 ppm value, it is based on a larger sample size.
• For shark, the concentration value employed in the calculations is recent but it is not based on a large sample size (N=12).

5.0 Risk Characterisation

5.1 Comparison of the PDI to the Provisional Tolerable Daily Intakes of Methylmercury

PDI values for methylmercury must be considered against the Provisional Tolerable Daily Intake (pTDI) for methylmercury (see Section 2.0) as a step in evaluating whether there is increased risk of adverse health effects from methylmercury exposure. A comparison of the two parameters is conveniently achieved by determining the ratio of the PDI to the pTDI, expressed as a percentage (PDI / pTDI x 100%). Values approaching or exceeding 100% identify those exposure scenarios where the toxicological reference value may be exceeded and that require more careful evaluation. Other factors that must be assessed in evaluating the risk and the degree of confidence in the assessment conclusions include the quality of the methylmercury concentration data, the likelihood of possible long-term exposure to the particular source of fish, etc.

Notes to Table 5.1:

• Age groupings: "Gen" refers to adult men and adult women who are not of reproductive age. "Women" refers to females of reproductive age. "5-11 yrs" and "1-4 yrs" are the two age groupings for young children.
• µg/g or micrograms per gram is equivalent to ppm or parts per million.
• In the case of cusk and "tuna, canned (species not specified)", the methylmercury concentration was available for only one sample.
• For "barracuda B not from the U.S.", samples from the 1990's (results not shown in above table) were from Oman, New Zealand, and Portugal.
• For marlin, the concentration used in the exposure calculations was the average of a relatively large data set (N=53) and the value is consistent with values for the other comparatively large data sets (see Appendix II).
• For sablefish, the 0.71 ppm value from the 2002-2003 CFIA survey was not employed in the exposure assessment as it was for samples available prior to the implementation of a fisheries management strategy to eliminate the source of sablefish containing elevated mercury. (CFIA, 2003b)
• For sea bass, although the value employed in the exposure calculations is not as recent as the 0.31 ppm value, it is based on a larger sample size.
• For shark, the concentration value employed in the calculations is recent but it is not based on a large sample size (N=12).
• For sauger and fresh yellowfin, the weighted concentrations for sampling years 2003-04 and 2004-05 were employed.

The PDI values contained in Table 5.1 can be considered to represent the average exposure to methylmercury if an individual consumed each specified fish over a long-term period to the exclusion of all others. For example, a 14.4 kg (1-4 yr old) child consuming an average of about 10 g each day (70 g per week) of cusk containing 0.35 ppm methylmercury would consume 0.24 micrograms of methylmercury per kilogram body weight per day (0.24 µg/kg bw/day). This value approximates the pTDI for methylmercury (the PDI is 122% of the pTDI). However, cusk is a fish that is not widely available. Stocks are seriously depleted and it is no longer commercially fished in Canada. It is available in limited quantities as a by-catch from other fisheries (DFO, 2005 b). Therefore, the likelihood of this exposure scenario, where a small child consumes cusk on a regular basis, is considered remote. In addition, confidence in the methylmercury concentration value is low because only one sample was analysed.

5.2 Characterisation of the Risk

For members of the general adult population, based on the available data and the results shown in Tables 4.5(a) and 5.1(a), swordfish is the only fish for which regular weekly consumption would result in a PDI that exceeds the pTDI (%pTDI of 143%).

Regular consumption, by women of child-bearing age, of barracuda (from the U.S.), escolar, marlin, sea bass, shark, swordfish, bigeye tuna, and "fresh" tuna containing mercury at the levels shown in Tables 4.5(a) and 5.1(a) could result in the methylmercury pTDI being exceeded. These fish contained an average of 0.54 ppm or more total mercury, assumed to be 100% methylmercury. The intake of methylmercury from the regular consumption of grouper, orange roughy, and walleye, although somewhat high (80% of the pTDI), would not cause the pTDI to be exceeded. It is important to bear in mind that many of these fish were sampled early in the chain of distribution rather than at retail.

For young children (12 years of age and younger), regular consumption of fish that contain on average 0.3 ppm or more total mercury (assumed to be 100% methylmercury) could result in the PTDI being exceeded, as demonstrated in Table 5.1. In consideration of the relative popularity of different types of fish in general (as discussed in previous sections), it is not considered likely that a child would regularly consume the fish types listed in Table 5.1 except in the case of canned albacore tuna.

Canned tuna represents a unique case. It is widely available and relatively inexpensive. As discussed in Section 4.1, canned light tuna (e.g. yellowfin, skipjack, etc.) is a commonly consumed fish. However, its low mercury concentration leads to the estimation that it would not cause unacceptably high exposure to methylmercury in the diet. Canned albacore ("white") tuna, which is more expensive than canned light tuna, is less popular than canned light tuna. Nonetheless, the possibility exists that those regularly consuming canned tuna may be consistently choosing canned albacore tuna, which has a higher mercury content. This could potentially lead to an unacceptable high methylmercury exposure.

6.0 Knowledge Gaps/Undertainties

Certain pieces of information would allow the risk assessment to be further refined. These include the following:

• Information on the amount of specific types of piscivorous or predatory fish consumed by Canadians and whether any have particular importance to the diets of certain groups within the population. Such information would also further inform the development of the risk management strategy and risk communications.
• Total mercury levels in a larger number of recent samples of certain types of fish (e.g. orange roughy, barracuda, etc.) which would provide results that are more representative of that which is available at retail and which would decrease uncertainty in the risk assessment for those types of fish.
• Actual methylmercury concentrations in the fish samples which would reduce uncertainty in the risk assessment. Applying the assumption that all of the mercury in fish is present as methylmercury results in a more conservative assessment.
• Information on mercury levels in human blood in Canada could serve as an index of exposure to mercury, the largest source of which is fish for many people. Such information would be an important consideration in the context of both risk assessment and risk management. Statistics Canada, with the support of Health Canada, is collecting socioeconomic and demographic information and physical measurements from a sample of the Canadian population. Statistics Canada is also considering other variables such as blood mercury levels from a sub-sample of the population (Statistics Canada, 2004).
• Halibut, sea bass, grouper, and walleye were found to contain average mercury at levels somewhat similar to those found in canned albacore tuna. In the case of canned albacore tuna, the available information led to the conclusion that consumption of canned albacore tuna may be higher than other seafoods, which could lead to an unacceptably high exposure to mercury. BCS did not have information that would allow it to assess whether there are groups in the Canadian population that consume halibut, sea bass, grouper, or walleye at relatively high rates.

7.0 References

NOTE: Unless otherwise noted, all on-line documents were last accessed on February 23, 2007.

Agreement on the Application of Sanitary and Phytosanitary Measures. 1994.

Agriculture and Agri-Food Canada (AAFC). 2003. Fish and Seafood Market Overview - Republic of Korea.

Agriculture and Agri-Food Canada (AAFC). 2003a. Profile and Demand of Asian and Hispanic Food In Metro Atlanta.

Agriculture and Agri-Food Canada (AAFC). 2003b. Canada's Fish and Seafood Industry.

Atuna. 2004. Tuna Species Datasheet - Albacore Tuna.

Dabeka, R.W., A.D. McKenzie and P. Bradley. 2003. Survey of total mercury in total diet food composites and an estimation of the dietary intake of mercury by adults and children from two Canadian cities, 1998-2000. Food Additives and Contaminants. 20(7): 629-638

British Columbia Ministry of Environment. 2001. "Aquatic Pathogens."

Bumble Bee Seafoods. 2003. Portraits of Tuna Types.

Bureau of Chemical Safety, Health Canada. 2004. Fish Consumption: Review and Recommendation of Current Intake Figures for Canadian Consumers

Canadian Food Inspection Agency. 2002. Mercury in imported canned tuna - by species: 2002 CFIA Product verification project. Email attachment, from P. Chan to E. Desjardins regarding 'Request for Canned Tuna Mercury Data. Unpublished document.

Canadian Food Inspection Agency - Burnaby Laboratory. 2002b. Mercury Results for sampling period August 1990 - March 2002. Email attachment, from S. Liem to G. McGregor regarding 'Hg results on the follow-up investigation'.

Canadian Food Inspection Agency. 2002c. Fish Products Standards and Methods Manual, Chapter 2, Standard 1, Canned Tuna Standard.

Canadian Food Inspection Agency. 2002-2003 and 2003-2004. Total mercury results for sampling periods April 1, 2002 to March 31, 2003 and April 1, 2003 to October 7, 2004.

Canadian Food Inspection Agency. 2003. Contaminants- Western Area- Mercury Sampling. Included as an addendum to 'Agenda: HC/CFA Fish Chemistry Working Group Meeting, February 26, 2003, 9:00-12:00, Banting Bldg (Room C204).'

Canadian Food Inspection Agency. 2003b. Draft Sablefish Mercury Report - Investigation of mercury in B.C. Sablefish sampled between October 2002 and November 2003.

Canadian Food Inspection Agency. 2003c. Certification Requirements Mexico.

Canadian Food Inspection Agency. 2005. Certification Requirements Japan.

Canadian Food Inspection Agency. Undated. "List of Canadian Acceptable Common Names for Fish and Seafood."

Center for Food, Nutrition, and Agriculture Policy. 2005. "Executive Summary: A National Opinion Survey Conducted for The Center for Food, Nutrition, and Agriculture Policy." University of Maryland, College Park, U.S.A.

Chapman, L., and H.M. Chan. 2000. The Influence of Nutrition on Methyl Mercury Intoxication. Environmental Health Perspectives: Supplements. 108(Supplement 1): 29-56.

Clarkson, T.W. 2002. The Three Modern Faces of Mercury. Environmental Health Perspectives: Supplements. 110( Supplement 1): 11-23.

Clarkson, T.W., Magos, L., and G.J. Myers. 2003. The toxicology of mercury- current exposures and clinical manifestations. New England Journal of Medicine. 349(18): 1731-1737.

Clarkson, T.W., Strain, J.J. 2003. Nutritional Factors May Modify the Toxic Action of Methyl Mercury in Fish-Eating Populations. Journal of Nutrition. 133:1539S-43S.

Codex Alimentarius Commission. 2005. Codex Committee on Food Additives and Contaminants - Discussion Paper on Guideline Levels for Methylmercury in Fish.

Cohen, J.T. January 2006. Matters of the Heart and Mind: Risk-Risk Tradeoffs in Eating Fish Containing Methylmercury. In: Risk in Perspective, a newsletter of the Harvard Center for Risk Analysis. Volume 14, Issue 1.

Commission of the European Communities. 2001. Official Journal of the European Communities. Commission Regulation (EC) No 466/2001 of 8 March 2001 setting maximum levels for certain contaminants in food stuffs.

Commission of the European Communities. 2002. Official Journal of the European Communities. Commission Regulation (EC) No 221/2002 of 6 February 2002 amending Regulation (EC) 466/2001 setting maximum levels for certain contaminants in food stuffs.

Dabeka, R.W., A.D. McKenzie and P. Bradley. 2003. Survey of total mercury in total diet food composites and an estimation of the dietary intake of mercury by adults and children from two Canadian cities, 1998-2000. Food Additives and Contaminants. 20(7):629-638.

Dabeka, R.W., A.D. McKenzie, D.S. Forsyth, H.B.S. Conacher. 2004. Survey of total mercury in some edible fish and shellfish species collected in Canada in 2002. Food Additives and Contaminants. 21(5):434-440.

Department of Fisheries and Oceans. 2006. Fisheries Data provided in a personal communication from Warren Brown, DFO, e-mail January 13, 2006.

Department of Fisheries and Oceans. 2005. Statistical Services. Trade - Canadian Imports - Domestic imports of selected commodities by species group and species.
AND
Commercial - Landings (seafisheries and freshwater).

Department of Fisheries and Oceans. 2005b. Cusk: At a Glance.

Department of Fisheries and Oceans. 2005c. Canadian Atlantic Pelagic Shark Integrated Fisheries Management Plan.

Department of Fisheries and Oceans. 2005d. Atlantic Fisheries Research Document 96/75 - Abstract.

Department of Fisheries and Oceans. 2005e. The Canadian Atlantic Swordfish (Xiphias gladius) Fishery.

Department of Fisheries and Oceans. 1989. Data provided directly to the Chemical Evaluation Division (now the Chemical Health Hazard Assessment Division). Cited in "Mercury in Shark: Background information concerning mercury in fish guideline." Unpublished document.

Department of Justice Canada. 2004. Fish Inspection Regulations; Part IV Canned Fish.

Department of Justice Canada. 2005. Food and Drugs Act and Related Regulations.

Environment Canada. 2004. Mercury and the Environment - Fish Consumption.

Environment Canada. 2005. Canadian Water Quality Guidelines: Inorganic Mercury and Methylmercury.

European Commission, Health and Consumer Protection Directorate-General. April 25, 2003. Opinion of the Scientific Committee on Animal Nutrition on Undesirable substance in feed. (Last accessed July 5, 2004)

Feeley, M.M. and Lo, M.-T. Risk Assessment for Mercury in Health Canada - Development of the Provisional Tolerable Daily Intake (pTDI) Value. In: Proceedings of the Conference on Mercury in Eastern Canada and the Northeast States, 21-23 September, 1998. Ed. by Pilgrim, W., Burgess, N., Giguère, M.-F. 1998. Unpublished.

Florida Museum of Natural History. Undated. Biological Profile: Nassau Grouper.

Food and Agriculture Organization of the United Nations (FAO). 2005. Fisheries Global Information System.
Species Fact Sheet on Dicentrarchus labrax.
Species Fact Sheet on Dissostichus eleginoides.

Food Safety Authority of Ireland (FSAI). 2004. FSAI Issues Guidelines on Consumption of Shark, Swordfish, Marlin and Tuna.

Food Standards Australia and New Zealand (FSANZ). 2004. "FSANZ updates advice on mercury in fish (Australia only)."

Food Standard Australia New Zealand (FSANZ). Undated. Standard 1.4.1 Contaminants and Natural Toxicants.

Food Standards Agency of the United Kingdom. February 4, 2004. News Centre: Be My Healthy Valentine.

Food Standards Agency of the United Kingdom. July 24, 2003. Mercury in imported fish and shellfish, UK farmed fish and their products.

Food Standards Agency of the United Kingdom. Undated. Mercury in fish: your questions answered.

Food Standards Authority of Ireland (FSAI). 2004. "FSAI Issues Guidelines on Consumption of Shark, Swordfish, Marlin and Tuna."

Forsyth, D., Casey, V., Dabeka, R. W. and McKenzie, A. 2004. Methylmercury levels in predatory fish species marketed in Canada. Food Additives and Contaminants. 21(9): 849-856.

General Agreement on Tariffs and Trade (GATT). 1994.

Godknecht, A.J. June 15, 1999. "Requiem for smoked dogfish." Article appearing on the website of Shark Info, a media service based in Switzerland. Available on-line:
http://www.sharkinfo.ch/SI2_99e/requiem.html

Goyer, R.A. and T.W. Clarkson. 2001. Chapter 23: Toxic Effects of Metals. In Casarett and Doull's Toxicology: The Basic Science of Poisons. Ed. by C.D. Klaasen. Pages 811-867.

Grandjean, P., Weihe, P., White, R.F., Debes, F. Araki, S., Yokoyama, K., Murata, K., Sorensen, N., Dahl, R. And Jorgensen, P.L. 1997. Cognitive deficit in 7-year-old children with prenatal exposure to methylmercury. Neurotoxicology and Teratology. 19(6):417-428.

Harada, Masazumi. 1995. Minamata Disease: Methylmercury poisoning in Japan caused by environmental polluation. Critical Reviews in Toxicology. 25(1): 1-24.

He, K., Song, Y., Daviglus, M.L., Liu, K., Van Horn, L., Dyer, A.R., and Greenland, P. 2004. Accumulated Evidence on Fish Consumption and Coronary Heart Disease Mortality: A Meta-Analysis of Cohort Studies. Circulation. 109: 2705-2711.

Health Canada. 2002. Advisory - Information on mercury levels in fish.

Health Canada. 2004. Mercury B Your Health and the Environment: A Resource Tool. Health Canada Mercury Issues Task Group.

Health Canada. 2004a. It's Your Health: Mercury and human health.

Health Canada and Canadian Heart Health Initiative. June 10, 2004. Intake of Fish from Federal-Provincial Nutrition Surveys. Data generated by M. Vigneault, Division of Statistics and Epidemiology, Bureau of Biostatistics and Computer Applications, Health Products and Food Branch. Unpublished document.

Hines, S. June 2, 2005. "Dogfish down: Once plentiful fish now dwindling; June conference to address the issue." University Week , Volume 22, No. 30, University of Washington Office of News and Information.

International Hydrographic Organization (IHO). September 2004. IHO Hydrographic Committee on Antarctica.

Japanese Ministry of Health, Labour and Welfare. 2003. Advice for Pregnant Women on Fish Consumption concerning Mercury Contamination.

Kumar, M., Aalbersberg, B., Mosley, L. (Institute for Health Sciences, Suva, Fiji Island). Undated. Mercury Levels in Fijian Seafoods and Potential Health Implication - Report for the World Health Organization.

Lim, Kelvin K.P.; Low, Jeffrey K. Y. , Undated; A Guide to Common Marine Fishes of Singapore.

Market Facts of Canada. March 1991. Research Report: National Seafood Consumption Study. Conducted for: Health and Welfare Canada. #C388/JdeB.

Moszczynski, P. 1997. Mercury compounds and the immune system: a review. International Journal of Occupational Medicine and Environmental Health 10(3): 247 258.

Mozaffarian, D., Gottdiener, J.S., Siscovick, D.S. 2006. Intake of Tuna or Other Broiled or Baked Fish Versus Fried Fish and Cardiac Structure, Function, and Hemodynamics. The American Journal of Cardiology. 97(2): 216 222.

Mozaffarian, D., Psaty, B.M., Rimm, E.B., Lemaitre, R.N., Burke, G.L., Lyles, M.F., Lefkowitz, D., Siscovick, D.S. 2004. Fish Intake and Risk of Incident Atrial Fibrillation. Circulation. 110(4): 368-373.

National Fisheries Institute (NFI), 2005, "Top 10 U.S. Consumption by Species Chart." (The NFI is an advocacy group of the U.S. seafood industry.)

National Institute for Minimata Disease. Undated. Minamata Disease Questions and Answers.

North American Free Trade Agreement. 1994. Section B Sanitary and Phytosanitary Measures..

Oken, E., Wright, R.O., Kleinman, K.P., Bellinger, D., Amarasiriwardena, C.J., Hu, H., Rich-Edwards, J.W., Gillman, M.W. 2005. Maternal Fish Consumption, Hair Mercury, and Infant Cognition in a U.S. Cohort. Environmental Health Perspectives. 113(10): 1376-1380.

Ottawa Citizen, May 27, 2001

Region of Waterloo Public Health. March 2005. Fish Facts for Families: Choose Fish Low in Mercury.

Robinson, J., Schroff, J. November 2004. Observations on the levels of total mercury (Hg) and selenium (Se) in species common to the artisanal fisheries of Seychelles. Seychelles Medical and Dental Journal, Special Issue, Volume 7, Number 1.

Roulet, M., Lucotte, M., Farella, N., Serique, G., Coelho, H., Sousa Passos, C.J., de Jesus da Silva, E., Scavone de Andrade, P., Mergler, D., Guimarães, J.-R.D., and Amorim, M. 1999. Effects of Recent Human Colonization on the Presence of Mercury in Amazonian Ecosystems. Water, Air, & Soil Pollution. 112:297-313.

Shakeel, Hassan and Ahmed, Hudha. March 1996. Exploitation of reef resources, grouper and other food fishes. (Paper posted to the website of the Secretariat of the Pacific Community.)

Statistics Canada. 2004. Canadian Health Measures Survey.

Stern, A.H. 2005. A review of the studies of the cardiovascular health effects of methylmercury with consideration of their suitability for risk assessment. Environmental Research 98:113-142.

United States Department of Commerce. 1999. National Oceanic and Atmospheric Administration Technical Memorandum NMFS-NE-143. Essential Fish Habitat Source Document: Black Sea Bass, Centropristis striata, Life History and Habitat Characteristics.

United States Environmental Protection Agency. 1995. Reference Dose for Methylmercury. Integrated Risk Information System (IRIS). Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati, OH. Updated version of document.

United States Environmental Protection Agency. August 2002. Estimated Per capita fish consumption in the United States (EPA 821 C 02 003).

United States Food and Drug Administration (U.S. FDA). 2004b. Mercury in Fish: FDA Monitoring Program (1990 2004).

United States Food and Drug Administration (FDA) and the Environmental Protection Agency (EPA). 2004. "What You Need to Know About Mercury in Fish and Shellfish."

United States Tuna Foundation. 2005. About Tuna.

Virtanen, J.K., Voutilainen, S., Rissanen, T.H., Mursu, J., Tuomainen, T.-P., Korhonen, M.J., Valkonen, V.-P., Seppänen, K., Laukkanen, J.A., and Salonen, J.T. 2005. Mercury, Fish Oils, and Risk of Acute Coronary Events and Cardiovascular Disease, Coronary Heart Disease, and All Cause Mortality in Men in Eastern Finland. Arteriosclerosis, Thrombosis, and Vascular Biology 25:228 233.

World Health Organization (WHO). 2003. Joint FAO/WHO Expert Committee on Food Additives, Sixty-first meeting 10-19 June 2003, Summary and Conclusions.

World Health Organization (WHO). 1999.

1. International Programme on Chemical Safety, Environmental Health Criteria 210. Principles for the Assessment of Risks to Human Health from Exposure to Chemicals.
2. And the WHO and FAO Joint Project to Update and Consolidate Principles and Methods for the Risk Assessment of Chemicals in Food.

World Health Organization (WHO). 1972. WHO Food Additive Series: 52. Methylmercury (addendum).

World Trade Organization. Undated. SPS Handbook Training Module - Introduction.

Yamashita, Y., Omura, Y., and Okazaki, E. 2005. Total mercury and methylmercury levels in commercially important fishes in Japan. Fisheries Science. 71: 1029-1035.

Yeh, Shinn Pyng; Yang, Tony; Chu, Tah Wei. Undated. Marine Fish Seed Industry In Taiwan.

Appendix I

APPENDIX II

Notes:

• Blank entries indicate that the data was not available to allow calculation of the specified parameter or that only 1 sample was analysed.
• For sources of data where CFIA has been cited but the particular year does not appear in the list of references, the data has been provided directly to Health Canada as part of this review.
• Note that the tuna samples whose source is cited as either "CFIA, 02-03" or "CFIA, 03-04" were not originally identified in the tables provided to Health Canada as to whether they were canned or fresh/frozen. A request for this information has been made to the CFIA and to date, information has been received for yellowfin tuna and samples identified as "species not specified."
• For the CFIA tuna results, samples for which the species is not identified may be so either because the required information was not entered into the results table or because the information was simply not available to the person sampling to allow them to identify the tuna species.
• The CFIA results from Burnaby were reported separately for imported samples and for domestic samples, hence they are reported separately in the above table.
• For swordfish samples analysed by the CFIA in 2003-2004, 11 samples were actually listed in the original data table, one of which was reported to contain 0 ppm total mercury. This sample was not included in the above table (hence 10 samples are reported).
• For the shark samples described in Dabeka et al., 2004, the raw data files describe results for 12 samples although the publication indicates that 13 were analysed. The summary results in the above table are based on the raw data files provided by Dr. Dabeka.

Appendix III

Notes:

1. The samples described in Dabeka et al., 2003, were collected in 2002. For canned tuna in particular, the juice was NOT drained. Of the 13 samples of "Tuna (fresh or frozen)", one sample was specified as "yellowfin."
2. For sablefish, the 0.71 ppm value from the 2002-2003 CFIA survey was not employed in the exposure assessment as it was for samples available prior to the implementation of a fisheries management strategy to eliminate the source of sablefish containing elevated mercury.

Appendix IV

Fish Consumption: Review of the Current Intake Figures for Canadian Consumers and Further Recommendations

• Executive Summary
• Introduction
• Review of Data Sources
• Recommended Figures
  • Portion Size
  • Frequency of Consumption
  • Average Daily Consumption
• Distinct Population Subgroups
  • Coastal Provinces
  • Adult Men
  • Children
  • Pregnant Women and Women of Childbearing Age
  • Subsistence and Recreational Fishers
• References
• Appendix A. Summary of Data Sources

Bureau of Chemical Safety, Food Directorate
Health Products and Food Branch
Health Canada

Executive Summary

Following a review and analysis of the data on the daily intake figures and portion sizes of finfish, shellfish, and seafood (the combination of all finfish and shellfish), the following figures are recommended for the Canadian population. These figures will provide the basis for the more accurate estimate of contaminant exposure from the consumption of finfish and shellfish.

"All-persons" intake figures consider consumption values reported by entire number of respondents from a particular survey or study (i.e. those that reported eating and not eating the food in question). "Eaters-only" intake figures are averaged over the number of respondents that reported eating the food in question.

Table i. Consumption Figures for the Entire Population (all ages)

Fish Consumption: Review of Intake Figures for Canadian Consumers

Introduction

The Bureau of Chemical Safety (BCS), Food Directorate has employed the intake figure of 22 g/day to represent the finfish consumption of commercial consumers and 40 g/day to represent the finfish consumption of sportfish or subsistence consumers for a number of years. The present review was undertaken to verify and update these figures with current information that is available regarding fish consumption in Canada. This paper presents the key sources of data pertaining to Canadian fish consumption, discusses and evaluates the validity of the applicable data, and provides recommendations for appropriate fish consumption figures and portion sizes for the Canadian population. The purpose of this review paper was to develop standardized consumption figures that could be used to more accurately determine contaminant exposure from finfish and shellfish consumption as required by human health risk assessments.

Throughout this document, the term "seafood" is used to refer to the combination of all finfish and shellfish. However, there may be some diversions from this practice when the results of other studies are discussed, as they may not have used similar terminology. Additionally, the terms "intake" and "consumption" are used interchangeably throughout this paper; both refer to the amount of food eaten. "All-persons" intake figures consider consumption values reported by entire number of survey respondents (i.e. those that reported eating and not eating the food in question). "Eaters-only" intake figures are averaged over the number of survey respondents that reported eating the food in question. Additionally, all figures presented herein have been rounded to the nearest two significant figures except values in the hundreds, which have been rounded to three significant figures.

Review of Data Sources

Canadian data sources of fish consumption information are summarized in the following section. Intake data from other countries with similar geographical, cultural, and socioeconomic characteristics to Canada's can provide information on seafood consumption, but Canadian data are given priority unless they are inadequate or unavailable. Suitable sources of Canadian consumption data are available and will hence be the focus of this paper. These reports are summarized below as well as in Appendix A (Table A1).

1. Statistics Canada publishes an annual report, Canada Food Statistics (2002), which summarizes the food disappearance figures for Canada. The 2002 disappearance figure for the edible portion of seafood is 9.9 kg/person or 27 g/day, which consists of 4.3 g/day of fresh/frozen fish, 0.44 g/day of freshwater fish, 3.1 g/day of processed seafish, and 2.1 g/day of shellfish. Per capita data are used to represent general population consumption trends but do not permit individual intake patterns to be deduced. For example, eaters are not distinguished from non-eaters and no information can be gleaned on the consumption habits of specific consumer groups or consumption frequency. Furthermore, these disappearance figures represent the net supply of food available for consumption (imports, exports, and manufacturing uses are accounted for) but do not consider losses at the individual or household level due to spoilage, preparation, cooking, and waste. Therefore, unadjusted disappearance data tends to overestimate actual consumption.

The per capita consumption of seafood by Canadians in 2002 was, in the same report, adjusted for the estimated retail, household, cooking and plate loss and determined to be 7.2 kg/year or 20 g/day. Broken into categories, this was equal to 2.9 g/day of fresh/frozen fish, 0.29 g/day of freshwater fish, 2.7 g/day of processed seafish, and 1.4 g/day of shellfish.

2. Nutrition Canada (1973) conducted a nation-wide consumption survey for Health and Welfare Canada between 1970 and 1972. The Nutrition Canada Survey collected data from over 19 000 residents from ten provinces^{Footnote 14} and various aboriginal and Inuit groups; however, some regions, such as the Atlantic Provinces, were not as well-represented as others (Brulé, 1996). Respondents were interviewed in person at a community survey centre and clinical, anthropometric, and dental examinations were also undertaken. The amount of individual foods consumed over the last 24 hours was recorded using portion models and the frequency at which certain foods were consumed over the previous month was also noted.

Surveys based on 24-hour recall information do not capture normal variations in an individual diet and only capture information on the foods eaten on the survey day. While only a small proportion of people may report eating fish on the particular survey day, more people are likely to eat fish over the course of a week or month. Therefore, a particular food may be over or under represented relative to the actual frequency and magnitude of consumption. As a result, eaters-only data from 24-hour recall surveys are not deemed useful for determining average daily consumption unless frequency of consumption is considered. On the positive side, the sample size of this 24-hour recall survey was large and the eaters-only intake figures provide valuable information on portion size.

Fish intake data from this survey are summarized below. Table 1 lists the mean eaters-only and all-persons consumption figures of the entire population (> 1 year old) and for 1 to 4 and 5 to 11 year-old children.

The disparity between the all-persons and eaters-only intake figures that are highlighted in Table 1 is a result of seafood not being consumed by the majority of individuals combined with consumption data from high-end consumers (US EPA, 1999). Eaters-only seafood figures can therefore be much higher than those represented by all-persons figures, so caution should be exercised when using 24-hour recall data to estimate daily consumption means.

3. Data from the 1986 Statistics Canada Family Food Expenditure Survey were analyzed in a publication from the Department of Fisheries and Oceans (Sabry, 1990). This survey gathered data on household food purchases and used it as a proxy for food consumption. In 1986, the total amount of seafood purchased per household was 319 g/week and a household included 2.53 people. Average per diem consumption for each person in the household was calculated to be 18 g/day. This figure, however, represents the retail weight of the seafood and not the amount of seafood that is actually consumed or by whom. Nor does it represent the amount of seafood lost to spoilage, preparation, cooking, and waste. In addition, these figures do not consider seafood consumption from other sources such as restaurants or recreational fishing.

The most popular fresh or frozen seafood species reported by household were: salmon (29 g/week), flounder and sole (29 g/week), cod (21 g/week), haddock (12 g/week), all other sea fish (46 g/week), and freshwater fish (30 g/week). A total of 72 g/week of canned fish was consumed by each household, with the most popular being salmon (32 g/week) and tuna (30 g/week). Shellfish was consumed at a rate of 52 g/week by each household.

Geographical differences in seafood consumption were observed across the country. Consumption in the Atlantic and Pacific coastal regions was greater than in the Prairie Provinces. Shellfish consumption was highest in Prince Edward Island, Nova Scotia, New Brunswick, and British Columbia. The largest amount of cod was purchased by families in Newfoundland and the greatest amount of salmon was purchased in British Columbia.

4. Food consumption surveys were conducted in ten Canadian provinces (see footnote 1) in the 1990's as part of the joint Health Canada and Canadian Heart Health Initiative study. The interviewees consisted of approximately 2000 individuals from each province between the ages of 18 and 74 years, excluding residents of institutions, military camps, and First Nations reserves. Infants, children, and teenagers were also excluded. The surveys were based on 24-hour dietary recall data, a food frequency questionnaire, as well as height and weight measurements (Brulé, 1996). The same advantages and disadvantages of 24-hour recall surveys that were discussed with respect to the Nutrition Canada Survey (1973) also apply to these studies.

Results summaries from the Québec (Santé Québec, 1992) and Nova Scotia (Nova Scotia Department of Health, 1993) surveys have been provided to BCS and the pertinent results are summarized in Tables 2 and 3.

The data in the tables above illustrate that 11% of Québécois surveyed consumed finfish and 4.9% consumed shellfish in the previous 24-hour period. In Nova Scotia, 21% of the population surveyed consumed finfish and 11% consumed shellfish. Detailed information about the questionnaire and which fish species were classified into each of the two categories ('>' and '<' 6% fat) was not provided.

5. In the early 1990's Health and Welfare Canada commissioned an independent marketing organization, Market Facts Canada, to conduct a detailed survey on the fish consumption habits of Canadians aged one year and older. This survey (Market Facts, 1991) collected data from 3815 respondents nationwide over a 3-month period from December 1990 to February 1991. Each respondent was asked to record their seafood consumption, including which types of seafood were consumed, the form it was in when acquired (e.g. canned, fresh bought, fresh captured, frozen cooked), how it was prepared, the quantity eaten, and its origins (e.g. Canadian river/lake, Caribbean, fish farm, etc.). This information was recorded in diary format over a one-month period.

The Market Facts (1991) study summarized that 81% of Canadians reported eating seafood during the month, which is consistent with a Canadian Seafood Advisory Council estimate quoted by Sabry (1990) stating that 84% of Canadians consume fish. Of those that reported eating seafood, 49% consumed only finfish, 3% ate only shellfish, and the remaining 48% consumed both. For consumers, this was equivalent to 5 or 6 seafood meals each month. The average age of seafood eaters was 37 years and that of non-eaters was 28 years. Data were not provided on the frequency of consumption by gender, but men did consume slightly greater portions of seafood than women.

A total of 17 714 meals of seafood were reported to have been eaten over the month. Salmon was the most popular seafood consumed, being eaten in 17% of the meals reported. Next, in descending order, were tuna (14%), shrimp (11%), cod (9%), and sole (6%). One of every six seafood meals consumed was salmon, with canned salmon accounting for two-thirds of these meals. Similarly, one in seven seafood meals were tuna, with canned tuna accounting for 95% of these meals. A high correlation between 'canned fish' and 'consumed as sandwich' was reported.

Any form of canned seafood constituted one third of all seafood meals consumed. Fresh and frozen seafood constituted 18% and 19%, respectively, of consumed seafood meals reported. Of the fresh seafood consumed, 14% consisted of salmon, 12% of shrimp, 11% of cod, and less than 0.5% of fresh tuna. The proportion of frozen raw seafood was comprised of 18% sole, 17% cod, 9% haddock, and 8% shrimp. Twelve percent of seafood meals were frozen and pre-cooked. Approximately 15% of seafood meals were prepared outside the home. In addition, 5% of seafood meals were 'self-caught' with the most frequently reported species being pike (68%), pickerel (46%), trout (38%), perch (25%), whitefish (16%) and bass (16%).

Survey respondents knew where the seafood originated in 68% of the meals reported as eaten. In total, 55% of the meals were of domestic origin: 28% from Atlantic Canada, 19% from Pacific Canada, 7% from a river or lake, and 1% from a fish farm. For 32% of the reported seafood meals, respondents did not know where the seafood had come from, though this figure may be influenced by the fact that 15% of seafood meals were prepared outside of the home.

The Market Facts (1991) study has been criticized for its low sample size, 3815 respondents. As well, the food diary approach has the potential to introduce biases to the data that are collected. Other potential drawbacks of the study include the fact that the survey was conducted in the winter when seafood availability is lowest, and seafood costs are highest. Also, the duration in which the study was carried out was short (1 month), and there was lack of focus on the intake patterns of sportfishers.

While these are valid concerns, Market Facts (1991) did capture the most pertinent information out of all of the other studies summarized in this document. Information on the consumption habits, portion sizes, frequency of consumption, and fish types that are popular among specific consumer groups were collected. Frequency of consumption data are particularly important for the calculation of mean daily intake figures for eaters. The sample size, though small, still manages to capture data pertaining to a fair number of individuals from the Canadian population - over three thousand people in ten provinces. The form of a written diary is the only feasible way in which to constantly record each aspect of fish consumption for an individual over an extended period of time. While a year-long study would be an ideal method in which to capture monthly or seasonal variation in fish consumption, this is an option that is not always feasible due to time and financial constraints. Furthermore, a month-long dietary survey is better at capturing variation in consumption habits than either a 24-hour recall study or a week-long study, both of which form the basis of all other Canadian intake studies discussed herein. Although the Market Facts (1991) report is now dated, it is still the most recent and comprehensive source of Canadian fish consumption information.

Due to the merits of the Market Facts (1991) study in relation to the other studies discussed above, all of the consumption-related figures generated herein are based solely or partially on data and information obtained from the Market Facts (1991) report.

RECOMMENDED FIGURES

Portion Size

In the Market Facts (1991) study, data were gathered on the average portion size consumed per eating occasion. Respondents were asked to approximate seafood serving sizes in 'cups' and these were then translated into metric grams according to conversion figures provided by Health and Welfare Canada. The average portion size of a finfish meal was 148 g, that of a shellfish meal was 105 g, and a general seafood portion was 137 g. However, 57% of seafood meals were 100 g or less and 73% of the meals were less than 150 g. In 11% of cases, servings were greater than 250 g (4% between 251 g and 300 g, 2% between 301 g and 350 g, 3% between 351 g and 400 g, and 2% over 400 g).

Portion sizes were positively correlated with respondent weight. Consumer body weights and the corresponding average seafood portion sizes from the Market Facts (1991) study are presented graphically in Figure 1.

Figure 1. Eaters-only Mean Seafood Portion Size Relative to Body Weight
Source: Data from Market Facts (1991)
Lines of best fit: Finfish: y = 0.85x + 93 (R² = 0.94); Shellfish : y = 0.70x + 57 (R² = 0.82); Seafood: y = 0.77x + 88 (R² = 0.97).

Average portion size can be determined for consumers of any body weight by interpolating values from linear regression lines-of-best fit. The portion sizes of finfish, shellfish, and seafood meals for consumer body weights that are commonly used by BCS in risk assessments were interpolated from Figure 1 and are presented below in Table 4.

Finfish Portion Size: Adults (20-70 Year-Olds)

Finfish portion sizes are available from various Canadian sources. Twenty-four hour recall data from the Nutrition Canada Survey (1973) reported eaters-only finfish consumption for adults ranging from 50 g to 164 g, with a mean of 98 g^{Footnote 15}. Similarly, the provincial Nova Scotia (Nova Scotia Department of Health, 1993) and Québec (Santé Québec, 1992) surveys report mean, eaters-only daily adult finfish consumption of 86 g to 122 g, with an eaters-only mean of 107 g for the entire population considered (18 to 74 year-olds)5. In the Market Facts (1991) study the average finfish portion size for adults weighing between 61 and 90 kg is 161 g; data interpolated from Figure 1 illustrate that this figure could be closer to 145 g. A Canadian Food Inspection Agency (CFIA) reference document indicates that for "marine and freshwater animals without sauce," the suggested serving size is between 60 g and 100 g of cooked fish (CFIA, 2003).

A 150 g portion size has been adopted by various international agencies. Food Standards Australia New Zealand (FSANZ) considers a finfish serving size to be equal to 150 g for adults (FSANZ, 2004) and in the UK, a portion of fish is represented by 140 g cooked fish (UK Food Standards Agency, 2004). In addition, 24-hour recall data from the US EPA (2002) reported the eaters-only mean intake of all types of seafood in people 14 years and older to range from 83 g/day to 112 g/day.

Taking the above information into consideration, 150 g would be considered a reasonable, albeit conservative, estimate of an average finfish meal size for an adult.

Finfish Portion Size: Toddlers (1-4 Year-Olds)

Interpolated data in Table 4 indicate that a young child aged 1 to 4 years and weighing 14 kg would consume, on average, approximately 106 g of finfish per serving. However, this interpolation may overestimate portion size for toddlers as there is only one data point representing the mean finfish portion size of all children weighing less than 30 kg. Data from Market Facts (1991) illustrates that the finfish portion size per eating day for 1 to 5 year-olds is 90 g, which may slightly overestimate the portion of finfish consumed by 1 to 4 year-olds. Twenty-four hour recall data from the Nutrition Canada Survey (1973) indicate that the mean, eaters-only finfish portion size of 1 to 4 year-olds ranges from 16 g to 124 g, with a mean intake of 49 g (see footnote^{Footnote 15}).

A 75 g portion size for children aged 6 years and under is employed by FSANZ (FSANZ, 2004).

It is recommended that the figure of 75 g be used to represent a typical finfish meal of a 1 to 4 year-old child.

Finfish Portion Size: 5-11 Year-Olds

For a child aged 5 to 11 years, interpolated data in Table 4 indicate that a typical finfish portion size is approximately 116 g. Market Facts (1991) reported finfish portion size per eating day to be 130 g for 6 to 12 year-olds. The Nutrition Canada Survey (1973) presented portion sizes for 5 to 11 year-old eaters ranging from 44 g to 261 g, with a mean of 96 g (see footnote^{Footnote 15}).

It is recommended that 125 g conservatively be applied as a typical finfish portion size for a 5 to 11 year-old child.

Frequency of Consumption

The Market Facts (1991) study investigated the frequency of seafood consumption among eaters. Fifty percent of seafood eaters consumed seafood, on average, once a week or less. Thirty-three percent of consumers ate seafood more than once per week but less than twice per week, for a total of 83% of the respondents eating seafood less than twice per week. Sixteen percent consumed seafood more frequently than twice per week and 12% of the respondents reported eating seafood once every three days. For the entire population surveyed, seafood was consumed an average of 5 times over a 1-month period, or 1 ¼ times per week.

Seafood consumption was also observed to increase with age. Those aged 55 years and older consumed an average of 7 seafood meals each month while those under 18 years of age ate an average of 4 seafood meals each month.

Seafood was consumed by eaters in the entire population (> 1 year) 1 ¼ meals/week; those over 55 years ate 1 ¾ seafood meals/week and those under 18 years consumed only 1 seafood meal/week.

Average Daily Consumption Estimation

Eaters-Only

The average daily intake of a food is the figure that is commonly employed in human health risk assessments. Market Facts (1991) used consumption frequency and portion size information to develop mean monthly intake figures for finfish, shellfish, and seafood. Table 5 reports mean and maximum monthly intake figures for the entire population that was surveyed (> 1 year).

Table 5. Daily Intake Figures for Different Fish Types (Entire Population (>1 year))

Source: Market Facts (1991)

The values listed in Table 5 for the entire population (> 1 year) closely approximate the mean monthly intake data for adults that are presented in Market Facts (1991). However, the figures in Table 5 overestimate consumption in children (aged 12 and under) since the proportion of children surveyed by Market Facts (1991) constituted a relatively small percentage (15%) of the total population surveyed. Market Facts (1991) specifically reports that the eaters-only mean daily finfish consumption rate for children aged 1 to 5 years is 10 g/day and that for children aged 6 to 12 years is 14 g/day; these age classes are slightly greater than those commonly employed by BCS in health risk assessments (1 to 4 and 5 to 11 year-olds) therefore may slightly overestimate the finfish consumption.

The average daily finfish intake figure recommended for 1 to 4 year-olds is 10 g/day and for 5 to 11 year-olds is 14 g/day; the recommended average daily finfish intake figure for adults is 22 g/day.

All-Persons

All-persons intakes can also be calculated using information provided in the Market Facts (1991) study with the assumption that there is an arithmetic distribution between eaters. The following figures are used to calculate the all-persons mean daily finfish intake for the entire survey population (> 1 year old):

• 81% of the respondents eat seafood
• 97% of these respondents consume finfish
• 22 g/day is the eaters-only mean finfish intake
• All-persons mean finfish intake (> 1 year)
• = 0.81*0.97*22 g/day
• = 17 g/day

This figure, 17 g/day, is higher than the mean of 11 g/day from the Nova Scotia Provincial survey (1993) but is similar to those reported in the Canada Food Statistics survey, 16 g/day (Statistics Canada, 2002) and the Family Food Expenditure Survey, 15 g/day (Sabry, 1990). However, 17 g/day is much higher than the 1.0 g/day to 2.2 g/day finfish means listed in the Nutrition Canada Survey (1973), which, due to its 24-hour recall format, does not accurately capture true consumption trends.

All-persons shellfish consumption for the Canadian population can be calculated in the same manner as above, though it is thought that such a figure has a greater probability of being skewed as there were fewer shellfish consumers in the population surveyed by Market Facts (1991).

• 81% of the respondents consume seafood
• 51% of these respondents consume shellfish
• 9 g/day is the eaters-only mean shellfish intake
• All-persons mean shellfish intake (> 1 year)
• = 0.81*0.51*9 g/day
• = 3.7 g/day

This figure of 3.7 g/day is relatively close to the all-persons shellfish consumption figure of 2.9 g/day from the Family Food Expenditure Survey (Sabry, 1990) and is similar to the all-persons Canada Food Statistics consumption estimate of 3.8 g/day (Statistics Canada, 2002). This figure is lower than the provincial all-persons means of 4.3 g/day in Québec and 14 g/day in Nova Scotia (Brulé, 1996; Santé Québec, 1992; Nova Scotia Department of Health, 1993), which could be attributed to the fact that only adults were surveyed in the Market Facts (1991) study.

The all-persons population mean for all types of seafood can be calculated using the following information from Market Facts (1991):

• 81% of the respondents consume seafood
• 26 g/day is the eaters-only mean shellfish intake
• All-persons average seafood intake (> 1 year)
• = 0.81*26 g/day
• = 21 g/day

The figure, 21 g/day, is consistent with a number of studies, including the Canada Food Statistics estimate of 20 g/day (Statistics Canada, 2002) and the Family Food Expenditure Survey figure of 18 g/day (Sabry, 1990). This figure also falls within the range of the means from the Provincial surveys of 15 g/day in Québec and 37 g/day in Nova Scotia (Brulé, 1996; Santé Québec, 1992; Nova Scotia Department of Health, 1993). Also, this figure is appreciably higher than the 0.6 g/day to 2.2 g/day estimated in the Nutrition Canada Survey (1973), but this is due to the nature of the 24-hour recall survey that allowed a small percentage of actual eaters to be captured since seafood is infrequently consumed.

It is recommended that the following mean, all-persons intake figures for the entire population (> 1 year) be adopted: 17 g/day for finfish, 3.7 g/day for shellfish, and 21 g/day for seafood.

Distinct Population Subgoups

It is useful to have standard finfish, shellfish, and seafood consumption values to apply to the Canadian population. However, it should be recognized that there are unique subgroups of the population that may consume more or less seafood than that of the average Canadian. Such population sub-groups are discussed briefly below.

Coastal Provinces

A number of studies illustrate that the frequency and magnitude of fish consumption varies throughout Canada. Sabry (1990) noted that those living in the coastal areas consumed more fish than those in the Prairie Provinces. Hunter et al. (1988) observed from the Nutrition Canada Survey (1973) that those living in Atlantic Canada had nearly twice the probability of reporting fish consumption the previous day than those living in the Prairie Provinces. The majority of all-persons and eaters-only 24-hour recall consumption rates reported by Nova Scotians were higher by as much as 2-fold than those reported by Québécois. Furthermore, between 1.6 and 2.3 more survey respondents reported eating seafood in Nova Scotia relative to Québec (Nova Scotia Department of Health, 1993; Santé Québec, 1992). Market Facts (1991) identified general geographic areas of high fish consumption; the Atlantic Provinces constituted only 8% of the total population sampled but possessed 19% of the heaviest consumers. The population surveyed from Québec was comprised of 26% of the survey population and included 29% of the heaviest consumers.

The Market Facts (1991) study noted that fish consumers from the Maritime Provinces consumed significant amounts of cod. Additionally, Hunter et al. (1988) reported that Atlantic cod comprised 90% of the overall fish consumption of a small group of people living in coastal Newfoundland. In light of the cod fishery closures of the past decade, it is conceivable that Atlantic Canadians now consume other varieties of fish more frequently.

Adult Men

Demographic groups that consume greater-than-average amounts of seafood are generally older population sub-groups and males. In the Market Facts (1991) study, men over the age of 25 years consumed a greater quantity of seafood than women within the same age group. The greatest number of heavy consumers were men over 55 years of age. Since men weigh more than women, on average, they may consume larger portion sizes relative to their energy requirements and body weight. Moreover, it is conceivable that older men are advised by their physicians to eat greater amounts of fish in order to maintain their cardiovascular health. Sabry (1990) also suggested that older people may be more likely to follow traditional patterns of high fish consumption such as those of coastal fishers, First Nations groups, and certain minority groups.

Children

The Nutrition Canada Survey (1973) illustrated lower rates of fish consumption in children; such a trend has also been noted by a number of other sources, including Sabry (1990). This trend can be explained by the lower energy needs and smaller portion sizes consumed by children and the fact that a smaller proportion of young people consume fish.

The Market Facts (1991) study indicated that 75% of children 12 and under consumed seafood and 91% of consumers over 55 years consume seafood. Among children seafood consumers, the average portion size per eating day was estimated to be between 87 g (1 to 5 year-olds) and 121 g (6 to 12 year-olds). However, when frequency of consumption was factored in, the mean monthly consumption rate for children seafood eaters 12 years and younger was between 10 g/day and 14 g/day of finfish, 4 g/day to 5 g/day of shellfish, and 10 g/day to 15 g/day for seafood. The US EPA (2002) reports that those 14 years and under most commonly consume tuna (1.5 g/day)^{Footnote 4}, cod (0.66 g/day), shrimp (0.61 g/day), and marine salmon (0.28 g/day).

Pregnant Women and Women of Childbearing Age

The intake figures applied to pregnant women and women of childbearing age may vary from that of the general population. For example, intake of some large, predatory fish species may be lower in this sub-group due to the consumption advice that Health Canada has issued due to elevated mercury levels in these types of fish.

Subsistence and Recreational Fishers

In order to determine an intake figure for heavy adult seafood consumers, a number of studies that include intake data in heavy consumers, ranging from recreational fishers, subsistence fishers, First Nations peoples, and certain ethnic groups, were examined. These studies are outlined below as well as in Appendix A (Tables A2 and A3). As numerous studies on this topic exist, it was impossible to evaluate them all.

1. As part of a cardiovascular study in Montréal (Godin et al. 2003), 112 St. Lawrence River anglers were surveyed about their seasonal consumption of finfish. All of the respondents were men. Finfish consumption was reported to be between 55 and 77 sportfish meals/year, in addition to 35 meals of commercial fish, for a total of 90 to 112 fish meals/year. The average weekly consumption of sportfish was estimated to be 1.5 meals - the definition of a meal was not specified although the use of fillet models and thickness indicators were mentioned. If it is assumed that a finfish portion size is 150 g, the average daily intake of sportfish would equal 27 g/day and that of commercial fish would be 14 g/day, for a total finfish intake of 41 g/day. However, the authors suggested that reported intakes were overestimated due to memory lapses, current seasonal intakes affecting previous seasonal intakes, and the respondents tendency to inflate fish size.
2. A methylmercury exposure study of 94 recreational anglers on James Bay area estimated average finfish consumption to be 87 g/day (Loranger et al. 2002). However, reported consumption figures may have been inflated due to past recall biases, unrealistically high reporting of meal frequencies and portion sizes, as well as a tendency to exaggerate fish size. In addition, study participants were specifically chosen from a pool of 300 anglers who were known to regularly eat their catch.
3. An Ontario Fish and Wildlife Consumption Survey (Great Lakes Health Effects Program, 1997) was carried out over a period of 3 years in order to identify fish eaters and their consumption frequency as well as the species consumed from five Areas of Concern in the Great Lakes. The subpopulation interviewed included over 4500 respondents. Fifty-two percent of the respondents consumed less than 11 finfish meals/year, 22% consumed between 12 and 25 meals/year, and 21% consumed between 26 and 95 meals/year. Approximately 6% of those surveyed ate more than 97 meals/year. The definition of a meal was not defined. Since this was a draft document and because a number of questions were raised by BCS concerning the study methodology, it is recommended that this survey be given little weight in developing intake recommendations.
4. Results from a Lake Ontario Salmon/Trout Angler Intake survey conducted by the Ontario Ministry of the Environment (OMOE) Sportfish Contaminant Monitoring Program (1988) included data on finfish portion size and the number of finfish consumed each year. These data were also compared to the OMOE Guide to Eating Ontario Sportfish (1986), as outlined in Table 6 below.

1. Twenty immigrant Vietnamese women residing near Hamilton, Ontario participated in a questionnaire pertaining to their consumption of finfish from the Great Lakes. Their yearly consumption of freshwater fish was determined to be between 34 meal/year and 58 meals/year (Cavan et al. 1996). A marked seasonal effect demonstrated that consumption in the spring and summer was much higher than in the fall or winter. Assuming that the average number of meals per year is 46 and a typical finfish portion size is 150 g, the average daily fish intake would be 19 g/day.
2. Kostasky et al. (1999) reported that high-level sportfishers in Montreal eat 0.92 sportfish meals/week, approximating 18 kg/year or 50 g/day. Low-level sportfishers ate 0.38 meals/week or 3.3 kg/year which is equal to 9.0 g/day. None of the respondents reported eating fish more than twice each week.
3. Using two 24-hour diet recalls, Shatenstein et al. (1999) determined fish consumption among 18 Asian-origin sportfishers on the St. Lawrence River. Nine Bangladeshi fishers consumed an average of 47 sportfish meals/year while nine Vietnamese sportfishers consumed 41 meals/year. The definition of meal size was not given. If it is assumed that a typical finfish meal size is 150 g, the average daily finfish intake is 19 g/day for Bangladeshi fishers and 17 g/day for Vietnamese fishers.
4. Dewailly et al. (2003) summarized data from the Santé Québec (1992) 24-hour recall survey and estimated that southern Québécois consumed 13 g/day of fish and marine products, James Bay Cree consumed 60 g/day, and Nunavik Inuit consumed 131 g/day.
5. A study of Ontario Amerindians estimated fish consumption rates found a geometric intake mean of 16 g/day, with males consuming 19 g/day and women consuming 14 g/day (Richardson and Currie, 1993). Fish consumption rates were found to increase with age and vary with season, with the highest consumption reported in the summer and the lowest in the winter.
6. Chan et al. (1999) determined that the average daily intake of locally caught freshwater fish for all fishers in a Mohawk First Nations community was 23 g/day and the eaters-only daily intake was 33 g/day. Fish consumption was highest in the spring and summer, with a reported frequency of usually less than once per week.
7. Indian and Northern Affairs (2003) reported fish consumption among 207 women residing in two Inuit communities, Repulse Bay and Pond Inlet. Data were obtained using both a 24-hour recall as well as a week-long food frequency questionnaire. Only the results of the 1997 weekly diary are reported here. In Repulse Bay, weekly consumption of fish was equal to 304 g/person, consisting of 32 g of canned salmon, 18 g of frozen fish sticks, and 3 g of canned sardines. Daily intake can thus be calculated to be 43 g/day for women from Repulse Bay. In Pond Inlet, the average weekly consumption of fish was 259 g, consisting of 13 g frozen fish sticks, 10 g canned pink salmon, and 10 g canned sardines. Thus, the weekly intake was approximately 37 g/day for female residents of Pond Inlet.

Of the studies examined, it was observed that the average intake of recreational or subsistence fishers in Canada ranges widely, from 9.0 g/day (Kostasky et al.1999) up to 87 g/day (Loranger et al. 2002). In the reports on First Nations and Inuit consumption habits, the mean intake ranged from 14 g/day (Richardson and Currie, 1993) to 131 g/day (Dewailly et al. 2003). These studies were all conducted in different manners for diverse purposes therefore it is difficult to directly compare intake values generated from each. Nevertheless, the average daily adult consumption rate was calculated from the numerous different intake figures presented herein, yielding a value of 38 g/day.

Market Facts (1991) also provided data on high-level consumers. The figures reported for consumers in the 75^th percentile for consumption were 28 g/day (all-persons) and 33 g/day (eaters-only) of seafood. The upper 90^th percentile seafood consumption rates were 45 g/day (all-persons) and 49 g/day (eaters-only).

Based on the information above, it is recommended that the current figure of 40 g/day continue to be used to represent heavy adult seafood consumers.

It is not expected that children will consume the same amount of fish as heavy adult consumers, hence intake figures should be adjusted accordingly for this cohort. In the absence of representative data, it was assumed that a child of a recreational or subsistence fisher will have the same consumption rate (40 g/day) as their parent(s) but their portion size would be smaller. Using the portion sizes presented earlier in this document, the average daily intakes for heavy child seafood consumers can be calculated. For a 5 to 11 year-old, a typical serving of commercial fish (125 g) is approximately 83% of the size of a typical adult portion (150 g). Thus finfish intake for a 5 to 11 year-old child that is a heavy seafood consumer is 33 g/day (40*0.83). The typical portion of commercial fish (75 g) for a 1 to 4 year-old child is half the size of a typical adult portion (150 g), therefore the mean intake for heavy consumers within this age group is approximately 20 g/day (40* 0.5).

Based on the information above, it is recommended that the intake figure of 33 g/day be applied to 5 to 11 year-olds and the figure of 20 g/day be applied to 1 to 4 year-olds that are part of the subsistence or recreational fishing culture.

References

Brulé, D. 1996. Canadian food consumption surveys: A federal-provincial partnership. Report presented at 21st National Nutrient Databank Conference, Baton Rouge, Louisiana June 20-22, 1996. (Accessed April 21, 2004)

Canadian Food Inspection Agency (CFIA). 2003. Reference Amounts [Schedule M] and Serving Sizes: Table 6-3. 2003 Guide to Food Labelling and Advertising. http://www.inspection.gc.ca/english/fssa/labeti/guide/ch6e.shtml#tab6-3 (Accessed April 21, 2004)

Cavan, K.R., Gibson, B.C., Cole, B.C., and Riedel, D. 1996. Fish consumption by Vietnamese women immigrants: A comparison of methods. Archives of Environmental Health 51: 452-457.

Chan, H.M., Trifonopoulos, M., Ing, A., Receveur, O., and Johnson, E. 1999. Consumption of freshwater fish in Kahnawake: Risks and benefits. Environmental Research 80:S213-S222.

Dewailly, E., Blanchet, C., Gingras, S., Lemieux, S., and Holub, B.J. 2003. Fish consumption and blood lipids in three ethnic groups of Québec (Canada). Lipids 38: 359-365.

Food Standards Australia New Zealand (FSANZ). 2004. Mercury in Fish: Further information. (Accessed April 21, 2004)

Godin, C. Shatenstein, B., Paradis, G., and Kostasky, T. 2003. Absence of cardiovascular benefits and sportfish consumption among St. Lawrence River anglers. Environmental Research 93: 241-247.

Great Lakes Health Effects Program. 1997. Draft: Sport fish and wildlife consumption study in Areas of Concern.

Hunter, D.J., Kazda, I.k Chockalingam, A., and Fodor, J.G. 1988. Fish consumption and cardiovascular mortality in Canada: An inter-regional comparison. American Journal of Preventive Medicine 4: 5-10.

Indian and Northern Affairs Canada. 2003. Change in Nutrition and Food Security in Two Inuit Communities. Table 24. Average weekly consumption (grams) of selected country foods: Food frequency questionnaire. (Accessed April 21, 2004)

Kostasky, T., Przybysz, R. Shatenstein,B., Weber, J.P., and Armstrong, B. 1999. Fish consumption and contaminant exposure among Montreal-area sport fishers: Pilot study. Environmental Research 80: S150-158.

Loranger, S., Schetagne, R., Plante, M., Carrier, G., Sauvé, S., Emard, B., Piché, L., Babo, S. 2002. Evaluation of a questionnaire-based method for the estimation of methylmercury exposure of recreational anglers in the James Bay Territory (Québec, Canada). Human and Ecological Risk Assessment 8: 559-571.

Market Facts of Canada. March 1991. Research Report: National Seafood Consumption Study. Conducted for Health and Welfare Canada. #C388/JdeB.

Nova Scotia Department of Health. 1993. Nova Scotia Nutrition Survey Report. Canadian Heart Health Initiative Study.

Nutrition Canada. 1973. National Survey: A report by Nutrition Canada to the Department of National Health and Welfare.

Ontario Ministry of the Environment (OMOE). 1988. Sportfish Contaminant Monitoring Program. Survey regarding Guide to Eating Ontario Sport Fish. A comparison of Sport Fish Consumption between Lake Ontario Salmon and Trout Anglers and 1986 Guide Questionnaire Respondents.

Ontario Ministry of the Environment (OMOE). 1986. Guide to Eating Ontario Sport Fish.

Richardson, G.M. and Currie, D.J. 1993. Estimating fish consumption rates for Ontario Amerindians. Journal of Exposure Analysis and Environmental Epidemiology 3: 23-38.

Sabry, J.H. 1990. Nutritional Aspects of Fish Consumption: A report prepared for the National Institute of Nutrition on behalf of the Economic and Commercial Analysis Directorate of the Department of Fisheries and Oceans.

Santé Québec. 1992. Québec Nutrition Survey Report. Canadian Heart Health Initiative Study.

Shatenstein, B., Kostasky, T., Tapia, M., Nadon, S., and Leclerc, B.S. 1999. Exploratory assessment of fish consumption among Asian-origin sport fishers on the St. Lawrence River in the Montreal Region. Environmental Research 80: S57-70.

Statistics Canada. 2002. Canada Food Statistics: Food Consumption in Canada. (Accessed April 21, 2004)

United Kingdom Food Standards Agency. March 2004. Tuna advice updated. (Accessed April 21, 2004)

United States Environmental Protection Agency (US EPA). 1999. Asian and Pacific Islander Seafood Consumption Study (EPA 910/R-99-003). (Accessed July 21, 2004)

United States Environmental Protection Agency. 2002. Estimated per capita fish consumption in the United States (EPA-821-C-02-003). (Accessed July 21, 2004)

Appendix A. Summary of Studies