Annex 2: Case Study of Methylmercury in Fish
In the decades since Minamata, several epidemiological studies of populations with a diet either high in fish or in fish and marine mammals have provided evidence that typical levels of methylmercury in some types of fish, not unusually high levels associated with pollution, pose some health hazards, again with a focus on the developing brain.24 There is some evidence that methylmercury exposure from a diet rich in fish and seafood may adversely affect cognitive function in adults.25 Nevertheless, damage associated with prenatal exposure is considered the most sensitive effect and is the central concern of risk management. Evidence that these potential health risks may be associated with “normal” levels of fish consumption has led to both national and international efforts to assess the risks from methylmercury in fish, and to establish guidelines for safe maximum exposure.
Methylmercury risks may be a concern for any national or subnational population that consumes large amounts of fish. Different fish species tend to accumulate methylmercury to different degrees, and the degree of exposure to methylmercury will vary depending on which fish species are important in a population’s diet, and how much methylmercury is present in the specific fish species consumed locally. Risk assessments, in particular the exposure assessment part, must therefore be population-specific. If excessive methylmercury exposure is found, risk management can be challenging. Fish consumption has many nutritional benefits, and fish is the main source of dietary protein for some populations. Reducing fish consumption to avoid methylmercury exposure might therefore damage public health in the broader sense. Risk communication, in particular educating consumers so that they can choose low-mercury fish species, is an important risk management tool for managing methylmercury risks.
This case study briefly reviews two examples of risk analyses for methylmercury in fish.
- The United States Environmental Protection Agency (EPA) has established a Reference Dose (RfD), which is a safe upper intake limit, similar to a Tolerable Daily Intake. The United States has also established an Action Level, which is a guideline for a maximum acceptable mercury level in fish, and has issued fish consumption advice.
- The Joint FAO/WHO Expert Committee on Food Additives (JECFA) has established a safe upper intake limit, called a Provisional Tolerable Weekly Intake (PTWI), based on a scientific review and risk assessment, and the Codex Alimentarius Commission (CAC) has established Guideline Levels for Methylmercury in Fish (CAC/GL 7-1991).
The population group most often considered at risk from methylmercury exposure are women of childbearing age because damage to the developing foetal brain is currently considered to be the health risk of greatest concern, i.e. the most sensitive endpoint. However, methylmercury has other toxic effects (e.g. it affects the nervous system in adults).26
Therefore, concern is not strictly limited to potential effects on the foetal brain; people who eat a great deal of fish may also be at some risk for as yet sparsely documented effects. In some countries, only a small subset of the total population consumes enough fish to warrant any health concerns, while in other countries, where fish is the primary source of dietary protein, “high-end” consumers may include much of the general population.
The risk profile developed by the EPA focuses on women who are or may become pregnant, and on a handful of particular fish species that accumulate fairly high levels of methylmercury. The JECFA/Codex approach recognizes that methylmercury in fish may be a public health concern for many member countries, and also that a specific risk profile needs to be developed for each individual country contemplating action, since fish consumption patterns and thus the associated risk vary from country to country. These risk profiles were developed primarily by risk assessors (JECFA for FAO/WHO and Codex; government scientists in the USA), who were working and communicating with the risk managers in each case.
risk assessment to help them choose among risk management options so much as they needed an updated and more precise definition of a “safe” level of exposure to methylmercury to support their determinations of the appropriate level(s) of protection for exposed populations.
At the international level, JECFA has reviewed methylmercury on several occasions during the period from 1972 to 2006. At its 2000 session, and at the request of the CAC, JECFA noted that evidence was accruing from two major ongoing epidemiological studies, and agreed that an additional review be conducted, specifically to advise on whether the existing PTWI should be revised in light of recent evidence, when additional data became available. That review occurred at the 61st JECFA meeting, in 2003. Thus, in the United States, the need for a risk assessment was driven primarily by risk managers planning a policy action, while internationally, risk assessors, monitoring emerging scientific evidence, determined that the time had come to update the risk assessment, knowing that risk managers were prepared to review the related risk management decisions.
At the international level, JECFA communicates closely with CCFAC, the risk managers who apply the PTWI in managing risks of methylmercury in fish. Since CCFAC and JECFA each meet once a year at different times and in different countries, communication between them mostly occurs through the JECFA Secretariat. Subsequent to the 2003 JECFA review, CCFAC posed some specific questions to JECFA, which were taken up at the JECFA session in 2006. The discussion at CCFAC is continuing and further interaction with JECFA may occur as the process moves forward.
A key step in commissioning a risk assessment is to assemble the risk assessment team. Finding qualified experts who are knowledgeable about the specific problem but are not committed to a predetermined point of view can be a challenging task for risk managers. The EPA put together a group of scientists drawn from its health effects research staff. The NAS/NRC assembled a group of experts from the national scientific community, following procedures (described on the NAS web site)27 to ensure appropriate expertise, to balance viewpoints and to exclude those with possible biases or conflicts of interest. Internationally, the JECFA Secretariat assembled an expert group from FAO and WHO rosters of experts, drawn from the worldwide scientific community, in accordance with FAO/WHO procedures to balance expertise and screen out potential conflicts of interest.28
safety problems that all need to be managed, and have limited resources. However, enough knowledge already exists to establish that methylmercury is a serious public health concern, and it has been a priority for risk management for many years. The risk ranking step therefore was not necessary either in the United States or internationally in this case.
Unlike the examples presented in Chapters 2 and 3, which describe how changes in risk associated with given increases or decreases in exposure are quantified and used to determine an Appropriate Level of Protection, the risk assessors in these methylmercury cases used a somewhat different approach. In each case, the (limited) available dose response data were used to calculate a Benchmark Dose Lower Confidence Limit (BMDL) or to estimate a No-Observed Effect Level (NOEL). Uncertainty factors were then applied to estimate the nominally “safe” dose (RfD by the EPA, PTWI by JECFA).
The EPA and NAS/NRC each concluded, after reviewing the new epidemiological evidence, that a long-term study in the Faeroe Islands, testing for methylmercury effects in children born to women with a diet rich in fish and whale meat,30 provided the best available evidence on potential adverse health effects. The Faeroe Islands study has associated prenatal methylmercury exposure with observed effects on brain nerve signal transmission and on several indices of cognitive development. Neither of the risk assessments in the United States relied on a similar study of a population with a high-fish diet in the Seychelles Islands,31 which has examined children for effects comparable to those studied in the Faeroe Islands, but has to date not identified statistically significant adverse effects, and thus was not deemed suitable for the risk assessment EPA wished to perform. JECFA, on the other hand, relied on both studies to derive an average dose from a BMDL (Faeroe Islands) and the NOEL (Seychelles).
The EPA next estimated a variety of BMDLs using several models and associations between methylmercury doses and neurological developmental outcomes from the Faeroe Islands study. One BMDL was then selected and a ten-fold default uncertainty factor was applied to account for the variability in individual sensitivity, and a RfD of 0.1 µg/kg of body weight (µg/kg-bw) per day, or 0.7 µg/kg-bw per week was established which corresponds to a blood mercury level of 5.8 µg/litre.32 JECFA, relying on the same evidence, used a slightly different approach. The committee calculated a steady-state intake of methylmercury of 1.5 µg/kg-bw per day from a maternal hair mercury level of 14 mg/kg, which is the average dose from the two studies. It was the lower confidence limit of the benchmark dose from the Faeroe Islands study, and the calculated NOEL from the Seychelles study. JECFA then applied a data-derived, 6.4-fold uncertainty factor to calculate a PTWI for exposure of pregnant women of 1.6 µg/kg-bw per week.33 This value is slightly lower than the previous JECFA PTWI of 3.3 µg/kg-bw per week, which was derived based on the lowest effect levels noted in earlier studies of populations exposed to methylmercury contamination via food.
The recommendations reached by experts in the USA and JECFA cases described here differed by approximately a factor of two. However, in view of the uncertainties in the scientific evidence and the different approaches taken by the two groups of risk assessors who made those determinations, these recommendations are actually quite close.
JECFA assembled data from five national exposure studies, and calculated possible methylmercury intake associated with the five WHO GEMS/Food-regional diets, using estimated average fish intake and data on the average mercury content of fish submitted by various member governments. JECFA estimated that high-end fish consumers in most of the countries for which it had data were exposed to methylmercury doses greater than the PTWI. The highest estimate for the average methylmercury dose from the five GEMS/Food-regional diets (JECFA did not say which regional diet was highest) was 1.5 µg/kg-bw per week, just below the new PTWI of 1.6 µg/kg-bw per week, indicating that almost half the people with that diet would exceed the tolerable level of methylmercury intake.38
JECFA did not characterize the risk for particular regions or countries, but clearly suggested that exposure to methylmercury doses above the PTWI is relatively commonplace in countries where fish is important in the diet, and that national governments may now need to carry out population-specific exposure assessments.
Risk characterizations of the type developed for methylmercury are relatively imprecise; risk is not quantitatively characterized in terms of the probability and severity of adverse health effects relative to defined levels of exposure, but rather, presumptively “safe” exposure levels are estimated (see Chapter 3, Risk assessment, for discussion). Such “safety assessments” can nonetheless provide a basis for risk management decisions.
risk communication and in the USA, extensive communication took place among the interested government agencies, the scientific community, and a variety of stakeholders, ranging from fishing industry interests to NGOs concerned about methylmercury hazards in foods.
As attention returned to risk management aspects, the process in the United States was open to participation by stakeholders.40 Some of those stakeholders have communicated aggressively, both with the government and with the public at large. For example, fishing interests, especially the United States tuna industry, have criticised the EPA risk assessment and RfD as excessively precautionary, denied that methylmercury in fish poses risks to public health, and spent millions of dollars on public relations and advertising campaigns to persuade people to ignore methylmercury risks and eat more fish.41 Public health, environmental and consumer organizations have concluded, in contrast, that methylmercury risks are a significant public health concern, and sought in their own ways to inform the public and persuade policy-makers of their view.42 There has been so much risk communication on the methylmercury problem in the United States that an intense public controversy exists.
Communication about the JECFA risk assessment has been somewhat less intense. When CCFAC received the JECFA recommendation for a lowered PTWI, the committee initiated a review of the Codex guidelines for methylmercury in fish. Some CCFAC members had questions, seeking clarification of JECFA’s reasoning on certain points.43 In particular, some members were uncertain whether JECFA intended that the new, lower PTWI should be applied to everyone in the general population, or whether it applied only to women who were or might become pregnant. JECFA considered this request in 2006 and clarified that the previous PTWI of 3.3 µg/kg-bw had, in fact, been withdrawn in 2003. JECFA confirmed the existing PTWI of 1.6 µg/kg-bw, set in 2003, based on the most sensitive toxicological end-point (developmental neurotoxicity) in the most susceptible species (humans). However, the Committee noted that life-stages other than the embryo and foetus may be less sensitive to the adverse effects of methylmercury. In the case of adults (with the exception of women of childbearing age for protection of the developing foetus), JECFA considered that intakes of up to about two times higher than the existing PTWI of 1.6 µg/kg-bw would not pose any risk of neurotoxicity. For infants and children JECFA could not identify a level of intake higher than the existing PTWI that would not pose a risk of developmental neurotoxicity for infants and children, hence for this age group the new PTWI applies.
The CCFAC, based on the new JECFA PTWI, is now considering further appropriate actions it might pursue. At its 2004 meeting, CCFAC asked a drafting group to prepare a discussion paper, outlining possible risk management options that national governments might consider. The paper,44 prepared with the leadership of the European Commission, focused on both the Codex Guideline Levels for Methylmercury in Fish, and on providing information to stakeholders, especially consumers, as a risk management option. It was discussed at the 2005 CCFAC session,45 which agreed to organize a workshop on risk communication as a risk management tool. This workshop was held in conjunction with the CCFAC session in April 2006.
WHO is also currently drafting a document to provide advice to member governments on how to conduct risk analysis for methylmercury in fish. International advice on this subject will be drawn from national experiences. The rest of this section, therefore, examines the national aspect of this case study, the experience in the United States.
Among actions that can be taken by national food safety authorities, the following are some risk management options that could be considered:
- The sale of certain fish species that are very high in methylmercury could be banned.
- A maximum contaminant level could be set for mercury or methylmercury in fish, and used to restrict sale and consumption of fish that exceed the established limit.
- The fishing industry and fish processors and retailers could be required to implement a code of Good Hygienic Practice or a HACCP system designed to prevent fish with potentially harmful levels of methylmercury from reaching consumers.
- Consumers can be educated and informed about methylmercury levels in fish and the associated risks, so that they can manage their own methylmercury exposure.
The United States adopted an “Action Level,” a guideline value for the acceptable upper limit of methylmercury concentrations in fish, in 1969. Originally set at 0.5 parts per million (ppm), the Action Level was raised to 1.0 ppm in 1979, after the fishing industry successfully sued the FDA. The court ruled that FDA’s exposure assessment and resulting safety assessment which it used as the justification for the 0.5 ppm level were unnecessarily conservative and inappropriate. Many other national governments, and CCFAC, have issued similar guidelines, generally set either at 0.5 or 1.0 ppm.46
In the United States, the Action Level is rarely if ever enforced; FDA concedes, for instance, that a significant portion of swordfish sold in national markets contains more than 1.0 ppm of mercury. While such a limit can, in theory, be used to prevent sale of fish that exceed it, in practice the United States Action Level has proved difficult and costly to enforce, and if strictly enforced, it could have negative socioeconomic effects similar to those discussed for a ban, above. Also, since the level of mercury in fish is just one of several factors that determine risk, efforts to keep high-mercury fish off the market cannot, by themselves, effectively reduce exposure and the associated risk. Someone who ate a great deal of fish with, for example, 0.25 ppm mercury could exceed the safe intake limit by a wide margin, while someone else who ate swordfish once or twice a year, for instance, might not be particularly at risk. Since the Action Level cannot be adjusted to take into account other factors that determine risk, enforcing it has not been a high priority. In sum, while it is seen as a useful guideline, the United States Action Level for mercury in fish has not significantly reduced exposure.
GHP or HACCP approaches that could help fish and seafood industries reduce the amount of methylmercury in products they sell appear to have significant potential for mitigating the problem, but this approach has not been pursued to date in the USA.
A few other private-sector initiatives have had modest effects. Some retail grocery chains are working with state governments and NGOs in the United States to provide information on the mercury content of different fish at the point of sale (e.g. at supermarket fish counters). Other sellers of fish, including chefs at famous restaurants, have promised to stop offering certain high-mercury species.
Information-based options have been the recent focus of risk management for methylmercury in the USA. Because the risk depends on multiple factors (including who is consuming the fish, which fish they choose to consume, how much of each fish species they eat, and how much methylmercury the fish in question contain) education and risk communication have attracted great interest as risk management options. These approaches can address the complexity of the problem, do not require costly and impractical enforcement efforts, can be implemented relatively quickly and at relatively minimal cost, and hold at least the potential for reducing methylmercury exposure substantially, without adverse nutritional or economic consequences.
Several State Health Departments within the United States have also issued consumer advice on methylmercury in fish, as have some professional organizations and numerous NGOs. American consumers have no shortage of advice and “educational” information on this topic; in fact, one concern has been that differences in the advice from different sources may be confusing consumers. The 2004 joint FDA/EPA “advisory” was in part undertaken as an effort to get the federal government, at least, to speak with a single voice on this subject.
Since implementation is a responsibility of national authorities, there is no section on this phase of risk management in the JECFA/Codex risk analysis for methylmercury.
The “advisory” option being pursued now in the USA was implemented in 2004, and there has not been enough time to determine most of its expected effects. For example, a key indicator of effectiveness of the EPA/FDA “advisory” will be whether national surveys show that a decreasing percentage of women have blood mercury levels above the EPA reference level, but such data are not expected to be available for several years.
Nevertheless, some efforts to assess the effects of the informational approach in the USA are now under way. Before it issued the 2004 advisory, the government conducted sessions with consumers (“focus groups”) to assess how they would understand and respond to both the information and the advice provided. Since the advisory was issued, a concern has arisen that warnings about contaminants like methylmercury in fish may make consumers afraid to eat fish, and cause them to lose important nutritional benefits associated with fish in the diet. Whether this is true or not is far from clear at this point,48 but the question has attracted a great deal of attention from academic researchers, state and federal governments, and interested stakeholders. Investigations now under way may lead to fine-tuning the advice offered to consumers, so that they can continue to consume low-mercury fish for their nutritional benefits, while minimizing their mercury exposure.
Huddle, N., M. Reich and N. Stisman. 1987. Island of Dreams: Environmental Crisis in Japan. Rochester, VT: Schenkman Books, Inc.; 2nd Edition.
Grandjean, P., et al. 1997. Cognitive deficit in 7-year-old children with prenatal exposure to methyl mercury. Neurotoxicol Teratol 19:417-428; National Research Council. 2000. Toxicological Effects of Methylmercury. Washington, DC: National Academy Press.
For example, Yokoo, E.M., et al. 2003. Low level methylmercury exposure affects neuropsychological function in adults. Environmental Health: A Global Access Science Source 2:8. Also, Newland, C.M. and E.B. Rasmussen. 2003. Behavior in Adulthood and During Aging Is Affected by Contaminant Exposure in Utero. Current Directions in Psychological Science 12(6):212-217.
For a review of the relevant literature, see National Research Council. 2000 (footnote 24 above). Also see JECFA’s 2003 assessment, WHO Food Additives Series, 52, Safety evaluation of certain food additives and contaminants, Prepared by the 61st meeting of the Joint FAO/WHO Expert Committee on Food Additives. International Programme on Chemical Safety, World Health Organization, Geneva, 2004. Pages 565-623, Methylmercury.
Further information about FAO/WHO rosters of experts is available in the FAO/WHO Framework for the Provision of Scientific Advice on Food Safety and Nutrition (to Codex and member countries) (at: www.fao.org/ag/agn/proscad/index_en.stm) as well as on the JECFA web site (at www.fao.org/ag/agn/jecfa/experts_en.stm and www.who.int/ipcs/food/jecfa/experts/en/index.html).
For a description of the EPA risk assessment, see Rice, D.C., R. Schoeny and K. Mahaffey. 2003. Methods and rationale for derivation of a reference dose for methymercury by the US EPA. Risk Analysis 23(1):107-115. For a description of the NAS/NRC risk assessment, see National Research Council. 2000 (footnote 24 above). For a description of the JECFA risk assessment, see WHO Food Additives Series, 52, Safety evaluation of certain food additives and contaminants, Prepared by the 61st session of the Joint FAO/WHO Expert Committee on Food Additives. International Programme on Chemical Safety, World Health Organization, Geneva, 2004. Pages 565-623, Methylmercury.
Grandjean, P., et al. 1997. Cognitive deficit in 7-year-old children with prenatal exposure to methyl mercury. Neurotoxicol Teratol 19:417-428; Murata, K., et al. 2004. Delayed Brainstem Auditory Evoked Potential Latencies in 14-year-old Children Exposed to methylmercury. J. Pediatr. 144:177-183.
Myers, G.J., et al. 2003. Prenatal methylmercury exposure from ocean fish consumption in the Seychelles child development study. Lancet 361:1686-1692.
See Rice et al., footnote 29 above.
JECFA report, cited in footnote 29 above, page 615.
Carrington, C.D. and P.M. Bolger 2002. An exposure assessment for methylmercury from seafood for consumers in the United States. Risk Anal. 22:689-699.
For FDA data on mercury levels in fish; see http://www.cfsan.fda.gov/~frf/sea-mehg.html.
Mahaffey, K.R., R.P. Clickner, and C.C. Bodurow. 2004. Blood Organic Mercury and Dietary Mercury Intake: National Health and Nutrition Examination Survey (NHANES), 1999 and 2000. Environ Health Perspect 112: 562-570; Schober, S.E., et al. 2003. Blood Mercury Levels in US Children and Women of Childbearing Age, 1999-2000. JAMA 289(13):1667-1674; see Jones, R.L., et al. 2004. Blood Mercury Levels in Young Children and Childbearing-Aged Women – United States, 1999-2002. Morbidity and Mortality Weekly Reports 53(43):1018-1020, November 5, 2004. United States Centers for Disease Control and Prevention.
For a review of this evidence, see Mahaffey, K.R. 2005. “Update on Mercury,” presentation at the 2005 Fish Forum, September 19, 2005 (available at http://epa.gov/waterscience/fish/forum/2005/presentations/Monday%20Slides%200919/afternoon/Mahaffey_Fis h%20Forum%202005%20-%20Mahaffey%20Final.ppt).
See JECFA report (cited in footnote 29 above) pp. 607-609.
These reports are cited in footnote 29 above.
Mercury in fish was discussed extensively at a December 10, 2003 meeting of the FDA’s Food Advisory Committee (transcript available at http://www.fda.gov/ohrms/dockets/ac/cfsan03.html). It was addressed in written comments submitted by industry groups and by Consumers Union among others.
Many examples of denial of the evidence of mercury risks and promotion of increased fish (and specifically, tuna) consumption are accessible on the United States Tuna Foundation web site. For example, see http://www.tunafacts.com/news/eat_more_fish_081505.cfm. Also see, http://www.fishscam.com, an industry-funded web site created by a public relations firm in an effort to discredit mercury risk concerns.
For example, see Groth, E. 2005. Risks and Benefits of Fish Consumption: Yes, Mercury is a Problem. Report prepared for Oceana and the Mercury Policy Project, December, 2005 (available at http://www.oceana.org/fileadmin/oceana/uploads/mercury/Final_Report_12-5.pdf).
See the report of the 2005 CCFAC meeting, ALINORM 05/28/12, paragraphs 201-205.
Available at ftp://ftp.fao.org/codex/ccfac37/fa37_35e.pdf
See Report of 2005 CCFAC session (cited in footnote 43 above).
CCFAC has adopted a two-tiered system, with a list of species that should not exceed 1 ppm, i.e. large predatory fish that tend to accumulate relatively high mercury levels, and a second list that should not exceed 0.5 ppm, i.e. fish that tend to accumulate moderate but still relatively significant amounts of mercury.
“What You Need to Know About Mercury in Fish and Shellfish,” the current (2004) EPA/FDA advisory (available at http://www.epa.gov/waterscience/fishadvice/advice.html). For an Australian example, see the Food Safety Authority of New South Wales’s advice on mercury in fish for women who may become pregnant (available at http://www.foodauthority.nsw.gov.au/pregnancy.htm?lk=consinfo).