Mercury: risk assessment

Later, in the mid 1960’s a similar neurological disease developed along the Agano River in Niigata Prefect of Japan. The cause of Minamata Disease was not at first understood, but was traced to methylmercury poisoning. In each case, the sources of the mercury in these mercury poisonings was a single chemical industrial facility discharging mercury-containing waste into surface waters. Prior to the Minamata Disease being attributed to methylmercury, industrial discharge of mercury was not considered a human health concern. Studies of the Minamata Bay mercury exposure for the first time demonstrated the susceptibility of the fetus to damage, as evidenced by some children with severe neuromotor defects acquired by in-utero mercury exposure, while the mothers only suffered mild effects. The World Health Organization and the U.S. Food and Drug Administration use data from the Japanese studies to set standards for maximum mercury content in food.

The Japanese government has established a research institute to study the effects of Minamata Disease and to treat and support the victims. The web site for the Japanese Ministry of the Environment, National Institute for Minamata Disease is: http://www.nimd.go.jp/english/index.html

Iraq Mercury Poisoning Incident

During the winter of 1971-1972, mercury poisoning occurred in rural areas of Iraq. Wheat seeds, intended for crop planting, which had been treated with methylmercury as a fungicide, were distributed free in rural areas. Some seeds were ground into flour, baked into bread and consumed. Of an estimated 50,000 people exposed to the contaminated bread, 459 died, and 6,530 were hospitalized. This exposure also demonstrated the susceptibility of the fetus to in-utero exposure to methylmercury. 10 ppm mercury in maternal hair was shown to have an adverse dose-response by the children born after the exposure of the mothers. Studies of this exposure are the current basis for the U.S. EPA’s Reference Dose (RfD). Below is a dose-response developed based on the data from Iraqi exposure study.

Dose-response curve from article: Cox C., D. Marsh, G. Myers, and T. Clarkson, (1995). Analysis of data on delayed development from the 1971-1972 outbreak of methylmercury poisoning in Iraq: assessment of influential points. NeuroToxicology, 16 (4), 727-730.

In the Iraq poisoning, of an estimated 50,000 people exposed to the contaminated bread, 459 died, and 6,530 were hospitalized.

Photographs from Neurotoxicology, 1995, Vol. 16, No. 4

Documentation of the Iraq study are contained in the following journal articles:.
[Marsh, Arch. Neurol, 44 1017-1022, 1987]
[Cox, Environ.Res. 31, 640-649, 1989]
[Cox, Neurotoxicology 16(4) 727-730, 1995]

Low Dose Studies

In the mid 1990’s, several long-term studies were designed to better quantify the lower exposure effects of methylmercury. Populations were selected for study that had high intake of methylmercury from a seafood diet with a minimum of confounding factors. The two most significant studies were the Seychelle Islands study and the Faroe Islands study. Both studies identified in-utero exposure of children and followed their neuro-development. The Seychelle Island study concluded that no adverse effect could be shown for a chronic low dose of methylmercury, while the Faroe Island study did demonstrate a negative developmental effect at 11 ppm maternal hair concentration of mercury.

An evaluation of several studies was reviewed by an expert panel sponsored by the National Institute of Environmental Health Sciences in Raleigh, NC, in November 1998. This conference primarily addressed differences found between the Faroe Island study and the Seychelles Island study. The New Zealand study was not included in the Raleigh review because a peer review had not been completed. The report from the conference is located at: http://ntp-server.niehs.nih.gov/main_pages/PUBS/MeHgAppIA.html

Subsequent review of the two studies did identify possible reasons for differences in results (v)Error! Bookmark not defined., including:

Differences in population groups resulting in potential differences in genetic susceptibility.
Differences in age of children at evaluation.
Type of testing performed to assess development in children. In the Seychelle Island study, generalized developmental assessments were performed similar to IQ tests. In the Faroe Island study, more specific motor and developmental tests were performed as an assessment. The Faroe Island study tests are reported to be more sensitive to detecting neurological effects.
Differences in the distribution of mercury exposure. The Faroe Islanders eat fewer meals of fish and marine mammals, but the mammal meals (pilot whale) contained a higher dose of methylmercury. The Seychelle Island population did not eat marine mammals, but had a larger intake of seafood.
Differences in exposure to PCBs. The Faroe Island study group was also exposed to PCBs from their diet, whereas the Seychelle Island group’s PCBs were below detectable levels.
Differences in the biomarker used to test for mercury exposure. The Faroe Island study used Hg concentration in umbilical cord blood as the primary measurement of prenatal exposure and maternal hair mercury concentration as a secondary exposure measurement. It was evaluated by NRC that cord blood may be a more sensitive measurement of in-utero exposure to the fetus. Therefore, the Faroe Island study may have been more precise at measuring the exposure for the unborn child.

The disparity of results between the Faroe Island and the Seychelle Island studies focused attention on an older study done in New Zealand. The New Zealand study had not undergone a peer review and, consequently, had not received much scientific attention. A peer review was later completed and the study was found to be sound. The New Zealand study did find adverse mental development in children exposed in-utero to methylmercury at a maternal hair mercury concentration of 6 ppm. The New Zealand study had used generalized developmental testing much like the Seychelle Island study, and showed an effect at 6ppm mercury in maternal hair samples. The New Zealanders’ diet was also more similar to that of the Seychelles Islanders, without the potential for exposure peaks due to the consumption of marine mammals and without significant exposure to PCBs. The three studies together tend to rule out some of the more convincing theory why the Seychelles and Fareo Island studies found different results.

The New Zealand data was published in a peer review article in 1998, Crump, Risk Anal. 18(6) 701-713.

U.S. EPA contracted with the National Research Council to review existing information on the toxic effects of Mercury, and to make a recommendation on future revision of regulatory standards for mercury exposure. The NRC report was published in 2000, and based on the weight of the evidence and by using a conservative bias to public health decisions, recommends the use of the Faroe Island Study for the reevaluation of the mercury RfD used by the U.S. EPA. (ii)Error! Bookmark not defined. The NRC report is contained in a book, Toxicological Effects of Methylmercury, which can be accessed or purchased at: http://books.nap.edu/books/0309071402/html/index.html

Risk Assessments by Governmental and International Agencies

Below is a summary of a chart contained in Table 1-1 of the NRC publication, Toxicological Effects of Methylmercury, National Research Council, National Academy Press, 2000. It lists a number of government/international agencies and the corresponding evaluation of acceptable mercury exposure. An interesting fact contained in the chart is the listed uncertainty factors (UFs) for mercury exposure, which range from 3 to 10. Normally UF are in the range of 100-1000 when animal testing is used, however since there is a large number of human exposure studies of the most sensitive human group, the UF is reduced.

Agency Key Study Critical Dose

Agency	Key Study	Critical Dose ug/kgbw/d	Uncertainty Factor	Acceptable Level ug/kgbw/d
EPA	Iraq Study	1.1	10	0.1
ASTDR	Seychelles Study	1.3	4.5	0.3
FDA	Japanese data	4.3	10	1ppm in food equivalent to 0.5 ug/kgbw/d
IECFA	Japanese	4.3	10	0.5
Health Canada	Seychelles, New Zealand	1.0	5	0.2
North Carolina	Seychelles Study	0.5	3	.2
Washington State	Faroe Island	0.8	10	0.8

EPA’s risk assessment, which uses the Iraqi study, identified an uncertainty factor, for human population variability, of 3, and an additional factor of 3 to account for the lack of data on reproductive effects and sequelae adult paresthesia. The two UFs were multiplied together for a total UF rounded up to 10. The ATSDR evaluation, which used the Seychelles study, included UFs of 1.5 for pharmacokenetic and 1.5 for pharmacodyamic variability of the population. An additional 1.5 was added to the UF to account for possible less sensitive neurological domain testing used in the Seychelles study. ATSDR added the component UFs for a total UF of 4.5.

The graph below plots the regulatory limits from NRC Table 1-1. Notwithstanding the source of the data, all protective limits are contained in a relative close one half of an order of magnitude. This reflects the large quantity of relatively good data on human exposure.

Risk Management

Mercury is a natural element and, as such, the total mass of mercury on the earth is not increased or decreased by human activity. Anthropogenic activity can, however, allow mercury to be released from its depositories in the earth’s crust and through industrial activities, resulting in concentrations of mercury in environmental zones that are detrimental to human and animal health. Risk management of mercury focuses on two schemes: one to reduce the human intake of mercury from environment sources, and the second to reduce the mobilization and concentration of mercury from industrial activities. The U.S. EPA has published extensive information on Mercury Exposure and Risk Management in its Mercury Study Report to Congress, EPA-452/R-97-009. This detailed report is available at: http://www.epa.gov/oar/mercury.html

Consumptions Limits

Intake is reduced by limiting the allowable mercury concentration in foods. The U.S. FDA current limit on commercial foods is 1 ppm. Many states have also limited mercury ingestion by issuing fish consumption advisory limits based on mercury concentrations found in fish. Even with these limits, the U.S. EPA report to congress estimates that 7% of U.S. women may consume mercury above the U.S. EPA reference dose of 0.1 ug per kg of body weight per day. Based on this estimate, the NRC has estimated that 60,000 children per year in the U.S. may be born which having been exposed in-utero to mercury levels above the protective limit (iii, iv).

Atmospheric Mercury

Any risk management of the source of anthropogenic mercury needs to be based on an understanding of the source and transportation mechanism of airborne mercury. The chart below gives the chemical-physical state of mercury and some characteristics which are important to atmospheric transportation.

Type of Mercury	Residence Time in Atmosphere	Transport Distance	Method for Deposit
Hg&Mac176; (elemental mercury vapor)	One Year	Global	Wet precipitation
Hgp (Mercury in atmospheric particulates)	Hours up to years, dependent on particle size	Local, regional, and global	Dry and wet precipitation
Hg2+ (divalent mercury)	Days	Regional	Dry and wet precipitation

The global atmospheric mercury from anphropogenic sources is estimated between 40 to 75% of the total atmospheric mercury content. This is a 2-4 times increase in atmospheric mercury. This increase may have an impact on human health for fish-eating populations, and therefore is a risk management concern as shown by the Faroe Island Study and the New Zealand Study. This broad spread of the mercury risk is of global concern and has motivated the U.S. EPA to seek regulations to reduce the industrial causes of mobilization of mercury.

The long residence time in the atmosphere results in mercury vapor (Hg&Mac176;) and some mercury particles (Hgp) being distributed globally. The majority of Hg&Mac176; released in the United States escapes the U.S. boundaries and joins the global atmospheric mercury reservoir. This reservoir in turn is responsible for 30% of mercury deposited in the U.S. Divalent mercury and the majority of particulate mercury precipitate out of the atmosphere and are deposited close to the release source. Therefore, the majority of Hg2+ and Hgp released in the US is deposited on the US land mass. This source of mercury precipitation accounts for the remaining 70% of U.S. mercury deposit. Thus, mercury air emissions have both a local deposition effect and a global distribution dependent on the type of mercury being released.

Predominately, the source for mercury released is from three combustion sources, specifically: coal combustion, burning of municipal solid waste (MSW) and incineration of medical waste (MW). Mercury contamination can be expected downwind from these types of facilities. An EPA computer has predicted mercury deposition in the 48 states. That map shows Hg deposition in the industrialized regions of the U.S. (vii).

In 1995, waste incineration accounted for 33% or the airborne mercury released. Airborne mercury released from MSW combustors and MW incinerators has been limited by excluding mercury from products, which find their way into the waste stream, and by removing mercury from the flue gases of the combustion units. Efforts to reduce mercury in consumer products have been successful in reducing the introduction of mercury into the MSW waste stream as well as into medical waste. Additionally, regulatory action by the U.S. EPA has required the waste incinerators and MSW burners to install mercury capture devices on the flue gases. Releases from these two sources is estimated to have been reduced by 90% as of 2000.

Coal-burning power plants and industrial boilers contributed to 46% of airborne mercury released in 1995, and are the largest contributors of mercury to the environment. Coal burners were not included in the actions taken at the waste burners mentioned above (viii). As other sources are reduced, the percentage of mercury from coal combustion will become significantly higher. Therefore coal combustion is the primary future target of the U.S. EPA effort to reduce mercury emissions to the environment.

Mercury is a natural component of coal and poses a difficult problem to reduce. The concentration of Hg in coal is small relative to the concentration found in MSW and MW. Therefore the amount of flue gas, which would have to pass through an emission control device, is orders of magnitudes larger to achieve the same amount of mercury removal. The cost associated with reduction of 90% of mercury from coal combustion in US utility boilers alone is estimated to be 5 billion dollars annually or $70,000 per pound of mercury removed. This is in comparison to the $870 per pound in a MSW combustor. It is likely that the high cost may result in selecting a solution from a number of alternative approaches, including fuel switching to natural gas, and selecting a reduction goal below 90% removal. The U.S. EPA is scheduled to publish proposed rules for utility coal boilers on December 15, 2003, with final rules scheduled for a year later.

Aquatic Mercury

Mercury in an aquatic environment can be converted into methylmercury by microorganisms. Methylmercury is readily absorbed by biota and can be passed up the aquatic food chain, resulting in bio-concentration of the mercury in the predator species. Subpopulations that rely on fish from the affected water may receive toxic levels of mercury. Therefore industrial sources with the potential to release high concentrations of mercury into a confined water body are a risk management concern. The most notable examples of mercury poisoning from industrial sources occurred in Minamata Bay, Japan, and Niigata Prefect, Japan. Similar poisoning of subpopulations has occurred in Canada, where subsistence Indian populations consumed fish from contaminated waters.

Control of point-sources of mercury contamination of surface waters is attainable within standard regulatory schemes such as water discharge permits, run-off control requirements, and controls on the use of chemicals released to the environment. Industrial processes have also been changed to reduce the use of mercury or to substitute a process that does not require mercury to be used. Developed countries have adopted control measures to prevent mercury discharges which lead to Minamata Disease.

References

Mahaffey, K. R. (2000). Recent Advances in Recognition of Low-level Methylmercury Poisoning. Current Opinion in Neurology, 13, 699-707.
National Research Council. Toxicological Effects of Methylmercury- Chapter 1, Washington D.C., National Academy Press, 2000.
National Research Council. Toxicological Effects of Methylmercury- Chapter 8, Washington D.C., National Academy Press, 2000.
Myers, G.J., P.W. Davidson, C. Cox, C. Shamlaye, E. Cernichiari, T.W. Clarkson. (2000). Twenty-Seven Years Studying the Human Neurotoxicity of Methylmercury Exposure. Environmental Research, 83, 275-285.
Jacobson, J.L. (2001). Contending with Contradictory Data in a Risk Assessment Context: The Case of Methylmercury. NeuroToxicology, 22, 667-675.
Dourson M. L., A. E. Wullenweber, K. A. Poirier. (2001).Uncertainties in the Reference Dose for Methylmercury. NeuroToxicology, 22, 677-689.
EPA (U.S. Environmental Protection Agency). (1997). Mercury Study for Congress Volume III, Fate and Transport of Mercury in the Environment, page 2-3, EPA-452/R-97-003, December 1997.
EPA (U.S. Environmental Protection Agency). 1997. Mercury Study for Congress Volume I, Executive Summary Table 3-1, Best Point Estimates of National Mercury Emission Rates by Category, EPA-452/R-97-003, December 1997.
EPA (U.S. Environmental Protection Agency). (1997). Mercury Study for Congress Volume I, Executive Summary Table 4-2, Potential Mercury Emission Reductions and Costs for Selected Source Categories, EPA-452/R-97-003, December 1997.