Human & Experimental Toxicology (2007) 26: 185
190

http://het.sagepub.com

Risk assessment of dietary exposure to

methylmercury in fish in the UK
BJ Maycock and DJ Benford
Food Standards Agency, 125 Kingsway, London WC2B 6NH, UK

Risk assessment of chemicals in food is generally based much fish can be consumed without undue risk from the
upon the results of toxicological studies in laboratory contaminants.
animals, allowing for uncertainties relating to inter- Consumption by adults of one weekly portion (140 g) of
species differences, human variability, and gaps in the shark, swordfish or marlin, would lead to an exceedance
database. Use of quantitative human data is preferable if
available, as in the example of methylmercury. Methyl-

of the guideline intake for methylmercury of 40 90%, set
to protect the developing fetus, without considering
mercury is a neurotoxic environmental contaminant, for intake from the rest of the diet. Pregnant women and
which fish is the main source of dietary exposure. Human women who may become pregnant within 1 year were,
data from poisoning incidents and epidemiological stu- therefore, advised to avoid consumption of these species.
dies have been used by expert committees to derive a Intakes in other adults would be within a higher guide-
guideline intake level for methylmercury, based on the
line intake, set to protect groups of the population other
susceptibility of the most sensitive lifestage, the develop-
ing fetus. than the developing fetus. However, consumption by
In the UK, an expert group of nutritionists and children of one weekly portion of these species could
toxicologists was formed to review the benefits and risks lead to an exceedance of this guideline intake by up to
associated with fish consumption. A formal risk benefit
analysis was not possible because the nutritional data

60%, without considering intake from the rest of the diet.
It was, therefore, advised that consumption of these
were not sufficiently quantitative. The Committee on species by children should be avoided. Human & Experi-
Toxicity of Chemicals in Food, Consumer Products and mental Toxicology (2007) 26, 185
190
the Environment (COT), therefore, modified the risk
assessment approach to derive different guideline intake
levels for different subgroups of the population. The COT Key words: fish; mercury; neurotoxins; risk assessment; risk

opinion was used to provide targeted advice on how benefit

Introduction
Fish is a nutritious food, being a good source of levels in oily fish that are higher in the aquatic
protein, rich in certain vitamins and minerals, and food chain. Methylmercury tends to accumulate in
containing long chain n3 polyunsaturated fatty long-lived predatory fish, such as tuna, shark and
acids (LC n3-PUFAs). Intake of LC n3-PUFAs from swordfish, and fish is the major source of dietary
fish or fish oils has been linked to a reduced risk of exposure. Other contaminants of possible concern
cardiovascular disease, leading to recommendations found in fish include arsenic, brominated flame-
on the minimum amounts of fish that people should retardants, cadmium and lead. Risk assessment of
consume per week, and to benefits on pre- and post- chemical contaminants in fish is required to deter-
natal neurodevelopment. mine levels of fish intake that may safely be
However, fish can also be a significant dietary consumed without unnecessarily restricting con-
source of a number of environmental contaminants. sumption of this nutritious food. This paper dis-
Dioxins and polychlorinated biphenyls are lipophi- cusses an example of how human data have been
used in the risk assessment of neurotoxic contami-
lic persistant organic pollutants, found in higher
nants found in fish, focusing on the risk assessment
in the UK of methylmercury.
*Correspondence: Dr Diane J Benford, Food Standards Agency,
Room 508C, Aviation House, 125 Kingsway, London WC2B 6NH,
UK Methylmercury
E-mail: diane.benford@foodstandards.gsi.gov.uk

Received 31 October 2005; revised 2 April 2006; accepted Mercury is widespread in the environment due to its
3 May 2006 uses in many products and industrial processes, and
– 2007 SAGE Publications 10.1177/0960327107070565
Downloaded from het.sagepub.com at PENNSYLVANIA STATE UNIV on May 11, 2016
 Risk assessment of neurotoxicants in fish
BJ Maycock and DJ Benford
186

its natural release from the Earth’s crust. Mercury of marine and freshwater fish,1 with higher levels
entering the aquatic environment, in whatever found in local areas of mercury contamination due
form, is converted into methylmercury by micro- to industrial processes.7
organisms.1,2 Methylmercury is able to enter fish
either directly from the water or by being concen-
trated up the food chain. It has a half-life of Risk assessment
approximately 2 years in fish, and therefore, larger,
older fish, particularly predatory species, contain Risk assessment of chemicals in food consists of four
the highest levels of methylmercury. Consumption stages
referred to as hazard identification, hazard
of fish contaminated with methylmercury is the characterization, exposure assessment, and risk
major source of exposure of the population to characterization.8 Hazard identification and charac-
methylmercury.2 terization involves identification of adverse health
Methylmercury is rapidly absorbed from the effects and characterization of the dose
response
relationship, and may result in the establishment of
gastrointestinal tract. Approximately 95% of the
a health-based guidance value, such as a tolerable
methylmercury ingested is absorbed, and peak
intake. In general, these steps rely on the results of
blood methylmercury levels are reached within 6
toxicological studies in animals. Quantitative hu-
hours. Methylmercury is able to cross plasma
man data are the preferred basis, where available.
membranes more readily than inorganic mercury
Methylmercury is an example of a chemical for
compounds, and readily crosses the blood
brain
which a large amount of human data are available,
barrier and the placenta. Animal studies have
and have been used in the risk assessment.
shown that methylmercury is particularly able to
Excessive intakes of methylmercury in laboratory
enter the fetal brain, since the blood
brain barrier is
animals have resulted in neurotoxicity, renal toxi-
not fully developed. Methylmercury accumulates in
city, hepatotoxicity, reproductive and developmen-
the brain and is slowly converted to inorganic
tal toxicity, immunotoxicity, carcinogenicity and
mercury.2,3 decreased body weight. It is possible to determine
Methylmercury is neurotoxic in humans and levels of intake that are without adverse effects in
animals, though the mechanism by which this some of the animal studies; however, there are many
occurs is still unclear. Cases of neurotoxicity, and uncertainties in extrapolating the results of these
some fatalities in humans, have been reported since studies to humans. For example, the most sensitive
the late nineteenth century.4 These cases were due endpoint differs in different species. In a 52-month
to laboratory exposure, industrial exposure during feeding study in Rhesus and Mullata monkeys, it
the manufacture of anti-fungal seed dressings, and was weight loss,9 whereas in a 24-month study in
isolated incidents of consumption of methylmer- rats, it was renal toxicity;10 and the doses required to
cury-dressed grains intended for planting. It was not produce toxic effects vary considerably between
until the second half of the twentieth century that it species. It has been observed, for example, that
was appreciated that methylmercury could also doses of about 1 mg/kg bw/day have frequently
present risks to the general population. been tested in rodents with minimal effects, whereas
In the 1950s, an epidemic of a new disease this is in the range of intakes that has been
occurred in and around the coastal Japanese town associated with severe neurotoxicity in humans.11
of Minamata.5 The residents of Minamata and In addition, developmental neurotoxicity, which is
surrounding villages relied largely on fish caught considered the most sensitive endpoint in humans,
in Minamata bay as a protein source. In 1932, a has not been adequately investigated in animals.
Minamata company began making acetaldehyde, Most of the available human data were originally
and inorganic mercury, used as a catalyst in the provided by several mass methylmercury-poisoning
production process, started leaking into the bay. outbreaks. These included the incidents in the areas
Some methylmercury was also produced in the surrounding Minamata and Niigata in Japan, result-
process and released. In the early 1950s, production ing from the consumption of fish heavily contami-
boomed, but signs of neurotoxicity began to appear nated with methylmercury, and the consumption of
in residents. This had reached epidemic proportions grain dressed with methylmercury compounds as a
by 1956, but it was 1959 before methylmercury fungicide, by a large number of people in Iraq. The
poisoning was diagnosed and linked to the release effects of these very high level exposures ranged
of mercury into the bay. In the mid-1960s, a simi- from transient paresthesia to severe sensory and
lar outbreak developed along the Agano River in motor impairment and death in exposed children
Niigata, Japan.6 Low levels of methylmercury have and adults, and included mental retardation, cere-
subsequently been detected globally in most species bral palsy, deafness, blindness and dysarthria in

Downloaded from het.sagepub.com at PENNSYLVANIA STATE UNIV on May 11, 2016

 Risk assessment of neurotoxicants in fish
BJ Maycock and DJ Benford
187

individuals who had been exposed in utero . Renal mercury concentrations of 0
3 mg/g; of the two with
toxicity also occurred in more severely affected low hair mercury concentrations, maternal fish
cases. consumption had been more than three times/week
The Joint Expert Committee on Food Additives for one and less than three times/week for the other.
(JECFA) of the Food and Agriculture Organization Twenty-six tests were administered, assessing the
and World Health Organization originally consid- domains of general intelligence, language develop-
ered methylmercury at its Sixteenth meeting in ment, fine- and gross-motor co-ordination, academic
1972.12 The JECFA established a Provisional Toler- attainment and social adjustment, and results were
able Weekly Intake (PTWI) of 5 mg/kg bw for mercury, adjusted for a wide range of potential confounding
of which two-thirds (3.3 mg/kg bw) was allocated to factors. Maternal hair mercury concentration was
methylmercury. The PTWI was based primarily on associated with poorer scores in full-scale IQ,
human data from the Minamata and Niigata poison- language development, visual-spatial skills, and
ing incidents. The lowest hair and blood mercury gross motor skills. Regression analyses, using the
levels reported in adults with neurological impair- actual measurements of mercury in hair rather than
ment were converted to a steady state mercury intake grouping data, did not find significant associations
from the diet, to determine the PTWI. The conver- between maternal hair concentration and test per-
sion was based on data provided by several small formance; however, the results were highly affected
human studies on the relationship between hair by the results for a single child whose mother had
mercury content and mercury intake, and mercury the highest hair mercury concentration. When this
level in blood cells and mercury intake. result was excluded, the results of six of the tests
In 1989, JECFA noted that developmental neuro- were significantly associated with maternal hair
toxicity appeared to be the most sensitive endpoint mercury concentration.11,14
and, therefore, pregnant and nursing mothers were In the Seychelles study, evaluations of neurode-
likely to be more susceptible to methylmercury.13 velopment were made in a large cohort at up to 8
However, the available data were insufficient to years of age (717 children assessed at up to 5.5
recommend an intake specifically for those sub- months, and 643 at 8 years). At 5.5 years of age, six
groups, and the PTWI of 3.3 mg/kg bw for methyl- neurodevelopmental tests were performed
the
mercury was retained at this and a number of McCarthy Scales of Children’s Abilities, the Pre-
subsequent discussions. school Language Scale, the Woodcock-Johnson Ap-
Subsequently, data have become available from plied Problems and Letter and Word Recognition
epidemiological studies, which have investigated Tests of Achievement, the Bender Gestalt test and
whether exposure during pregnancy to methylmer- the Child Behaviour Checklist. Average maternal
cury, through high levels of fish consumption, may hair mercury concentration at pregnancy was 6.8 mg/
be associated with subtle neurodevelopmental ef- g, equivalent to approximately 27 mg/L in blood. No
fects in children. These included prospective cohort associations were found between maternal hair
studies in New Zealand, the Faeroe Islands, and the mercury concentration and test scores.15 At 8 years
Seychelles. of age, a greater number of tests were used, measur-
The New Zealand study evaluated 237 children, at ing 21 endpoints. Statistically significant differ-
4 and 6 years of age, exposed in utero to methyl- ences at the 5% level were found in 2/21 tests;
mercury through a maternal diet high in fish.11,14 A these were decreased performance in the grooved
‘high-dose’ group was defined as children of women pegboard test with the non-dominant hand, and
who had hair mercury concentrations during preg- better score for the hyperactivity index of the
nancy of
/6 mg/g. At 4 years of age, abnormal or Connors Teacher Rating Scale with increased
questionable scores were observed in the Denver methylmercury exposure. It was concluded that
Development Screening Test (DDST) in 52% of high- the overall results were consistent with there being
dose children compared to 17% of controls, pair- no effect of prenatal methylmercury exposure in this
matched as far as possible for maternal ethni- cohort.16
city, hospital of birth, maternal age and child age In contrast to the results from the Seychelles
(P B/0.05). study, in the Faeroe Islands study (917 children
At 6 years of age, each high-dose child was assessed), prenatal mercury exposure was related to
matched to three controls for ethnic group, sex, neuropsychological dysfunction in the children of
maternal age, maternal smoking, areas of maternal mothers with maternal hair concentrations of ap-
residence, and duration of maternal residence in proximately 10 mg/g, when evaluated at 7 years of
New Zealand. One control had a hair mercury age.17 Increased methylmercury exposure was asso-
concentration of 3
6 mg/g, and two had low hair ciated with poorer performance in finger tapping,

Downloaded from het.sagepub.com at PENNSYLVANIA STATE UNIV on May 11, 2016

 Risk assessment of neurotoxicants in fish
BJ Maycock and DJ Benford
188

reaction time on a continuous performance test, confidence limit (BMDL) for the Faeroe Islands
digit span in the revised Wechsler Intelligence Scale study were calculated and averaged. This averaged
for children, the Boston Naming test and the hair concentration was converted to a steady state
California Verbal Learning test for children. When dietary methylmercury intake using a one compart-
an alternative approach to adjusting for confounding ment pharmacokinetic model, which used values
factors was used, prenatal methylmercury exposure appropriate for pregnancy. The resulting intake was
was also associated with decreased performance in divided by uncertainty factors of 2 and 3.2, to
the Block Design test of the revised Wechsler account for inter-individual variation in hair:blood
Intelligence Scale for children, and errors on copy- ratios and the association between blood mercury
ing in the Bender Gestalt test. concentration and intake. This resulted in a PTWI of
It has been suggested that the difference between 1.6 mg/kg bw.
the results of the Seychelles and Faeroe Islands
studies may be due to confounding of the Faeroe
Risk benefit analysis
Islands study results, by co-exposure to polychlori-
nated biphenyls (PCBs). While the relatively high The UK Food Standards Agency sought advice from
level of methylmercury exposure in the Seychelles a subgroup of members of the UK advisory commit-
is due to frequent consumption of a range of fish, in tees on nutrition and toxicity on the benefits and
the Faeroe Islands the source of highest exposure to risks of fish consumption, particularly of oily fish, to
methylmercury is pilot whale, which is consumed bring these views together in order to provide clear
occasionally. Pilot whale contains high levels of and helpful advice to consumers. A formal risk

methylmercury, but also high levels of PCBs. In benefit analysis was not possible because the nutri-
the Faeroe Islands study, analysis of cord tissue at tional data were not sufficiently quantitative. There-
birth demonstrated a strong correlation between fore, toxicological guidelines were developed for the
mercury concentration and PCB concentration, contaminants that would not be associated with
although the reliability of cord measurements com- appreciable risk to the health of different subgroups
pared to those of plasma or serum lipids is un- of consumers.21 As part of this process, the COT
known. PCB exposure has been associated with considered the toxicity of methylmercury in light of
adverse effects on neurodevelopment in a number the PTWI established by JECFA in 2003. The COT
of studies,18 although dose
response relationships agreed that the revised PTWI was protective of all
have not been adequately characterized. effects, including neurodevelopmental effects. How-
Further analyses of the Faeroe Islands data were ever, it found no evidence to suggest that the
performed to explore this issue. Based on 436 previous JECFA PTWI of 3.3 mg/kg bw/week was
analyses of umbilical cord PCB levels, children insufficiently protective for non-developmental ad-
were divided into tertiles of PCB exposure. The verse effects. Taking into account the half-life of
strength of the association between test score at 7 methylmercury, the COT considered that the JECFA
years of age and methylmercury exposure was not PTWI of 1.6 mg/kg bw/week should be used in
significantly affected by PCB tertile.19 assessing dietary exposure to methylmercury of
In additional analyses, the results for four of the women who were pregnant or likely to become
neurodevelopment tests, which indicated the possi- pregnant in the next year. The guideline of 3.3 mg/
ble effects of PCB intake, were stratified by methyl- kg bw/week could be used for other groups.
mercury intake. The greatest effects of PCBs were
observed in the highest methylmercury tertile. Tests
Exposure assessment and risk characterization
for interaction did not reach statistical significance;
however, the authors suggested that methylmercury The intake guidelines were compared with UK
exposure may affect the neurotoxic effects of PCB survey data on methylmercury content of a wide
exposure, but not the reverse.20 range of domestic and imported fish and shellfish,
The data from the larger Faeroe Islands and and products made from fish and shellfish.22
Seychelles studies were used by JECFA to determine The highest levels of methylmercury were found
a new PTWI, which would be protective of the in shark (mean: 1.52 mg/kg), swordfish (mean:
developing fetus, as well as other groups of the 1.36 mg/kg) and marlin (mean: 1.09 mg/kg)
population.3 Maternal hair mercury concentrations (Table 1). Tuna (mean: 0.40 mg/kg fresh tuna,
corresponding to a no observed adverse effect level mean: 0.19 mg/kg tinned tuna) contained higher
(NOAEL) from the Seychelles study (mean maternal levels of methylmercury than other commonly eaten
hair concentration in the highest exposed children fish, but lower levels than shark, swordfish and
in the cohort) and to a benchmark dose lower marlin.

Downloaded from het.sagepub.com at PENNSYLVANIA STATE UNIV on May 11, 2016

 Risk assessment of neurotoxicants in fish
BJ Maycock and DJ Benford
189
Table 1 Examples of differences in levels of mercury in different may be at risk from a small exceedance. It was,
fish species21
therefore, considered that maternal consumption of
Hg (mg/kg fish, fresh weight) one 140-g portion/week of these species of fish
during pregnancy could present a risk to the fetus.
Minimum Maximum Median Mean
Other adults consuming one portion or less of
Oily fish these species per week would be unlikely to exceed
Salmon 0.03 0.08 0.05 0.05 the COT guideline intake and, therefore, would not
Trout 0.01 0.1 0.05 0.06
Tuna
fresh 0.1 1.5 0.31 0.40 be expected to be at risk of neurotoxicity. However,
Swordfish 0.15 2.7 1.4 1.4 it was estimated that consumption of one portion
Orange roughy 0.53 0.65 0.60 0.60 per week of these species by children would result
Non-oily fish in a methylmercury intake of 3.0
5.2 mg/kg bw/
Halibut 0.04 0.62 0.29 0.29
Hoki 0.08 0.31 0.19 0.19 week, without considering intake from the rest of
Monkfish 0.1 0.30 0.20 0.20 the diet, resulting in an exceedance of up to 60% of
Pollack 0.007 0.02 0.01 0.01 the COT guideline intake. There could, therefore, be
Sea bass 0.03 0.09 0.07 0.07
Sea bream 0.05 0.06 0.05 0.05 a risk to children from regularly consuming these
Shark 1.0 2.2 1.4 1.5 species of fish.
Marlin 0.41 2.2 0.88 1.1 Consumption of up to two 140-g portions/week of
Tuna
canned 0.03 0.71 0.15 0.19
fresh tuna or four 140-g portions/week of canned
tuna would not lead to an exceedance of the COT
guideline intake in children or the PTWI in women
Dietary estimates of mercury exposure were made
who were pregnant or likely to become pregnant in
for fish species for which reliable consumption data
the next year.
were available (salmon, prawns and canned tuna)
from UK dietary surveys.23
25 The COT concluded
that these intakes were unlikely to be associated Conclusions
with adverse effects in the developing fetus or at
Risk assessment of methylmercury is based almost
other life stages.21
entirely on human data, originally from large,
As reliable consumption data were not available
poisoning incidents, and more recently from epide-
for shark, swordfish or marlin from the UK National
miological studies. The critical effect is neurotoxi-
Diet and Nutrition Surveys, intake estimates were
city, and the developing nervous system is most
based on average fish portion sizes (Table 2). Con-
sensitive. Establishment of tolerable intakes gener-
sumption by adults of one weekly portion (140 g) of ally aims to protect the most sensitive individuals
shark, swordfish or marlin would result in an intake against the most sensitive adverse effects. If the
in the range 2.2
3.0 mg/kg bw, without considering tolerable intakes are set to protect against effects on
intake from the rest of the diet. This intake would the fetus, they are likely to be over-precautionary for
exceed the revised JECFA PTWI by up to 90%. The individuals who will not become pregnant. Devel-
PTWI is not a threshold for risk, and it is not opment of different intake guidelines for different
possible to quantify any risk from a particular groups of the population enables risk managers
exceedance. Instead, there is uncertainty regarding to target dietary advice, such as restrictions on
the risk, which reduces with increasing exceedance. eating certain species of fish, to the appropriate
However, since the uncertainty factors used to set sub-groups, without unnecessarily restricting con-
the PTWI were to account for the potential for inter- sumption by other subgroups. Different population
individual variability, the most sensitive individuals subgroups already receive specific dietary advice,

Table 2 Mercury intake from one weekly portion of shark, swordfish, marlin, fresh tuna or canned tuna (consumers of fish only)21

Age group Body weight Average portion size* Weekly mercury intake assuming one portion of fish
(years) (kg) (g) per week (mg/kg bw/week)

Shark Swordfish Marlin Fresh tuna Canned tuna

1.5
4.5 14.5 50 5.24 4.62 3.79 1.38 0.66

4
6 20.5 60 4.44 3.90 3.22 1.17 0.56
7
10 30.9 85 4.17 3.69 3.04 1.10 0.52
11
14 48.0 140 4.44 3.92 3.21 1.17 0.55
15
18 63.8 105 2.51 2.21 1.82 0.66 0.31
Adults 70.1 140 3.04 2.68 2.20 0.80 0.38

*The average portion size that each age group of the population would consume at a single meal event for fish consumption.23
26

Downloaded from het.sagepub.com at PENNSYLVANIA STATE UNIV on May 11, 2016

 Risk assessment of neurotoxicants in fish
BJ Maycock and DJ Benford
190

particularly women who are pregnant, breastfeeding different population sub-groups is unlikely to be
or trying to conceive, and those feeding young overly confusing to consumers, providing the key
children. Therefore, providing different advice to messages are communicated clearly.

References

1 Clarkson TW. The three modern faces of mercury. months of age in the Seychelles Child Development
Environ Health Perspect 2002; 110: 11
23. Study. J Am Med Assoc 1998; 280: 701
707.
2 National Research Council. Toxicological effects of 16 Myers GJ, Davidson PW, Cox C, Shamlaye CF, Palumbo
methylmercury. National Academy Press, 2000. D, Cernichiari E, Sloane-Reeves J, Wilding GE, Kost J,
3 Joint FAO/WHO Expert Committee on Food Additives. Huang LS, Clarkson TW. Prenatal methylmercury
Safety evaluation of certain food additives and con- exposure from ocean fish consumptions in the Sey-
taminants . WHO Food Additives Series No. 52. World chelles child development study. Lancet 2003; 361:
Health Organisation, 2004. 1686
92.
4 Hunter D, Bomford RR, Russell DS. Poisoning by 17 Grandjean P, Weihe P, White RF, Debes F, Araki S,
methylmercury compounds. Q J Med 1940; 9: 193
Yokoyama K, Murata K, Sorensen N, Dahl R, Jorgensen
213. PJ. Cognitive deficit in 7-year old children with
5 Watanabe C, Satoh H. Evolution of our understanding prenatal exposure to methylmercury. Neurotoxicol
of methylmercury as a health threat. Environ Health Teratol 1997; 19: 417
28.
Perspect 1996; 104: 367
79. 18 Longnecker MP, Wolff MS, Gladen BC, Brock JW,
6 Myers GJ, Davidson PW, Weiss B. Methylmercury Grandjean P, Jacobson JL, Korrick SA, Rogan WJ,
exposure and poisoning at Niigata, Japan. Seychelles Weisglas-Kuperus N, Hertz-Picciotto I, Ayotte P,
Med Dent J 2004; 7: 132
33. Stewart P, Winneke G, Charles MJ, Jacobson SW,
7 Grandjean P, White RF, Nielsen A, Cleary D, de Dewailly E, Boersma ER, Altshul LM, Heinzow B,
Oliveira Santos EC. Methylmercury neurotoxicity in Pagano JJ, Jensen AA. Comparison of polychlorinated
Amazonian children downstream from gold mining. biphenyl across studies of human neurodevelopment.
Environ Health Perspect 1999; 107: 587
91. Environ Health Perspect 2003; 111: 65
69.
8 Renwick AG, Barlow SM, Hertz-Picciotto I, Boobis AR, 19 Budtz-Jorgensen E, Keiding N, Grandjean P, White RF.
Dybing E, Edler L, Eisenbrand G, Greig JB, Kleiner J, Methylmercury neurotoxicity independent of PCB
Lambe J, Muller DJ, Smith MR, Tritscher A, Tuijtelaars exposure. Environ Health Perspect 1999; 107: A236

S, van den Brandt PA, Walker R, Kroes R. Risk 37.
characterisation of chemicals in food and diet. Food 20 Grandjean P, Weihe P, Burse VW, Needham LL, Storr-
Hansen E, Heinzow B, Debes F, Murata K, Simonsen H,
Chem Toxicol 2003; 41: 1211
71.
Ellefsen P, Budtz-Jorgensen E, Keiding N, White RF.
9 Kawasaki Y, Ikeda Y, Yamamoto T, Ikeda K. Longterm
Neurobehavioral deficits associated with PCB in 7-
toxicity study of methylmercury chloride in monkeys. J
year-old children prenatally exposed to seafood neu-
Food Hyg Soc Japan 1986; 27: 528
52.
rotoxicants. Neurotoxicol Teratol 2001; 23: 305
17.
10 Munro IC, Nera EA, Charbonneau SM, Junkins B,
21 Scientific Advisory Committee on Nutrition and Com-
Zawidzka Z. Chronic toxicity of methylmercury in
mittee on Toxicity of Chemicals in Food, Consumer
the rat. J Environ Pathol Toxicol 1980; 3: 437
47. Products and the Environment. Advice on fish con-
11 Joint FAO/WHO Expert Committee on Food Additives. sumption: benefits and risks . TSO, 2004. http://
Safety evaluation of certain food additives and con- www.food.gov.uk/multimedia/pdfs/fishreport2004-
taminants . WHO Food Additives Series No. 44. World full.pdf
Health Organisation, 2000. 22 Food Standards Agency. Methylmercury in imported
12 Joint FAO/WHO Expert Committee on Food Additives. fish and shellfish and their products . Food Surveil-
Evaluation of certain food additives and the contami- lance Information Sheet No. 40/03. Food Stan-
nants mercury, lead and cadmium . WHO Technical dards Agency, 2002. http://www.food.gov.uk/science/
Report Series No. 505. World Health Organisation, surveillance/fsis-2003/fsis402003
1972. 23 Henderson L, Gregory J, Swan G. The national diet and
13 Joint FAO/WHO Expert Committee on Food Additives. nutrition survey: adults aged 19 to 64 years ; Volume 1:
Evaluation of certain food additives and contami- Types and quantities of foods consumed . TSO, 2002.
nants. WHO Technical Report Series No. 776. World 24 Gregory J, Collins DL, Davies PSW, Hughes JM, Clarke
Health Organisation, 1989. PC. National diet and nutrition survey: children aged
14 Crump KS, Kjellstrom T, Shipp AM, Silvers A, Stewart one-and-a-half to four-and-a-half years ; Volume 1:
A. Influence of prenatal mercury exposure upon Report of the national diet and nutrition survey.
scholastic and psychological test performance: bench- HMSO, 1995.
mark analysis of a New Zealand cohort. Risk Anal 25 Gregory J, Lowe S, Bates CJ, Prentice A, Jackson LV,
1998; 18: 701
13. Smithers G, Wenlock R, Farron M. National diet and
15 Davidson PW, Myers GJ, Cox C, Axtell C, Shamlaye C, nutrition survey: young people aged 4 to 18 years ;
Sloane-Reeves J, Cernichiari E, Needham L, Choi A, Volume 1: Report of the national diet and nutrition
Wang Y, Berlin M, Clarkson TW. Effects of pre-natal survey. TSO, 2000.
and postnatal methylmercury exposure from fish con- 26 Mills A, Tyler H. Food and nutrient intakes of British
sumption on neurodevelopmental outcomes at 66 infants aged 6
12 months . HMSO, 1992.

Downloaded from het.sagepub.com at PENNSYLVANIA STATE UNIV on May 11, 2016