IPCS INCHEM Home

WHO FOOD ADDITIVES SERIES: 52

METHYLMERCURY (addendum)

First draft prepared by

Dr D. Bellinger
Harvard Medical School, Boston, Massachussetts, USA

Dr M. Bolger
United States Food & Drug Administration, College, Maryland, USA

Dr M. Dinovi
United States Food & Drug Administration, College, Maryland, USA

M. Feeley
Health Canada, Tunney’s Pasture, Ottawa, Ontario, Canada

G. Moreau
Health Canada, Tunney’s Pasture, Ottawa, Ontario, Canada

Prof A. Renwick
Clinical Pharmacology Group, University of Southampton, Southampton, United Kingdom

and

Dr J. Schlatter
Swiss Federal Office of Public Health, Zürich, Switzerland

Explanation

Biological data

Biochemical aspects

Absorption, distribution and excretion

Transfer from mother to offspring

Clearance

Pharmacokinetics

Effects on enzymes and other biochemical parameters

Cleavage of the carbon-mercury bond

Complexes with thiol radicals

Interaction with selenium

Toxicological studies

Acute toxicity

Long-term studies of toxicity and carcinogenicity

Reproductive toxicity

Single-generation studies

Developmental toxicity

Special studies

Neurotoxicity

Renal and hepatic toxicity

Immunomodulation

Other effects

Observations in humans

Study in the Seychelles

Pilot cohort neurodevelopmental evaluations at age 8 years

Neurodevelopmental evaluations of the full cohort at age 8 years

Study in the Faroe Islands

Other studies in children with a high intake of fish

Reproductive toxicity

Neurotoxicity in adults

Immunotoxicity

Cardiotoxicity

General health of residents of the Minamata City area

Safety assessments: analyses of levels of exposure and corresponding responses

Estimated dietary intake

Environmental sources of mercury and methylmercury

Biomarkers of exposure

Concentrations of methylmercury in seafood

Intake assessment

National estimates of intake

France

New Zealand

Australia

Japan

United States

International estimates of intake based on WHO GEMS/Food diets

Comments

Absorption, metabolism and excretion

Toxicological studies

Observations in animals

Observations in humans

Dose–response assessments

Estimated dietary intake

Evaluation

References

1. EXPLANATION

Methylmercury was evaluated by the Committee at its sixteenth, twenty-second, thirty-third and fifty-third meetings (Annex 1, references 30, 47, 83, 144). At the last meeting, the Committee reaffirmed the previously established provisional tolerable weekly intake (PTWI) of 200 µg of methylmercury (3.3 µg/kg bw) for the general population, but noted that fetuses and infants might be more sensitive than adults to its toxic effects. The Committee concluded that data from studies undertaken in the Seychelles and the Faroe Islands, which were evaluated at its fifty-third meeting, did not provide consistent evidence concerning neuro-developmental effects in children of women whose methylmercury intakes had resulted in burdens of mercury in hair of 20 mg/kg and below. Adverse effects on neurodevelopment were reported in the study in the Faroe Islands, but not in that in the Seychelles; however, different methods for assessing neurobehavioural effects had been used in the two cohorts. The Committee recommended that methylmercury be re-evaluated at a subsequent meeting when the results of the analysis of neurodevelopmental effects in the Seychelles cohort after 8 years and other relevant data had become available. The Committee noted that fish make an important nutritional contribution to the diet, especially in certain regions, and recommended that nutritional benefits be weighed against the possibility of adverse effects when limits were being considered for concentrations of methylmercury in fish or for fish consumption, nutritional benefits should be weighed against the possibility of adverse effects. Studies published since the fifty-third meeting were considered at the present meeting.

2. BIOLOGICAL DATA

2.1 Biochemical aspects

2.1.1 Absorption, distribution and excretion

Details regarding the absorption and distribution of methylmercury in various species of experimental animals have been provided previously (Annex 1, reference 144). Briefly, methylmercury is effectively absorbed from the gastrointestinal tract and readily crosses both the blood–brain barrier and the placenta.

(a) Transfer from mother to offspring

Transport across the placenta

Methylmercury passes readily through the placenta to the fetus, with concentrations in the fetal blood and brain being generally greater than the corresponding maternal concentrations at parturition.

The distribution of methylmercury during early neurogenesis was examined in groups of two or three pregnant Sprague-Dawley rats that were treated subcutaneously with methylmercury hydroxide at doses of 5–22 mg/kg bw on day 11 of gestation. The animals were sacrificed on day 12, 13, or 14 of gestation. Total mercury reached maximum concentrations in most tissues/organs (including the embryo) within 48 h after dosing, except for the maternal blood and kidney, in which concentrations peaked by 24 h after dosing and then declined over the next 2 days. Concentrations in embryonic brain and body were similar on day 14 of gestation, indicating that there had been no functional impairment of transfer by the fetal blood–brain barrier. At 5 mg/kg bw, partitioning coefficients for mercury were generally similar to those found in previously-reviewed toxicokinetic studies at later gestational times. Concentrations of total and inorganic mercury in the embryo did not increase when the administered dose of mercury was >12 mg/kg bw. However, concentrations of inorganic mercury in the placenta did increase superlinearly with dose, indicating the possible presence of a sink preventing exposure of the embryo. In maternal liver, total mercury increased as inorganic mercury decreased with dose administered (Lewandowski et al., 2002).

Lactation

The transport of methylmercury from blood to milk is less efficient than transport across the blood–brain and blood–placenta barriers. Exposure of offspring to methylmercury via lactation is lower than exposure in utero, with rodent milk : blood ratios usually being 0.03–0.2 (Annex 1, reference 144).

Female Long-Evans rats were given drinking-water containing methylmercury chloride at a concentration of 0, 0.5 or 6 mg/l, 28 or 49 days before mating and during gestation and lactation until postnatal day 16. The duration of dosing prior to mating had no effect on concentrations of mercury in brain or blood of the off-spring. Measurements at birth and at weaning showed brain : blood ratios of 0.14 and 0.24, respectively. Concentrations of total mercury in blood and brain showed a linear relationship to cumulative exposure only after exposure was terminated after weaning. Exposure via lactation was minimal; brain concentrations decreased by 10–20-fold between birth and weaning (Newland & Reile, 1999).

(b) Clearance

Mercury is predominantly cleared in faeces and, to a lesser extent, in urine. In rat offspring, clearance is slow and increases with time as the ability to excrete bile develops, usually by 2–4 weeks after birth.

Groups of five or six male Wistar rats given boiled whale blubber in a single oral dose of 2 g/kg bw, equivalent to a dose of total mercury of 3.96 mg/kg bw (methylmercury, 0.047 mg/kg bw), excreted almost 95% of the administered dose in faeces and <0.05% in urine within 72 h after dosing. The maximum concentration of mercury in kidney was approximately 0.18 µg/g compared with only 0.07 µg/g in liver. Most of the mercury in liver was thought to exist as insoluble mercury selenide, with the pattern of absorption and distribution being similar to that of inorganic mercury (Endo et al., 2003).

(c) Pharmacokinetics

Groups of 10 female B6C3F1/HSD mice were given drinking-water containing methylmercury chloride at a concentration of 0, 1 or 3 mg/l, 4 weeks before mating (with CBA/J HSD males) and during mating and gestation. Offspring were exposed to the same doses as the mothers from birth until postnatal day 13 (perinatal groups), or for 26 months (lifetime groups). Concentrations of total mercury in blood and brain decreased substantially between birth (postnatal day 4) and weaning (postnatal day 21), and to a greater degree in the perinatal groups, owing to the diminished importance of lactation as a source of exposure and to the higher body weights of the pups. The ratio of total mercury in brain : blood at both doses in the perinatal groups averaged 1.64 at birth and 1.54 at weaning. Brain : blood ratios at both doses in the lifetime groups remained comparable at weaning and at 26 months, averaging 0.63. Between 14 and 26 months, concentrations of mercury in both blood and brain increased substantially in the group exposed to the higher dose. In the group exposed to the lower dose, concentrations of mercury in the blood remained constant while those in the brain increased slightly. The authors concluded that these results suggest a possible toxicokinetic interaction between ageing and level of exposure (Stern et al., 2001).

Young et al. (2001) developed a physiologically-based pharmacokinetics model for methylmercury, incorporating pharmacokinetic data from 12 animal species (mouse, hamster, rat, guinea-pig, cat, rabbit, monkey, sheep, pig, goat, cow, and human). The resulting model for humans was consistent with those for animals and was supported by data from human autopsies, with regard to fate and distribution of mercury and methylmercury. The model predicts high concentrations of inorganic mercury in the kidneys after the elimination of methylmercury from the blood.

2.1.2 Effects on enzymes and other biochemical parameters

(a) Cleavage of the carbon–mercury bond

Of the organic mercury chemicals (methyl, dimethyl, phenyl, mercuric acetate), methylmercury is considered to be chemically the most stable. Bacterial demethylation of methylmercury can take place in the intestinal tract, and involves the generation of both enzymatic and reactive oxygen species, resulting in decreased absorption of methylmercury (National Academy of Sciences/National Research Council, 2000).

The reaction by which methylmercury is demethylated was investigated using several reactive oxygen modulators in a model using rat liver slices in vitro. The rate of conversion of methylmercury to inorganic mercury in vitro is similar to experiments in rats in vivo, indicating that the model is suitable for reproducing the bio-transformation of methylmercury. Results with various reagents suggest that the demethylation of methylmercury may be aided by superoxide anions produced by the electron transfer system in the inner mitochondrial membrane (Yasutake & Hirayama, 2001).

(b) Complexes with thiol radicals

Methylmercury forms complexes with thiol ligands, including glutathione and L-cysteine, facilitating its extracellular transport and ability to bind to intracellular proteins, targeting sulfhydryl enzymes. Complexing with L-cysteine increases the concentration of methylmercury in brain, liver, and kidney while a decrease in renal gamma-glutamyl transpeptidase activity results in a decrease in renal methylmercury and increased urinary excretion.

(c) Interaction with selenium

Any potential effect of selenium on the toxicity and tissue deposition of methylmercury depends on dose, chemical form, and route of administration (National Academy of Sciences/National Research Council, 2000). Under certain experimental conditions, the simultaneous administration of equal doses of selenium and methylmercury is known to result in decreased toxicity of both compounds. The possible mechanisms for the protective role of selenium may involve the formation of a mercury–selenium complex (Agency for Toxic Substances and Disease Registry, 1999).

ICR mice were fed diets containing selenium at a low (<0.02 ppm), marginal (0.05 ppm) or adequate (0.4 ppm) concentration for 4 weeks before breeding, and subsequently were given methylmercury at a dose of 5 mg/kg bw on day 12 of gestation, or 3 mg/kg bw on days 12–14 of gestation. Decreased dietary selenium, both alone and in combination with treatment with methylmercury, caused a decrease in the activity of glutathione peroxidase in the neonatal brain at postnatal day 3. A reduction in glutathione peroxidase activity in the brain has been associated with deficits in neurobehavioural function (retarded development of walking ability, greater preference for warmer environments in a thermogrid) (Watanabe & Satoh, 1994). Neurobehavioural outcomes (reflex and motor development) were most affected in the group given the diet that was deficient in selenium together with the highest total dose of mercury (three doses of 3 mg/kg bw or 9 mg/kg bw total) of methylmercury (Watanabe et al., 1999b).

In a study with a similar experimental design (but no diets deficient in selenium), there was a decrease in the activity of 5’-iodothyronine deiodinase (5’-DI) and an increase in the activity of 5-iodothyronine deiodinase (5-DI) selenoenzymes in the fetal brain at the high dose, while there was an increase in the activity of 5-DI in the placenta at both high and low doses. The activity of glutathione peroxidase was inhibited in fetal brain and placenta, but not in maternal brain. Dietary intake of selenium did not influence concentrations of methylmercury in fetal brain or placenta, indicating that the effect of mercury on these selenoenzymes could not be explained by a reduction in the availability of selenium (Watanabe et al., 1999a).

Offspring from groups of 10 female Sprague-Dawley rats, who had been maintained on a diet supplemented with sodium selenite at a concentration of 1.3 mg/kg for 8 weeks before breeding and then treated by gavage with methylmercury at a dose of 0, 2 or 6 mg/kg bw on days 6–9 of gestation, were assessed for behavioural effects at age 2 months. At postnatal day 2, the pups whose mothers had been maintained on a diet containing sodium selenite had increased concentrations of blood mercury (1.4 and 2.8-fold at 2 and 6 mg/kg bw, respectively) compared with controls, but no increase in brain mercury (Fredriksson et al., 1993). Partial attenuation by selenite was seen in the behavioural alterations induced in the group receiving the highest dose of methylmercury. Although the selenium found in foods appears to have a protective effect in experimental animals, no association has been confirmed in humans (National Academy of Sciences/ National Research Council, 2000).

2.2 Toxicological studies

2.2.1 Acute toxicity

The LD50 of methylmercury in rodents treated orally is usually 10–40 mg/kg bw. Methylmercury is also considered to be corrosive at high doses (Annex 1, reference 144).

Alterations in various serum and urine biochemical parameters within 72 h after dosing were observed in male rats given mercury as a single oral dose (total mercury, 3.96 mg/kg bw; methylmercury, 0.047 mg/kg bw) in boiled whale blubber. The activity of serum lactate dehydrogenase was increased two-fold and there were significant increases in urinary volume, activity of N-acetyl-beta-D-glucosaminidase and albumin compared with the control group, which received total mercury in a single dose of 0.4 µg/kg bw (Endo et al., 2003).

2.2.2 Long-term studies of toxicity and carcinogenicity

Long-term exposure to methylmercury has induced renal tumours in mice, but only at doses at which significant nephropathy was also evident. No significant increase in any tumour type, including renal tumours, has been reported in rats (Annex 1, reference 144).

2.2.3 Reproductive toxicity

In mice and rats, methylmercury induces abortions, increases resorption and malformations, and reduces offspring viability.

(a) Single-generation studies

Mice

Groups of six pregnant CD-1 mice were given a single intraperitoneal dose of methylmercury chloride of either 0 or 10 mg/kg bw on day 9 of gestation, and litters were examined for teratological effects on day 17 of gestation. Co-exposure to methylmercury chloride and PK11195 (4 mg/kg), a ligand for the mitochondrial peripheral-type benzodiazepine receptor, reduced the overall incidence of fetal malformation by approximately 2.5-fold (47.0% versus 19.2%, respectively) (O’Hara et al., 2002). In embryos collected from dams 3 h after treatment with methylmercury chloride at 5 mg/kg bw on day 9 of gestation, expression of mitochondrial 16S rRNA was decreased approximately three-fold compared with controls, while co-treatment with PK11195 (4 mg/kg) prevented any decrease.

Rats

Groups of three to four male Wistar rats received methylmercury chloride by subcutaneous injection at a dose of 0 or 10 mg/kg bw per day for 8 days, and were sacrificed 3, 6, 9 or 14 days after the first dose. Sperm production was suppressed (by 27%), testicular (by 28%) and prostatic lobe weights (by 52–65%) were decreased, and plasma testosterone was reduced (by 13–29%) compared with control animals. The development of spermatocytes and spermatids was affected at stages VII–VIII and IX–XI, respectively. Loss of spermatids was reported to occur at stages I and XII–XIV. Methylmercury impaired spermatogenesis as a result of germ cell deletion via cell- and stage-specific apoptosis (Homma-Takeda et al., 2001).

(b) Developmental toxicity

(i) In vitro

A 2 h exposure of primary cultures of chick forebrain neurons to methylmercury at a sublethal concentration of 0.25 or 0.5 µmol/l (62.7 or 125.5 µg/l) inhibited axonal morphogenesis in vitro. A similar treatment with methylmercury at 1.0 µmol/l (251 µg/l) resulted in substantial cell death by day 3, while treatment with 0.1 µmol/l (25.1 µg/l) had little effect on axonal morphogenesis. This effect was thought to be associated with the developmental neurotoxicity of methylmercury (Heidemann et al., 2001).

The effect of methylmercury on the proliferation of neuroepithelial cells was studied in primary rat embryo cells of the central nervous sytem. At a concentration of 1–6 µmol/l (0.25–1.51 mg/l), methylmercury affected signalling pathways involved in regulation of the cell cycle, contributing to reduced cell proliferation and cell death. The authors suggested that inhibition of cell cycle progression might be linked to the neuronal hypoplasia observed in vivo after exposure to methylmercury during gestation and the enhanced susceptibility to methylmercury of the fetus relative to that of the adult (Faustman et al., 2002).

Prolonged exposure (6 days) to nanomolar concentrations of methylmercury (0.1–30 nmol/l or 0.025–7.53 µg/l) impaired cationic channel function in differentiating pheochromocytoma cells. The lack of morphological changes suggests that ion channels are sensitive targets for neurotoxicity caused by methylmercury (Shafer et al., 2002).

Developing cortical neurons from rat pups (aged 4 weeks) treated in utero and during lactation with methylmercury at a maternal dose of 0.375 mg/kg bw per day, from mating to end of lactation, were found by microelectrophysiological assessment to be more excitable in vitro (Vilagi et al., 2000).

Methylmercury can facilitate the formation of reactive oxygen species and has been shown to decrease intracellular concentrations of glutathione. Adequate uptake of cysteine is essential for maintaining proper intracellular concentrations of glutathione. Methylmercury at a concentration of 5 or 10 µmol/l (1.25 or 2.51 mg/l) inhibited cortical uptake of cysteine in vitro via inhibition of specific cysteine transport systems in astrocytes isolated from newborn Sprague-Dawley rats, but no effect was seen in hippocampal neurons (Shanker et al., 2001).

(ii) In utero

Groups of 8–11 pregnant mice of each of three strains (BALB/c, C57BL/6J, C57BL/6Cr) received methylmercury chloride at a dose of 3 mg/kg bw per day by gavage on days 12–14 of gestation. Strain-specific neurobehavioural changes were investigated in male offspring only. In comparison to their respective controls, BALB/c and C57BL/6Cr mice, but not C57BL/6J, showed decreased open-field activity. BALB/c mice also showed a change in emotional status in open field activity and a disturbance in nocturnal rhythm of spontaneous activity in the home cage. In the Morris water maze, prolonged latency in locating a submerged platform was observed in strains C57BL/6J and C57BL/6Cr, but not in BALB/c (Kim et al., 2000b).

Methylmercury chloride was administered by gavage to groups of 10 pregnant female CD1 mice in a single dose of 12.5 mg/kg bw on day 10 of gestation and the animals were sacrificed on day 18 of gestation. Maternal toxicity was not significant. Fetuses examined for malformations on day 18 of gestation showed cleft palate and decreased fetal weight. Other groups received binary and tertiary mixtures of lead and arsenic. At the doses tested (lead nitrate, 25 mg/kg bw; sodium arsenite, 6.0 mg/kg bw), the interactive effects of lead and arsenic on developmental toxicity induced by methylmercury were not greater than additive (Belles et al., 2002).

Groups of 20 pregnant Sprague-Dawley rats were given methylmercury chloride in a single oral dose of 0 or 8 mg/kg bw on day 8 of gestation. Mothers were asymptomatic, while the offspring showed deficits in learning and memory (passive avoidance, object recognition, water maze) on day 60. Methylmercury also caused altered gene expression of N-methyl-D-aspartate (NMDA) receptors in the hippocampus (Baraldi et al., 2002). In the same experiment, the hippocampus of off-spring at postnatal days 21 and 60 showed an increase in tryptophan, while concentrations of anthranilic acid were reduced, and concentrations of quinolinic acid were increased at postnatal day 21 but not at postnatal day 60. The authors suggest that methylmercury-induced changes in the kynurenine pathway could contribute to alterations in brain development (Zanoli et al., 2001).

Groups of five pregnant Wistar rats were given methylmercury chloride in a single oral dose of 20 mg/kg bw on day 11, 13, 16, or 21 of gestation. One hour later, 5-bromo-2-deoxyuridine (BrDU) was given intraperitoneally at a dose of 20 mg/kg. The offspring had no symptoms of neurological impairment (observations not stated). At postnatal day 21, the central nervous system of offspring exposed during the neurogenesis (days 16–21 of gestation) showed that labelled cells were abnormally located in cortical layers II–IV when compared with the normal inside-out pattern of neuronal migration. The cells from day 16 of gestation appeared to have over-migrated while the cells from day 21 of gestation appeared to have ceased migration before reaching their normal position. Abnormal cell migration was not observed on days 11 or 13 of gestation. No cytoarchitectural abnormalities were apparent. The authors concluded that the results indicate a disruption in neuronal migration caused by the administration of methylmercury during the neurogenic period (Kakita et al., 2002).

Groups of six female Wistar rats were given methylmercury chloride at an oral dose of 0, 1, 2, or 3 mg/kg bw per day for 5 days, and 0, 1 or 2 mg/kg bw per day for 12 days before pregnancy, and up to day 19 of gestation (total periods of exposure, 24 or 31 days). Body-weight gain was decreased, and unsteadyness and flexion of hindlimbs appeared on days 15 and 16 of gestation in the group receiving the highest dose for 24 days. A slight decrease in body-weight gain and slow movement were reported in the group receiving the highest dose for 31 days. On day 22 of gestation, fetal brains showed neuronal degeneration in the brain stem, cingulate cortex, thalamus, and the cerebral basal area, including the hypothalamus. From dams monitored after parturition, only offspring from the group receiving a dose of 1 mg/kg bw per day for 24 days survived to postnatal day 3 and developed, albeit with mild lesions of the brain. The distribution of brain lesions in rats exposed in utero was different from those in rats exposed postnatally and in adulthood (Kakita et al., 2000a).

Groups of six female Wistar rats were given methylmercury chloride by gavage at a dose of 1 mg/kg bw per day for 5 days before mating, and during days 1–19 of gestation. On postnatal day 2, offspring from treated dams were cross-fostered with untreated females who had delivered litters on the same day. On postnatal days 1 and 3, degenerative neurons were observed in the brain stem and limbic system, including hippocampus and amygdala, of the offspring. On postnatal days 7 and 14, degenerative neurons were indiscernible but reactive astrocytosis remained in the brain stem. By postnatal days 70 and 180, brains appeared normal; however, at age 6 months, rats showed a learning disability in the passive avoidance response. Also, fewer neurons were present in the hippocampus and amygdala (Kakita et al., 2000b).

Fourteen offspring of female rats exposed to drinking-water containing methylmercury chloride during mating, gestation, and lactation (0 or 0.375 mg/kg bw per day) were tested for the maintenance and spread of induced (3-aminopy-ridine) epileptiform activity. Seizure duration was increased in offspring at age 3 months, while epileptic discharges were reduced in pups at age 1 month (28 days) (Szasz et al., 2000).

Leptomeningeal glioneurol heteropia (LGH), a developmental malformation of the brain was induced experimentally for the first time in the offspring of pregnant Wistar rats treated orally with methylmercury chloride at a dose of with 20 mg/kg bw on day 8, 11, 13, 16, 18, or 21 of gestation. The brains of offspring were examined on postnatal day 7 and 28. LGH was almost exclusively restricted to rats treated on day 13 of gestation, at the early stage of neurogenesis. The authors noted that LGH has been reported in Iraqi children exposed to methylmercury during the second and third months of gestation, as a result of maternal consumption of bread and cereals treated with a fungicide containing alkyl-mercury (Kakita et al., 2001).

Groups of three offspring of female squirrel monkeys with blood methylmercury concentrations of 0.7–0.9 mg/l showed reduced sensitivity to changes in the source of reinforcement, indicating learning impairment at age 5–6 years. Maternal dosing (by gavage) to obtain the required blood mercury concentration was initiated between weeks 11–14.5 of gestation, while offspring were removed from their mothers at birth to prevent lacational transfer (Newland et al., 1994).

(iii) Exposure in utero and postnatal exposure

Rats

Groups of seven to nine pregnant Long-Evans rats were given drinking-water containing methylmercury at a concentration of 0, 0.5 mg/l (equivalent to 40–50 µg/kg bw per day) or 6.4 mg/l (equivalent to 500–700 µg/kg bw per day) for 4 or 7 weeks before mating, and up to postnatal day 16. Brain concentrations of mercury in offspring of rats receiving low and high doses were 0.5 and 9.1 µg/g at birth and 0.04 and 0.52 µg/g at weaning, respectively. The F1 adults were trained to respond under differential reinforcement of high rate (DRH) 9 : 4 extinction schedule of food reinforcement (DRH 9 : 4; 9 responses within 4 s delivers the food reinforcement). Behaviour was monitored until rats were aged >912 days. All rats displayed increases in delayed responses with age. While most control animals showed a 50% decline in the reinforcement rate by age 950 days, rats receiving the lower and higher doses reached this level of decline at about age 800 and 500 days, respectively. Birth weight and rates of growth and survival were normal (Newland & Rasmussen, 2000).

In the same experiment, various drugs (d-amphetamine, 0.3–10 mg/kg bw; scopolamine, 0.1–3.0 mg/kg bw; pentobarbital, 1–20 mg/kg bw; haloperidol, 0.03–1.00 mg/kg bw; and dizocilpine, 0.01–0.30 mg/kg bw) were administered to females aged 13–15 months, for 7 months. The groups receiving methylmercury at a dose of 0.5 or 6.4 mg/kg were sensitive to behavioural disruption by amphetamine under the DRH 9 : 4 schedule, while the group receiving 6.4 mg/kg was relatively insensitive to pentobarbital. No interaction between methylmercury and the other drugs was noted. The results suggested an involvement of catecholaminergic activity and gamma-aminobutyric acid (GABA) systems in neurotoxicity caused by methylmercury (Rasmussen & Newland, 2001).

Groups of 10 female Wistar rats were given a diet containing methylmercury at a concentration of 5 mg/kg (approximately equivalent to 0.32 mg/kg bw) before mating and during gestation and lactation. Offspring were provided with the same diet after weaning. All offspring matured, normal body-weight gain included, without physical signs of poisoning. Blood and brain mercury concentrations in the newborn were 33 and 4.5 ng/g, respectively, and approximately 1.5-fold higher than those of their mothers. Exposure through lactation was less than during gestation. Eight offspring at age 5 and 6 weeks showed a significant deficit in motor coordination in the rotarod test and a learning disability in the test for passive avoidance response. Histopathological abnormalities (focal cerebellar dysplasia and heterotopic location of Purkinje and granule cells) were observed (Sakamoto et al., 2002).

Groups of eight pregnant rats were fed a diet containing a normal (20%) or low (3.5%) percentage of protein during gestation and lactation. Weaned pups were given the same diets as their mothers and given methylmercury chloride by gavage at a daily dose of 0 or 7.5 mg/kg bw for 10 days, starting on postnatal day 21 or 60. Offspring (both weanings and young adults) treated with methylmercury on the low-protein diet developed hind-limb crossing earlier and more severely than rats on a diet containing sufficient protein. For example, 100% of rat pups on the low-protein diet exhibited hind-limb crossing 1 day after the last dose of methylmercury compared with 5 days for the pups fed the diet containing sufficient protein. Rats fed the low-protein diet accumulated more mercury in different brain regions. The rate of protein synthesis in liver, kidney and various brain regions was reduced in animals exposed to methylmercury and receiving either a normal or a low-protein diet, although the reduction was greater in the latter group (Chakrabarti & Bai, 2000).

(iv) Exposure after parturition

Rats

Groups of five Sprague-Dawley rat pups received a subcutaneous dose of methylmercury chloride of 0 or 7 mg/kg bw per day on alternate days from post-natal day 3 to 13 (six doses) and were killed on postnatal day 15, 30, or 60. Compared with controls, body weights were decreased by 20% in the rats treated with methylmercury at postnatal day 15 and 30, but were normal at postnatal day 60. No additional signs of toxicity were observed. Neural cell adhesion molecule (NCAM) protein expression in cerebellar homogenates was elevated on postnatal day 30 but returned to control levels by postnatal day 60, while Golgi sialyltransferase activity was reduced on postnatal day 15 but not at later times. The precise temporal and spatial expression of NCAM is considered necessary for normal brain morphogenesis (Dey et al., 1999).

In the same experiment, cerebellar Golgi fractions obtained from untreated animals at postnatal day 15, 30, and 60 were incubated in vitro with methylmercury at a concentration of 0, 1, 2.5, or 7.5 µg/l for 2 h. While NCAM sialyltransferase activity decreases naturally as the developmental age of the rat increases, increasing concentrations of methylmercury further depressed sialyltransferase activity in fractions from postnatal day 15, but not later time-points. The authors concluded that alterations in expression of NCAM molecules during brain development might contribute to neurotoxicity caused by methylmercury (Dey et al., 1999).

The involvement of NMDA receptors in neurotoxicity caused by methylmercury was investigated in groups of seven Wistar rats, aged 2, 16, or 60 days, that were given condensed milk containing methylmercury at an oral dose of 0 or 10 mg/kg bw per day for 7 days and sacrificed 4 days after the last dose. The most severe neuronal degeneration was observed in the occipital cortex on postnatal day 16. Co-exposure to methylmercury and MK-801 (0.1 mg/kg), a non-competitive antagonist of NMDA, resulted in substantially less neuronal damage. Treatment of rats with methylmercury on postnatal day 16 resulted in an accumulation of nitrotyrosine in the occipital cortex brain region compared with controls (not detected); co-treatment with MK-801 significantly reduced the amount detected. Generation of oxygen radicals through stimulation of NMDA receptors may be an important factor in neurotoxicity caused by methylmercury in developing cortical neurons (Miyamoto et al., 2001).

Non-human primates

Five macaque monkeys were given methylmercury at an oral dose of 0.05 mg/kg bw per day from birth to age 7 years and were examined for visual function at age 3–4 and 20 years. Blood concentrations peaked at 1.2 to 1.4 mg/l and declined after weaning from infant formula to 0.7 to 1.0 mg/l. Two monkeys exhibited a slight constriction of visual fields at age 18–20 years, an effect that was not present at age 3–4 years. A treatment-related decline in spatial contrast sensitivity was observed in one monkey during young adulthood but not in middle age. There was a lack of evidence for a combined treatment-age-related decrement in visual function (Rice & Hayward, 1999).

2.2.4 Special studies

(a) Neurotoxicity

Typical symptoms of methylmercury-induced neurotoxicity in rodents include degenerative nerve changes (peripheral and central), ataxia, coordination disorders and decreases in motor activity.

Mice

Groups of 10 mice were given methylmercury chloride at a concentration of 40 mg/g of diet, with or without melatonin at a concentration of 20 mg/ml in drinking-water, for 5 weeks. Greater numbers of animals treated with methylmercury alone showed signs of neurotoxicity (abnormal righting reflex, staggering gait, fallen posture) and a higher rate of mortality after 35 days of treatment than the group receiving methylmercury plus melatonin. It was suggested that the anti-oxidative capacity of melatonin reduces the toxicity of methylmercury (Kim et al., 2000a).

Groups of eight adult male mice (strain not specified) were treated orally with methylmercury at a dose of 0, 0.2, 2, or 10 mg/kg bw per day for 7 days. All doses caused elevation of the hearing threshold. All mice receiving the highest dose died. Eleven weeks after treatment, hearing loss persisted in the group receiving a dose of 2 mg/kg bw, while effects were reversed in the group receiving the lowest dose. Results also showed an increase in the production of nitric oxide and significant inhibition of the activity of Na+/K+-ATPase within the brainstem, indicating a possible pathway for the altered brainstem response induced by exposure to methylmercury (Chuu et al., 2001a).

Rats

Methylmercury, with and without various drugs affecting dopamine uptake, was applied locally in the striatum of female Sprague-Dawley rats in order to investigate the mechanism of dopamine release. A dose of methylmercury of 400 µmol/l (a 60 min perfusion at 2 µl/min) induced the extracellular release of dopamine via a mechanism described as transporter-dependent but calcium- and vesicular-independent (Faro et al., 2002).

Groups of adult male Sprague-Dawley rats were given methylmercury intraperitoneally at a dose of 0 or 2 mg/kg bw per day for 13 days. Treatment with methylmercury significantly decreased body weight and produced hypoactivity. Treated rats also showed an impaired active-avoidance response immediately and 33 weeks after treatment. Na+/K+-ATPase activity in the cerebral cortex was significantly inhibited immediately and 8 weeks after treatment (Chuu et al., 2001b).

Groups of six to eight adult female Wistar rats were given a solution of methylmercury at 0, 10 or 100 µmol/l through a microdialysis probe in the frontal cortex of awake animals at a perfusion rate of 2 µl/min for 90 min. The perfusion caused significant elevations in concentrations of extracellular glutamate (Juarez et al., 2002).

Non-human primates

Neurotoxicological effects induced by methylmercury in non-human primates resemble those seen in humans and support the concept that the developing fetus is more sensitive than adults. Typical effects involve neurological, sensory and motor functions with no-observed-adverse-effect levels/lowest-observed adverse-effect levels (NOAELs/LOAELs) in the range of 10–50 µg/kg bw per day (Annex 1, reference 144; National Academy of Sciences/National Research Council, 2000).

Groups of two adult male marmosets were given drinking-water containing methylmercury at a concentration of 0 or 5 µg/ml for 24–90 days. An additional group was treated for 21 days and then maintained for 2.5 years without exposure. By day 70–90 of treatment (blood mercury concentration, 8–10 µg/ml), the animals showed physical signs of intoxication (weight loss, ataxia) as well as, at necropsy, severe neuronal loss and spongy degeneration. In the animals treated for 24 and 31 days without clinical signs of poisoning (blood mercury concentration, 5–8 µg/ml), mild neuronal loss was evident along with gliosis of the calcarine cortex. Axonal degeneration of the sciatic nerve was also seen in both of these treatment groups. Brain (cerebrum, cerebellum) mercury concentrations ranged from approximately 13 to 30 µg/g). After acute poisoning with methylmercury, white matter oedema may compress sulcal arteries and produce focal vascular insufficiency in the calcarine cortex (Eto et al., 2001, 2002).

(b) Renal and hepatic toxicity

In long-term experiments, renal damage in male mice and rats usually occurs at lower doses of methylmercury and with greater severity than in females.

(c) Immunomodulation

Groups of 8–11 female mice of strains SJL/N, A.SW, B10.S, BALB/C, DBA/2, A.TL, or B10.TL were given subcutaneous injections of methylmercury chloride at a dose of 0 or 1 mg/kg bw every third day for 27 days. Methylmercury induced systemic autoimmunity, manifest as antinucleolar antibodies (ANoA) targeting nucleolar fibrillarin (AFA), and anti-ssDNA (single stranded DNA) and anti-DNP (deoxyribonucleoprotein) antibodies, but the effects on the immune system differed qualitatively and quantitatively from those observed with inorganic mercury (Hultman & Hansson-Georgiadis, 1999).

In groups of five to nine female CBA/J mice infected with Toxoplasma gondii, methylmercury at a single oral dose of 20 mg/kg bw increased the mean numbers of brain cysts by four-fold compared with controls. Cell viability in both the spleen and thymus was decreased (60% compared with controls), but to a similar extent for either the separate or combined treatments (King et al., 2003).

Female Sprague-Dawley rats were given drinking-water containing methylmercury at a concentration of 0, 2.5 or 10 µg/l for 16 days (average dose, 0, 0.45 and 1.8 mg/kg bw per day, respectively). Muscarinic cholinergic receptors (mAChRs) were assessed in various brain regions and in splenic lymphocytes either immediately or 2 weeks after cessation of treatment. The density of MAChRs (as assessed by quinuclidinyl binding) was increased, without affecting receptor affinity, in the hippocampus and cerebellum, suggesting upregulation of mAChRs, but only 14 days after exposure to methylmercury had ended. No effects were observed in the cerebral cortex. There was a marked increase in the density of mAChRs in splenic lymphocytes both immediately and 14 days after exposure to methylmercury (Coccini et al., 2000).

Groups of three to five male Wistar rats were given methylmercury chloride orally at a dose of 0 or 10 mg/kg bw per day for 5 days and sacrificed 8, 12, and 15 days after treatment, while another group was given methylmercury at a single oral dose of 20 mg/kg bw and sacrificed 24 and 72 h after exposure. Peritoneal mast cells were isolated from all rats at the time of sacrifice. Compared with controls, methylmercury modified mast cell function in vitro, as observed by the inhibition of histamine release after stimulation with compound 48/80 (0.125 µg/ml) or with calcium ionophore A23187 (0.125 µg/ml). With the latter chemical, methylmercury progressively decreased the stimulation of histamine release for the three time-points (Graevskaya et al., 2003).

(d) Other effects

In mitotic Chinese hamster fibroblast cells, methylmercury chloride at a concentration of as low as 2.5 µmol/l (0.62 mg/l) affected centrosome integrity in vitro by inducing multiple foci of tubulin and producing multipolar spindles and multi-nucleated cells, while inorganic mercury did not (Ochi, 2002).

The time sequence of morpho-functional changes induced by methylmercury hydroxide at concentrations of 10-8–10-5mol/l in C6 rat glioma cells revealed initially measurable reactive oxygen species, followed by oxidative DNA damage and mitochondrial depolarization, leading to apoptosis (Belletti et al., 2002). The lowest concentration of methylmercury hydroxide that effectively generated reactive oxygen species and 8-hydroxy-2’-deoxyguanosine DNA adducts was 10-7mol/l (25 µg/l).

2.3 Observations in humans

2.3.1 Study in the Seychelles

(a) Pilot cohort neurodevelopmental evaluations at age 8 years

In the Seychelles child development study, a subset (11%, 87 infants) of the cohort used as a pilot sample for the neurodevelopmental evaluations of infants aged 5.5 years was used as a pilot cohort for the evaluations at age 8 years. Selection appears to have been based on birth date only (children who were aged 8 years ±6 months between 1 June 1997 and 30 October 1997). Median concentration of mercury in maternal hair in this subset was 7.8 mg/kg (range, 0.6–35.4; interquartile range, 12.1 mg/kg; 41 children had a concentration of >9 mg/kg). The battery of tests was constructed to include some of the same tests given to the Faroese children at age 7 years: several subtests of the Wechsler intelligence scale for children-III (WISC-III), the California verbal learning test, the Boston naming test, the developmental test of visual-motor integration, the design memory subtest of the wide-range assessment of memory and learning, the grooved peg-board, the trail-making test, and the finger-tapping test. The statistical methods used were modelled on those applied to data collected at previous evaluations. A significant interaction between prenatal exposure to methylmercury and sex was seen on the Boston naming test, such that in males, the number of correct responses increased with increasing concentration of methylmercury. On the grooved pegboard test, boys’ performance for both the dominant and nondominant hands improved with increasing prenatal exposure to methylmercury. On the developmental test of visual-motor integration, boys’ scores also improved with increasing prenatal exposure to methylmercury. In girls, for the preferred hand, time to complete the grooved pegboard increased with increasing prenatal exposure to methylmercury. Two negative associations were found between postnatal exposure to methylmercury (hair mercury concentration at age 8 years), for short-delay and long-delay scores on the California verbal learning test (Davidson et al., 2000).

(b) Neurodevelopmental evaluations of the full cohort at age 8 years

At age 8 years, 643 children underwent an additional evaluation. This number represents 90% of the 717 children (of the original cohort of 779) who were considered eligible. As in the pilot study for the evaluations of 8-year-old children, the test battery included both global and domain-specific tests. In addition to the tests given in the pilot study, the following tests were administered: the letter-word identification and applied problems subtests of the Woodcock-Johnson tests of achievement, several subtests of the Bruininks-Oseretsky test of motor proficiency, a test of haptic matching, the Connors continuous performance test, and the Connors teacher rating scale. The mean concentration of mercury in hair at the time the children were tested was 6.1 mg/kg (standard deviation (SD), 3.5). A total of 21 end-points were examined. Children with higher prenatal exposure to methylmercury took significantly longer to complete the grooved pegboard with the nondominant hand, and had significantly better scores on the hyperactivity index of the Connors teacher rating scale. The authors commented that the distribution of p values for the 21 end-points was as would be expected under the null hypothesis of no association between prenatal exposure to methylmercury and child outcome. Analyses involving postnatal exposure to methylmercury and child outcome were not reported, although it was noted that significant adverse associations in females were found in "a few tests" and that additional analysis of these findings was ongoing. The authors conclude that they continue to find: ". . . no detectable adverse effects in a population consuming a wide variety of ocean fish" (Myers et al., 2003).

Several additional analyses of the results of follow-up examinations of children aged 5.5 years have been reported. These have consisted of efforts to respond to issues raised by peer reviewers and to explore alternative explanations for the apparent discrepancies between the findings of this study and those of the study in the Faroe Islands. These studies are summarized below.

In the analyses reported by Davidson et al. (1998), both prenatal exposure to methylmercury (as reflected in concentration of methylmercury in maternal hair during pregnancy) and postnatal exposure to methylmercury (concentration of mercury in the child’s hair, at age 5.5 years) were included in the regression models. Although these two indices of exposure to methylmercury were only modestly correlated (correlation coefficient, 0.15), a concern was raised that including both indices in the same model might attenuate any association either one might have with neurodevelopment. Therefore, separate regressions on the six primary end-points of the evaluation at age 5.5 years were carried out, considering each exposure index separately. The results were essentially unchanged from those initially reported (Cox et al., 1999).

Because three different examiners were used to collect data on neurodevelopment at age 5.5 years, regressions adjusting for examiner were run. Again the results were unchanged (Davidson et al., 2001).

In the original study reported by Davidson et al. (1998), analyses were conducted on children’s age-standardized scores. In the Faroe Islands, in contrast, many analyses were done on children’s raw neurodevelopmental scores, adjusting for age at examination. Therefore, an evaluation was performed to determine whether using the latter approach changed the results obtained: it did not (Davidson et al., 2001).

Generalized additive models were applied to the data acquired during the examinations of 5-year-old children in order to identify possible nonlinearities in the association between methylmercury and the six primary end-points. Unlike the regression approach applied in the analyses reported by Davidson et al. (1998), generalized additive models make no assumption about the functional form of the association. Minor nonlinearities were identified for three of the six end-points. On the preschool language scale (for which a higher score is better), children’s scores declined 0.8 points between concentrations of mercury in maternal hair of 0 and 10 mg/kg, then increased 1.3 points for concentrations of >10 mg/kg. Total score on the child behaviour checklist (for which a higher score is worse) increased by 1.0 point between concentrations of mercury in maternal hair of 0 and 15mg/kg, and declined 4 points with an increase in concentration of 15 to 20 mg/kg. For the general cognitive index on the McCarthy scales of children’s abilities, the score increased 1.8 points for concentrations of mercury in the child’s hair of up to 10 mg/kg, then declined 3.2 points for concentrations of between 10 and 20 mg/kg. Thus, overall there was no clear evidence for consistent adverse effects across the entire range of exposures represented in this sample. Two of the three trends suggested a beneficial effect of increasing prenatal exposure to mercury (Axtell et al., 2000).

The fact that in the study in the Faroe Islands the neurodevelopmental tests administered were "domain-specific" while they were "global" in the study in the Seychelles has been cited by some as a possible reason for their discordant findings. To address this issue, a re-analysis was performed on scores obtained on the McCarthy scales of children’s abilities, a global test of ability, applying a neuro-psychologically-based conceptual framework to recombine subscale scores into discrete domains: attention, executive functions, expressive language, receptive language, nonverbal memory, verbal memory, visual-spatial organization, visual-motor organization, and gross motor skills. As in the original analyses, however, no evidence of any adverse associations between domain scores and prenatal exposure to methylmercury was found. For many domains, the mean performance was best in the stratum with the highest prenatal exposure to methylmercury. Higher concentrations of hair mercury in children aged 5.5 years were significantly associated with higher memory scores (Palumbo et al., 2000).

Analyses of data from the evaluations of neurodevelopment at age 6.5, 19, or 29 months (0.5, 1.6, or 2.4 years, respectively) were conducted to determine whether an adverse association between prenatal exposure to methylmercury and children’s scores could be identified among subgroups of the study sample. Specifically, potential effect modification by sex, birth weight, maternal age, family income, intelligence of the caregiver, and the amount of stimulation available in early home environment were evaluated. Although a few significant interactions were noted, the slopes of the dose—effect relationship varied across strata of covariates in ways that are hard to interpret (Davidson et al., 1999).

Analyses of the children’s scores on the separate scales of the child behaviour checklist, a parent rating scale, were reported. No consistent trends related to mercury in maternal hair or in the child’s hair were found. The T-scores of the Seychellois children were consistently higher than those in the United States standardization sample, but scores showed the expected associations with other covariates (Myers et al., 2000).

2.3.2 Study in the Faroe Islands

Neuropsychological, neurophysiological, and neurological evaluations of the Faroe Islands study cohort at age 14 years were completed in mid-2001, but no findings have yet been published. However, Grandjean and colleagues have recently reported additional analyses of the data collected when the children were aged 7 years. In an effort to identify age-dependent variation in susceptibility, they compared the associations between outcomes at age 7 years and five sets of bio-markers of exposure to methylmercury: concentrations of mercury in cord blood, maternal hair, child’s hair at age 1 and 7 years, and child’s blood at age 7 years (Grandjean et al., 1999). As expected, given the previously reported findings, the regression coefficients for mercury in cord blood were generally larger, and associated with lower p values, than were those for the other biomarkers, particularly on tests assessing language, attention, and memory, and less so for visuospatial memory and motor function. For finger tapping (a test of motor function), mercury in maternal hair was the better predictor, and was presumed to reflect exposure to mercury in the second trimester of pregnancy. Higher concentrations of mercury in hair of children at age 1 year were associated with slower finger tapping and with slower reaction time. Higher concentrations of mercury in children’s hair and in blood at age 7 years were associated with worse recall on the Bender Gestalt test of visual-motor function. This pattern led the authors to conclude that, in general, the period of greatest susceptibility to neurotoxicity caused by methylmercury occurs during late gestation, but that, to some extent, it probably differs for different brain functions (Grandjean et al., 1999).

The possibility of spikes in exposure to methylmercury of pregnant Faroese women who consume meals of whale meat has been cited as a possible explanation for discrepancies in the findings of studies in the Seychelles and the Faroe Islands. To evaluate this hypothesis, samples of maternal hair collected at parturition were used to compare the concentration of mercury in full-length hair segments (reflecting average exposure to mercury during pregnancy) to the concentration of mercury in the segment closest to the scale (reflecting exposure closer to parturition). Although the correlation between concentrations of mercury in these two samples was high (0.93), for 10% of the children, the smaller concentration was less than 60% of the higher concentration, suggesting that exposure had been variable between early and late pregnancy. Excluding these children from analyses of the associations between mercury in cord blood and neuropsychological outcomes at age 7 years had little impact on the regression coefficients estimated for mercury in cord blood. In fact, the coefficients increased from 3% to 43% after exclusion of children with greatest variation in exposure to methylmercury in utero. The authors concluded that the associations seen in the study undertaken in the Faroe Islands were not caused by peaks in exposure to methylmercury resulting from the periodic consumption of whale meals (Grandjean et al., 2003b).

Considerable concern has been expressed that adverse outcomes associated with prenatal exposure to methylmercury in the study in the Faroe Islands might be caused by residual confounding by high exposure to polychlorinated biphenyls (PCB) in this population. The investigators addressed this issue by stratifying analyses by PCB tertile. Those analyses provided little evidence that the magnitude of the association between methylmercury and neuropsychological outcomes differed in the three PCB strata. In fact, the association appeared to be somewhat stronger in the tertile with lowest exposure to PCBs (Budtz-Jorgensen et al., 1999b).

In additional analyses, the four neuropsychological tests that measure parameters most strongly related to exposure to PCBs in this cohort were examined, stratifying by methylmercury tertile. For all four outcomes, the strongest effects of PCB were seen among children in the highest methylmercury tertile. The confidence intervals were quite wide, however, so the cross-product interaction terms did not reach statistical significance for any outcome. The authors concluded that these analyses do, however, suggest that increased exposure to methylmercury might augment the neurotoxicity of PCB, but not the converse (Grandjean et al., 2001b).

In the Oswego newborn and infant development project, it was reported that in 212 consumers of fish from Lake Ontario, for whom maternal concentrations of mercury in hair during pregnancy averaged 0.5 mg/kg, exposure to methylmercury was not significantly associated with children’s scores on the McCarthy scales of children’s abilities at age 3.2 years, except among women with detectable levels of PCBs. However, this apparent enhancement of the neurotoxicity of methylmercury at higher concentrations of PCB was not apparent when the children were re-tested at age 4.5 years. The authors observed that the highest quartile of exposure to PCB in the Oswego study was still lower than the lowest quartile of exposure to PCB in the Faroe Islands cohort, suggesting that it is important to bear in mind that the effects of methylmercury were observed in a cohort that, by comparison with most other cohorts, has high levels of exposure to PCB (Stewart et al., 2003).

The hypothesis that visual dysfunction associated with exposure to mercury was responsible for the deficits noted in children aged >7 years, in scores for neuropsychological tests, some of which depend on the processing of visual information, was reconsidered. Although the score for a test of visual contrast sensitivity (functional acuity contrast test) was associated with children’s scores on several other tests, visual contrast sensitivity did not operate as a significant confounder of the association between mercury in cord blood and these tests, as adjustment had little appreciable impact on the regression coefficients for mercury in cord blood (Grandjean et al., 2001a).

A second, smaller cohort was recruited in the Faroe Islands. This cohort included 182 infants, 64% of consecutive spontaneous full-term births during 1 year, using a catchment area that included villages with greatest access to fish and whales. Concentrations of mercury were measured in maternal hair (mean, 4.08 mg/kg; range, 0.36–16.3), cord blood (mean, 20.4 µg/l; range, 1.90–102), and cord serum (mean, 2.54 µg/l; range, 0.70–8.74). In addition, 18 pesticides or pesticide metabolites and 28 PCB congeners (lipid-adjusted) were measured in maternal serum and breast milk, selenium was measured in cord blood, and fatty acids (arachidonic acid, eicopentanoic acid (EPA), docosahexaenoic acid (DHA), total omega-3 fatty acids) were measured in cord serum. The primary outcome, measured at age 2 weeks, was the neurological optimality score (NOS). This test assesses an infant’s functional abilities, reflexes, and muscle tone. In addition, several indices of thyroid function were measured in maternal and cord serum. A significant inverse relationship was found between NOS score and mercury in cord blood. A ten-fold increase in concentration of mercury was associated with a deficit in NOS score that was equivalent to a 3-week reduction in gestational age. Adjustments for total PCBs or fatty acid concentrations did not appreciably affect the results, and selenium did not appear to modify the effect of methylmercury (Steuerwald et al., 2000).

The prenatal and postnatal growth of this second cohort was examined in relation to fatty acid intake and exposures to environmental contaminants. It was found that increased intake of fatty acids (particularly DHA) prolongs length of gestation, but does not result in higher birth weight. Furthermore, birth weight adjusted for gestational age appeared to decrease at higher fatty acid intakes (of EPA, in particular), an effect that was independent of concentrations of methylmercury and PCBs. Mercury in cord blood, in particular, was not significantly associated with length of gestation, birth weight, or placental weight (Grandjean et al., 2001b).

The postnatal growth of children in the second cohort was assessed in relation to their prenatal and lactational exposures to PCBs and methylmercury, in an attempt to explain the "weanling’s dilemma", that is, that the growth of infants who are breastfed for longer than a few months is attenuated. Body weight and standing height were measured at age 1.5 and 3.5 years. The slower postnatal growth of children who were exclusively breastfed was confirmed. In multiple regression analyses, after adjusting for PCBs and several other factors, higher concentration of mercury in cord blood was negatively associated with both height and weight at age 1.5 years. A doubling of concentration of mercury in cord blood was associated with a decrease of 0.19 kg in weight and a decrease of 0.26 cm in height. Similar associations were seen at age 3.5 years, although they were not significant if adjustment was made for height and weight at age 1.5 years. The negative relationship between breastfeeding and weight was reduced to non-significance after adjustment for lactational exposure to mercury. Lipid-adjusted concentration of PCBs in serum at age 4.5 years was negatively associated with weight and height at age 4.5 years. The authors concluded that contaminants of seafood in the maternal diet might adversely impact the infants’ postnatal growth, and that this might help to account for the decreased growth observed in children who are exclusively breastfed (Grandjean et al., 2003a).

2.3.3 Other studies in children with a high intake of fish

A small group of Inuit children who had been exposed to relatively high levels of methylmercury were recruited. Twenty-one out of 40 children aged 6–12 years for whom concentrations of mercury in cord blood had been measured were compared with 22 Inuit children aged 7–12 years residing in the same community. Hair samples from all children and their mothers were also analysed for mercury. The mean concentration of mercury in cord blood of the 21 children for whom it was available was 181 µg/l (range, 28–777); the mean concentration of mercury in maternal hair for the 43 children was 5.5 mg/kg, ranging up to 18.4 mg/kg; and the mean concentration of mercury in maternal hair was 15.5 mg/kg, ranging up to 32.9 mg/kg. The only significant association found was a positive association between concentration of mercury in maternal hair and error score in a test of hand–eye coordination. On several tests, higher concentrations of mercury were associated with worse performance, but results were not significance, presumably at least in part because of the small sample size. Even the covariates that generally predict neurobehavioural test scores also tended not to be significantly associated with children’s performance in this limited sample. Other factors of possible relevance include the need to rely on an interpreter in administering the tests, and the wide age range of the children. Visual evoked potentials were also recorded but were not significantly related to concentrations of mercury (Weihe et al., 2002).

A study in children exposed to relatively high levels of methylmercury via fish consumption was conducted in French Guiana. Children from three American Indian communities with different levels of exposure were studied: high exposure (n = 153), medium exposure (n = 69), and low exposure (n = 153). A standard neurological examination was carried out on all children aged 9 months to 6 years from the group with a high level of exposure to methylmercury (n = 97) and on a random sample of children (n = 69), matched for age and sex, from the groups exposed to medium and low levels of methylmercury. In the groups with high and low exposure, children aged 5–12 years underwent a battery of neuropsychological tests modified in light of the cultural and linguistic settings: finger tapping, leg coordination, Stanford-Binet copying and bead memory subtests, and digit span. It was frequently necessary to use an interpreter. Hair samples were collected from the child and his or her mother, with the concentration of mercury in maternal hair interpreted as a proxy for the child’s level of prenatal exposure. Mean concentration of mercury in maternal hair in the groups exposed to high, medium and low levels of methylmercury, respectively, were 10.2 mg/kg, 6.5 mg/kg and 1.4 mg/kg. Mean concentrations of mercury in maternal hair were 12.7 mg/kg, 6.7 mg/kg, and 2.8 mg/kg, respectively. Women in the three groups differed in many respects, including language spoken, education, income, consumption of alcohol, prenatal care, illness during pregnancy, and intelligence quotient (IQ). No major neurologic signs were observed in the children. The prevalence of increased tendon reflexes increased across mercury categories, although the trend was significant only for boys (<5 mg/kg, 2.6%; 5–10 mg/kg, 13.2%; >10 mg/kg, 27.9%). After adjusting for age, sex, and place of birth, the odds ratio (OR) for increased reflexes in boys was 5.2 (95% CI, 1.1–2.2). For girls, the OR was close to 1. After 9 months, only 3 out of 10 children who had been found to have slower tendon reflexes at the initial examination were found to have slower reflexes upon re-examination. For the neuropsychological tests, after adjusting for age, sex, examiner, and (for some end-points) parity and place of birth, children with higher concentrations of mercury in maternal hair had significantly lower scores on the copying test (both sexes). In the group exposed at a high level only, higher concentration of mercury in maternal hair was significantly associated with poorer scores on bead memory and leg coordination (boys only) and copying test (girls only). In general, concentration of mercury in maternal hair was more strongly associated with child outcomes than was concentration of mercury in the child’s hair. Thus, children’s scores for selected neuropsychological tests were associated with higher current maternal exposures to mercury, although some associations were sex-specific and many test scores were not associated with any index of exposure to mercury. In order to link the neurobehavioral deficits of the children to their prenatal exposures to mercury, it is necessary to assume that the mother’s diet and the concentrations of mercury in fish had not changed appreciably since the birth of her child (Cordier et al., 2002).

Children from several areas of Ecuador, including gold-mining settlements in which elemental mercury was used in the extraction process, were studied. Neuro-otological examinations were carried out on 114 children (32 from a remote gold-mining settlement, 37 from a gold-mining town, and 15 from a control, non-gold-mining community). The method by which the children were selected, and the size of the population from which they were sampled, is unclear. The mean concentrations of blood mercury in the three groups, in which the mean age was 9–11 years (range, 0.5–15), were 18.2 µg/l (range, 2–89), 4.9 (range, 1–10), and 2.4 (range, 1–6), respectively. The frequency of neurological complaints (e.g. headache, dizziness, memory loss, attention deficits) was high, although it is not clear how these symptoms were elicited for the very young children, nor were the reports obtained in an unblended fashion. No statistical analyses of the frequency of neuro-otological symptoms as a function of concentrations of mercury were reported. The authors could not determine whether the children’s exposures to mercury were primarily from elemental mercury vapour liberated during burning of amalgam or from seafood contaminated with methylmercury. In addition, the effects of many other neurological risk factors, including poor nutrition, intestinal parasites, or other neurotoxic exposures such as sodium cyanide, were not considered (Counter et al., 2002).

2.3.4 Reproductive toxicity

In a case—control study, the blood mercury concentrations of couples with and without fertility problems were compared. The group of couples with fertility problems comprised 155 male and female partners of married couples undergoing in vitro fertilization. Two subgroups within this group were identified: males with abnormal sperm (on the basis of concentration, motility, and morphology) and females with "unexplained infertility" (i.e. partners with normal semen, women without anovulation, endometriosis, tubal malfunction or peritoneal adhesions).

The group of 26 controls comprised members of couples recruited in the second trimester of pregnancy. Exclusion criteria included known occupational exposure to mercury or a known etiology for semen abnormality. Blood mercury at a concentration of >50 µg/l was considered to be elevated. The participation rate was 90% in the group with fertility problems but only 20% in the control group. Overall, higher concentrations of mercury in blood and higher frequencies of elevated concentrations of mercury were found in both males and females in the former group than in males and females in the control group (males with fertility problems, 40.6 µg/l, 35% elevated; control males, 31.2 µg/l, 15% elevated; females with fertility problems, 33.2 µg/l, elevated 23%, control females: 17.5 µg/l, 4% elevated). In logistic regression analyses adjusting for age, concentration of mercury in blood was a significant predictor of fertility status. The difference in rates of elevated concentration of mercury in blood was significant only for females, however. Mercury was not speciated in this study, although self-reported consumption of seafood was significantly (but modestly) correlated with concentration of mercury in blood (correlation coefficent, 0.21), and consumption of seafood was the only source of exposure that differed between the couples with fertility problems with high and low concentrations of blood mercury (Choy et al., 2002).

Changes in the sex ratio of births in Minamata City between 1950 and 1969 were evaluated. This was the period spanning the most severe episode of methylmercury poisoning in the region. For the period 1955–1959, the presumed height of the episode, the percentage of male births declined in a manner than appears to be dose-dependent. The control ratio, based on Japanese population data was 0.515. In Minamata City, the overall rate was 0.492; in mountainous areas of Minamata City, it was 0.501; in urban areas, 0.488; in seashore areas, 0.475; in the area in which cases of Minamata disease were most prevalent, 0.459; among female patients with Minamata disease, 0.393; among fishermen’s families, 0.382. This pattern was not seen for births during 1960–1964 or 1965–1969, or among fishermen who lived in surrounding cities during the same period. Similarly, the number of male stillbirths was 1.7 times higher than that of female still-births in the period 1955–1959, versus 1.2 times among controls. This was not true for the periods 1952–1955 or 1060–1964. The authors concluded that the fact that fewer males were born to female, but not to male, patients with Minamata disease indicates that this represents a direct effect on fetuses via maternal exposure. The fact that the sex ratio returned to control levels after the period of most severe pollution suggests that methylmercury affected the fetuses themselves rather than maternal reproductive organs (Sakamoto et al., 2001).

2.3.5 Neurotoxicity in adults

Quantitative measurements of touch thresholds were determined in three patients with Minamata disease, 32 residents of a fishing village (Ooura) experiencing long-term exposure to low doses of methylmercury, and 63 residents of a control village. The mean ages in each group were 66–68 years. Current concentrations of mercury in hair were not measured. In 1960, the mean concentration of mercury in hair of patients with Minamata disease was 54.0 mg/kg and 36.95 mg/kg among 16 residents of Ooura. Two-point discrimination thresholds, measured in 26 regions, were approximately two-fold higher in both patients with Minamata disease and Ooura residents than in residents of the control village. Thresholds were increased in proximal extremities and trunk as well as in distal extremities. The authors concluded that the distal extremity paresthaesias associated with exposure to methylmercury (glove-and-stocking type) are not caused solely by damage to peripheral nerves, but also involve diffuse damage to the somatosensory cortex (Ninomiya et al., submitted).

A cross-sectional study in 129 adults aged 17–81 years (mean, 35 years) was conducted in Brazil in six fishing villages downstream from an area of gold mining in which mercury was used in the amalgamation process. Using a complete census, individuals were randomly selected, proportionally by village and age group. Recent immigrants to the area were excluded, as were individuals who reported working in the gold mining-related activities. Mean concentration of mercury in hair was 4.2 mg/kg (range, 0.56–13.6). Tests of attention, memory, learning, manual speed and dexterity, graphomotor speed, and mood were administered. Adjusting for sex, age, educational level, smoking, and alcohol intake, higher concentrations of mercury in hair were significantly associated with memory, attention, learning, and manual dexterity. Dose-dependent decrements in test scores were noted. Stratifying test scores at their respective medians, the adjusted ORs associated with having hair containing mercury at a concentration of >6 mg/kg were significant, ranging from 1.6 to 3.5. The authors concluded that these findings suggest that adult cognitive function might be as sensitive to methylmercury as children’s cognitive function. As the authors acknowledged, the cross-sectional design of this study limits the conclusions that can be drawn about the critical dose—effect relationships. The current concentrations of mercury in hair might not accurately represent concentrations in the past, which might have been responsible for the neuropsychological deficits observed at the time the study was conducted (Yokoo et al., 2003).

The question of whether methylmercury accelerates ageing was evaluated in the context of a prospective cohort study of dementia in Sweden. The study population consisted of all persons born in 1912 and before (2368 participants), living in the Kungsholmen district of Stockholm in October 1987. In this phase of the study, 106 subjects were selected for additional follow-up, although it is not clear on what basis the selection was made. Follow-up assessments included measurement of concentrations of mercury in blood and blood pressure, and administration of the mini mental status examination (MMSE). The mean concentration of total mercury in blood in this group was 3.4 µg/l (range, 0.4–16), or 17 nmol (range, 2–80). Concentration of mercury in blood was not significantly related to either systolic or diastolic blood pressure, or to MMSE score. The authors concluded that because blood mercury is a relatively short-term index of exposure to mercury, it most likely did not reflect the appropriate period from the standpoint of neurotoxicity. In addition, it was surprising that MMSE scores were not related significantly to age in this cohort (range, 81–94 years, well above average life expectancies for both men and women in Sweden), perhaps indicating that the study sample was an exceptionally healthy subsample of the Swedish population (Johansson et al., 2002).

2.3.6 Immunotoxicity

In order to explore the potential immunotoxicity of methylmercury, exposure to mercury and host resistance to parasitic disease were evaluated in a cross-sectional study of 205 people aged >2 years, living in randomly-selected households in a town located downstream from a region of intensive gold-mining activity in which mercury is used in amalgamation. The sample represented 10% of the total population. Hair concentration of mercury was used as the biomarker of exposure, and the mean was 8.6 mg/kg (range, 0.3–83.2). The prevalence of malaria was determined from blood smears. All individuals were negative for malaria, although 66% reported a history of malaria within the previous 5 years. Controlling for age, sex, duration of residence in the town, and income, the OR for past infection with malaria was 4.2 for those with a history of working with mercury (16 persons). Concentration of mercury in hair was not, however, associated with a self-reported past infection with malaria. Because of the remote site, studies that would have provided useful data (studies of immune function, or measurement of plasma cytokine levels) could not be carried out. Without any active malaria infections, the association between concentration of mercury in blood and levels of parasitaemia could not be investigated. The present reviewer noted that the association between the risk of past infection with malaria and a history of working with mercury, but not hair concentration of mercury, raises the possibility that it was occupational exposure to mercury vapour, rather than to methylmercury, that was responsible for increasing an individual’s risk of malaria (Crompton et al., 2002).

2.3.7 Cardiotoxicity

Cardiovascular autonomic activity in nine patients with congenital Minamata disease was compared with that in 13 age-matched healthy control volunteers. Blood pressures and heart-rate variability (time and frequency domain analyses) were recorded for 5 min during spontaneous breathing after subjects had rested in a supine position for 10 min. The critical end-points were the mean, standard deviation, and coefficient of variability of the R-R intervals. Mean R-R interval was significantly lower (i.e. increased heart rate) in the group of patients with congenital Minamata disease. The other two end-points were also lower in this group, but not significantly so. The frequency domain analysis indicated that the high-frequency component, an index of parasympathetic activity, was also significantly lower in this group. The low-frequency component (sympathetic activity) was lower but not significantly so. Groups did not differ in systolic or diastolic blood pressures, although pulse pressure was significantly lower in the group of patients with congenital Minamata disease (Oka et al., 2003).

In a 7-year follow-up study of 1833 Finnish men aged 42–60 years who, upon enrollment, were free of clinical heart disease, stroke, claudication, or cancer, the mean concentration of mercury in hair was 1.92 mg/kg (range, 0–15.67). The main end-points were acute myocardial infarction, death from coronary heart disease, or cardiovascular disease. Adjusting for age, examination year, ischaemic exercise electrocardiogram (ECG), and maximal oxygen uptake, men in the tertile with the highest concentration of mercury in hair (>2 mg/kg) had a two-fold higher risk of acute myocardial infarction (95% CI, 1.2–3.1). A fish intake of >30 g/day doubled the risk of acute myocardial infarction, and each additional increase in fish intake of 10 g/day was associated with a 5% increase in the 5 year risk (Salonen et al., 1995).

An additional follow-up examination of 1104 of the men was conducted after 4 years. Both at baseline and follow-up, each man underwent an ultrasound examination of the carotid arteries to assess the extent and severity of atherosclerosis, operationalized as the change in intima-media thickness. The increase in intima-media thickness was 32% greater in men in the quintile with the highest concentration of mercury in hair (>2.81 mg/kg) than among men in the other four quintiles. The increase appeared to be similar in the four lower quintiles. The authors speculated that the atherogenic effect of methylmercury results from its enhancement of lipid perioxidation (Salonen et al., 2000).

A case–control study was conducted in 684 men aged < 70 years with a first diagnosis of myocardial infarction, living in any of eight European countries or Israel, and 724 controls frequency matched on age from the same areas. The bio-marker of exposure used was mercury in toenail clippings. The OR for myocardial infarction associated with the highest quintile of mercury in toenails was 2.16 (95% CI, 1.09–4.29), after adjustment for age, centre, DHA, body mass index (BMI), smoking, alcohol, high density lipoprotein (HDL), diabetes, hypertension, parental myocardial infarction, alpha-tocopherol, beta-carotene, selenium, and weight of toenail, (Guilar et al., 2002).

A case–control study nested within the health professionals follow-up study involved 33 737 male health professionals aged 40–75 years. The cases comprised 470 men with coronary heart disease (including surgery on coronary arteries, nonfatal myocardial infarction, fatal coronary heart disease). Controls were matched on age and smoking. The biomarker of exposure, mercury in toenails, was not associated with risk of coronary heart disease. The adjusted OR in the highest quintile was 0.97 (95% CI, 0.63–1.50). A potentially important aspect of this study, however, is that 58% of the study participants were dentists, and their mean concentration of toenail mercury was twice as high as that among non-dentists. This suggests that exposure to mercury vapour (metallic mercury) contributes significantly to concentration of mercury in toenails. How informative this study is regarding exposure to methylmercury and cardiovascular risk is thus open to question (Yoshizawa et al., 2002).

In a population-based case–control study, 440 incident cases of stroke were identified among individuals aged 40–85 years. After controlling for confounders, the OR for individuals in the highest quintile of fish consumption (>46 g/day) was 1.95 (95% CI, 1.14–3.33), compared with that for individuals in the lowest intake quintile (11 g/day). The OR associated with omega-3 fatty acid intake was somewhat lower and not statistically significant. The author speculated that low doses of omega-3 fatty acids protect against atherosclerotic thrombotic disease via several mechanisms but that high doses increase the risk of haemorrhage. Exposure to methylmercury was not measured in this study. The present reviewer noted that the possibility that contaminants of seafood, rather than omega-3 fatty acids, might play a role was not considered (Caicoya, 2002).

2.3.8 General health of residents of the Minamata City area

Annual follow-up health examinations have been conducted since 1984 on approximately 1500 persons aged >40 years in a town in which the prevalence of Minamata disease is higher (36.9 cases/1000 population) than it is in Minamata City. In a series of case–control analyses, it was determined whether patients with diagnoses of liver disease, renal disease, or diabetes mellitus were more likely to reside in fishing villages and could thus be assumed to have a history of higher exposure to methylmercury in the past. For none of the three classes of disease was the OR elevated for residents of fishing villages (Futatsuka et al., 2000).

Regional differences in the prevalence of 65 subjective complaints were evaluated. A comparison was made between older residents (aged >40 years) of two towns: town A (1304 participants) on the Yatsushiro Sea, where pollution with methylmercury was severe, and a similar town (town B, 446 participants) on the Ariake Sea, where no cases of Minamata disease had been recorded. Town A was divided into two areas: fishing villages and an internal control area. The rate of participation was only about 30% in both towns, and it is not stated whether the public health nurses conducting the interviews were blinded to the location of a respondent’s residence. Residents of fishing villages in town A reported a higher prevalence of complaints than did other residents of town A or residents of town B. For many symptoms, the rate was two to three-fold higher (up to 14-fold). Among males, significant regional differences were noted for stiffness, dysesthaesia, hand tremor, dizziness, loss of pain sensation, cramp, atrophy of the upper arm muscle, arthralgia, insomnia, and lumbago. Among females, significant differences were noted for leg tremor, tinnitus, loss of touch sensation, atrophy of the leg muscle, and muscular weakness (Fukuda et al., 1999).

2.3.9 Safety assessments: analyses of levels of exposure and corresponding responses

The identification of a "practical threshold" for developmental effects, particularly in regard to the fetus, can serve as a basis for an assessment of the safety of exposure to methylmercury. The end result of such an assessment is the identification of a safe level of exposure for adverse developmental consequences that is three- to five-fold lower than a similar safe dose based on the end-point for adults. The following is a summary of various assessments of the adverse consequences on fetal development using human environmental epidemiological studies conducted in the Faroe Islands, Iraq, the Seychelles and New Zealand.

In considering such a safety assessment, one approach is to use the classical no-effect level/no-observed-adverse-effect level/lowest-observed-adverse-effect level (NEL/NOAEL/LOAEL) approach in which a level of exposure is identified and judged to be without appreciable risk (WHO, 1987; Lu & Sielken, 1991; Beck et al., 1993). While this approach is certainly viable when considering animal studies in which discrete numbers of animals are assigned to a number of groups receiving different doses, it does become somewhat problematic when considering environmental epidemiological studies, such as those that are available for methylmercury, and which serve as the primary basis for any consideration of safety or risk (Beck et al., 1993; Barnes et al., 1995). In these types of studies, population exposures do not comprise a few discrete values, as is normally the case in a bioassay in animals, but are continuous. When an effect is noted in such a study, a mean exposure is often assumed to be the LOAEL, but such an assessment can only determine that an effect may have occurred within a range of levels of exposure, and may very well span several orders of magnitude. A similar uncertainty would also apply to the identification of a NOAEL where mean exposure may not be a good determinant of a NEL/NOAEL. An alternative to the classical approach in which a NOAEL/LOAEL is identified in an environmental epidemiological study is that of the benchmark dose (BMD), the dose that corresponds to a specified level of additional response above background (e.g. 1, 5 or 10%) called the benchmark response (BMR). A lower statistical confidence bound (95%) on the BMD is then used as a replacement for the NOAEL. It is generally defined as a calculated NOAEL. It has been proposed that the BMD approach is superior to the use of "average" or "grouped" exposure estimates when dose–response information is available, as are used in epidemiological studies on methylmercury. The BMD approach has several advantages over the NOAEL approach, including making use of the entire exposure–response range and reflecting sample size more appropriately. Another advantage, and this applies specifically to the study in the Seychelles, is that negative data can be used to estimate the lower limit of the 95% confidence interval of the benchmark-dose level (BMDL). With negative data, the point estimate for the dose–response trend can be zero, or in a positive or negative direction. If the estimate of trend is negative or zero then the BMD will be undefined or infinite, while if it is positive it will be finite. When a study gives negative results and it is possible that there is no cause–effect relationship, the statistical constraints of the study design determine the BMDL. The BMDL represents a conservative value (an overestimation) and with negative data it represents the lowest estimate of a NEL/NOAEL. The value of the BMDL reflects the potential magnitude of any small effect of methylmercury that might have gone undetected in a study (Crump, 1984; Beck et al., 1993; Lu & Sielken, 1991; Barnes et al., 1995; National Academy of Sciences/National Research Council, 2000).

Study in the Faroe Islands

Several BMD analyses have been conducted by the study investigators in order to identify doses of methylmercury that result in an increased probability of an abnormal response in the study population (Budtz-Jorgensen et al., 1999a, 2000, 2001, 2002). In these assessments, the background occurrence of an adverse effect in the population is generally fixed at 5%, while the BMR varies from 2% to 10%. For each of five major cognitive functions assessed in the study in the Faroe Islands, one neuropsychological test was selected. Benchmark doses were calculated using each of these variables as the response variable. The tests selected were: neurobehavioural evaluation system (NES2) finger tapping, preferred hand (representing motor ability); NES2 continued performance, reaction time (representing attention); the Bender test (representing visuospatial ability); the Boston naming test, with cues (representing language ability); and the California verbal learning test, long delay (representing memory ability).

The BMD analysis was performed using concentration of mercury in both maternal hair and cord blood as the exposure variable, and a uniform set of confounding variables was identified. The higher the BMR, the higher the BMD. Using concentration of mercury in maternal hair as the exposure variable for all five response variables, there is almost no difference in the fit between the three dose–response functions (linear, logarithmic and square root function) used in the analysis. However, some variation is seen in the calculated BMDL values. For a 10% BMR, the highest variation is seen for the continuous performance test (CPT) of reaction time, where the BMDLs vary by a factor of 2.6. As anticipated, the lowest BMDLs are seen for the response variables where the effect of mercury is strongest, i.e., the CPT reaction-time test (BMDL = 6–13 µg/g) and the Boston naming test (BMDL = 9.14 µg/g).

For the concentration of mercury in cord blood, the results depend more on the assumed dose–response function. Again, the highest variation in BMDL for a BMR of 10% is seen for the reaction-time test. For this response variable, the fit of the linear dose–response function is significantly worse than the fit of the logarithmic function. The difference in fit between the logarithmic and the square root function is not significant for any of the response variables, although the variation in the corresponding BMDL values is considerable. Again, the lowest BMDLs were seen for the CPT reaction-time test (BMDL = 3–96 µg/l) and the Boston naming test (BMDL = 9.66 µg/).

An ad hoc committee of the United States National Academy of Sciences/ National Research Council concluded that the study in the Faroe Islands was the most appropriate study to use for the derivation of a reference dose (RfD) for methylmercury. This recommendation was in turn adopted by the United States Environmental Protection Agency (Rice et al., 2003). The Committee concluded that it was best to identify what it called a "point of departure" (POD) and recommended the use of the 95% BMDL as the POD. The Committee did caution that the type of statistical analysis used could have a substantial effect on the estimated BMDL. The Committee also recommended that the K-power model with a constraint of K >1 be used to make such an estimation (see Table 1). On the basis of the analyses of cord blood, the lowest BMDL for a neurobehavorial end-point considered by the Committee to be sufficiently reliable is the Boston naming test. The Committee’s preferred BMDL estimate is 58 ppb for mercury in cord blood. With regard to the size of the uncertainty factor to apply to the BMDL, the Committee concluded that a factor of 10 would account for biological variability in dose estimation and for insufficiencies in the database (i.e. possible sequelae after low doses and latent effects, and immunotoxicity and cardiovascular effects).

Table 1. K-power model for calculations of benchmark dose (mercury in cord blood, ppb), study in the Faroe Islands

End-point

BMD

BMDL

Finger tapping

140

79

CPT reaction

72

46

Bender copying errors

242

104

Boston naming test

85

58

California verbal learning test: delayed recall

246

103

CPT, continuous performance test

BMD, benchmark dose

BMDL, benchmark dose level

BMDs were calculated on the assumption that 5% of the responses would be abnormal in unexposed subjects, assuming a 5% excess risk (BMR = 0.05)

Iraq

Several safety assessments have relied on the data acquired in the study of the episode of poisoning with methylmercury that occurred in Iraq in 1971–1972 as a result of the consumption of bread and cereals made from grain treated with fungicides containing alkyl-mercury (WHO, 1976; Crump et al., 1995: Dourson et al., 2001). There are limitations in this study that warrant caution in its use in such an assessment (Clewell et al., 1999). The number of mother–infant pairs (83) in the study is fairly small, and most of the mothers whose offspring were observed to have decrements in nervous system function had body burdens that exceeded the LEL for an adult, namely paresthaesia at a hair concentration of mercury of 50 ppm. Other limitations include uncertainties surrounding the similarity of the toxicokinetics of methylmercury in grain versus in fish, different nutrient interactions in grain versus in fish and differences in type and duration of exposure, namely short-term versus long-term exposure. The level of the background occurrence of either delayed walking or of symptoms affecting the central nervous system in a population has a considerable impact on the estimated "practical threshold" for methylmercury-induced effects on either response. The concentration of mercury in hair at a hypothetical NOAEL for developmental effects was determined by application of a BMD approach. The analysis used the combined incidence of all neurological effects in children exposed in utero.

The continuous data for the Iraqi population were grouped according to five doses, and incidence rates were provided for delayed onset of walking, delayed onset of talking, mental symptoms, seizures, neurological scores of > 3, and neurological scores of > 4 for affected children. Neurological scores were determined by clinical evaluation for cranial nerve signs, speech, involuntary movement, limb-tone strength, deep tendon reflexes, plantar responses, coordination, dexterity, primitive reflexes, sensation, posture, and ability to sit, stand and run. The effects of late walking, late talking, and neurologic scores of > 3 were also combined for additional analysis.

In order to adjust for background rates of effects, the estimates of BMD for methylmercury were calculated to estimate the dose associated with "extra risk". Conceptually, extra risk is the additional incidence at which an effect is observed above the background rate of incidence, relative to the proportion of the population of interest that is not expected to exhibit such an effect. Extra risk is then more easily interpreted than additional risk, because it applies the additional risk only to the proportion of the population that is not represented by the background rate. Estimates of benchmark dose were made by calculating the 95% BMDL on doses corresponding to the 1%, 5% and 10% extra risk levels, using a quantal Weibull model.

For each end-point and for the combined end-points, the incidence of response was regressed on the dose. A chi2 test of goodness-of-fit was used to test the null hypothesis that the predicted incidence was equal to the observed incidence, so that the null hypothesis would be rejected for p values of < 0.05. Results for individual effects and for all effects combined for children exposed in utero are given in Table 2.

Table 2. Estimates of benchmark dose for methylmercury (ppm in hair): maximum likelihood estimates and 95% lower confidence limits from a Weibull model, based on incidence of effects in children exposed in utero

Effect

5% BMR

10% BMR

Goodness-of-fit

MLE

95% CL

MLE

95% CL

p value

Late walking

16.7

10.9

34.3

22.4

0.16

Late talking

22.1

12.3

43.8

25.3

0.79

Mental symptoms

61.0

32.8

125.4

67.5

0.63

Seizures

60.4

34.3

124.2

70.5

0.86

Neurological score >3

28.8

17.0

59.1

34.9

0.58

Neurological score >4

41.4

23.7

84.9

48.7

0.4

All end-points

8.3

5.4

17.1

11.1

0.94

BMR, benchmark response

MLE, maximum likelihood estimates

95% CL, 95% lower confidence limit

Goodness-of-fit, p value for testing the null hypothesis

Seychelles

In the derivation of the minimal risk level (MRL), the United States Agency for Toxic Substances and Disease Registry/Centers for Disease Control and Prevention used results of the Seychelles child development study for children aged 5.5 years (Davidson et al., 1998). As was noted above, children were exposed pre-and postnatally via the consumption of fish. The NOAEL was considered to be the mean concentration of total mercury, 15.3 ppm, in maternal hair in the group of children with the highest exposure in the test cohort at age 5.5 years, and was used as the basis for derivation of an oral MRL. It was assumed that the concentration of mercury in hair is 250 times that in blood, and that the corresponding concentration of mercury in blood would therefore be:

1/250 × 15.3 µg/g × 1 mg/1000 µg × 1000 g/l = 0.061 mg/l

The concentration of mercury in blood was then related to steady-state ingestion using the following equation:

C = f × d/b × V = AD × Ab = d/b × V

where:

C = concentration in blood, 0.95 µg/l

F = fraction of daily dose taken into blood, 0.05

d = daily intake (µg/day)

b = elimination constant, 0.014

V = volume of blood in a 60 kg woman, 4.2 litres

AD = percentage of mercury that is absorbed from the diet

Ab = percentage of absorbed amount of mercury that enters the blood

Using a body weight of 60 kg for women, the estimated dose that would result in a mean concentration of mercury in hair of 15.3 ppm is 0.075/60 kg = 0.0013 mg/kg per day and was defined as the NOAEL from the most heavily exposed group (95 children) at age 5.5 years.

An uncertainty factor of 4.5 was identified on the basis of the following issues.

An alternative considered by the Agency for Toxic Substances and Disease Registry was a physiologically-based toxicokinetic approach using the mean concentration of total mercury of 6.8 ppm in maternal hair for the entire Seychellois cohort. Using the same formula for the hair : blood ratio as cited above, the corresponding steady-state ingestion is 0.6 µg/kg bw per day. Several uncertainty factors were then applied. An uncertainty factor of 1.5 was used to account for the toxicokinetically-based variability of hair : blood ratios in pregnant women and fetuses in the population of the United States (Clewell et al., 1999). The size of the study population in combination with an uncertainty factor of 1.5 was considered adequate to encompass the full range of kinetic and dynamic variability within the human population. An independent modifying factor of 1.5 was also used to take into consideration the positive results of the domain-specific tests administered in the study in the Faroe Islands. The uncertainty factor of 1.5, multiplied by the modifying factor of 1.5, yields a total aggregate value of 2.25. Applying the factor of 2.25 to the daily intake of 0.0006mg/kg per day derived from the NOAEL yields an oral MRL value of 0.0003 mg/kg per day.

An assessment of BMD was used by Crump et al. (2000) to model the data from the tests on children aged 0.5, 1.6, 2.4, and 5.5 years in the Seychellois population. Developmental milestones, including age of walking and talking, were also used in this assessment. A number of dose–response models were considered, together with sets of covariates and definitions of background responses. Responses were modelled as continuous or quantal data. The 95% lower bound on the 10% benchmark-dose level (BMDL) was deemed to represent a conservative estimate of the traditional NOAEL. The modelling of the benchmark dose using a variety of BMD analysis methods yielded a range for BMDL10 of 21.6 to 26.6 ppm for methylmercury in maternal hair. This BMDL was converted to an expected distribution of daily ingestion rates across a population of women of child-bearing age in the United States, by using a Monte-Carlo analysis with a physiologically-based pharmacokinetics model for methylmercury. This analysis addresses the impact of interindividual pharmacokinetic variability on the relationship between ingestion rate and concentration of methylmercury in hair. The resulting distribution had a geometric mean value of 0.00160 mg/kg per day (SD, 0.00133). The first, fifth, and tenth percentiles of that distribution were 0.00086, 0.00104, and 0.00115 mg/kg per day, respectively. The authors concluded that the fifth percentile of 0.00104 mg/kg per day provides a scientifically sound and conservative basis for an intake limit, that it incorporates the pharmacokinetic variability across the population of women of child-bearing age in the United States, and that no other uncertainty factor for interindividual variability is needed.

Table 3. Benchmark (BMDLa) concentrations of mercury in maternal hair (ppm), study in the Seychelles

Type of data

Definition of abnormal

Covariates

Model

No. of BMDLs calculated under each set of assumptions

Average

Range

Continuous

Response > SDs from the mean response

None

Weibull

12

26.6

25.3–29.0

     

Logistic

12

26.7

25.3–29.1

     

K-power

12

26.1

24.7–27.8

 

Lowest 5% of responses

 

Weibull

12

25.0

23.1–27.2

     

Logistic

12

25.0

23.0–27.3

     

K-power

12

24.4

20.0–26.7

     

Weibull

12

25.4

23.5–30.1

     

Logistic

12

25.4

23.3–30.3

     

K-power

12

24.5

19.4–26.8

     

Weibull

12

24.8

23.4–27.2

     

Logistic

12

24.7

23.2–27.2

     

K-power

12

24.9

23.2–26.9

Quantal

   

Weibull

17

21.6

15.8–23.7

a

95% statistical lower bound on concentration of mercury in maternal hair associated with an increase of 0.1 in the probability of an adverse effect

SD, standard deviation

New Zealand

Calculations of BMD and additional regression analyses of data from the study in New Zealand in which scores from 26 scholastic and psychological tests administered to 237 children aged 6 and 7 years were correlated with prenatal exposure to mercury as measured by concentration of mercury in maternal hair. The original analyses of five test scores found an association between high prenatal exposure to mercury and decreased test performance, using categorical variables for exposure to mercury. Regression analyses, which used the actual concentration of mercury in hair, did not find significant associations between mercury and children’s test scores. However, this finding was highly influenced by the results for a single child whose mother had the highest concentration of mercury in hair (86 mg/kg) in the cohort. When this child was excluded, the results were more indicative of an effect of mercury and scores for six tests were significantly associated with the concentration of mercury in the mother’s hair. BMDs for mercury in maternal hair, calculated from results of five tests, ranged from 32 to 73 mg/kg, and corresponding BMDLs (lower 95% confidence limits on BMDs) ranged from 17 to 24 mg/kg. When the child with the highest concentration of mercury in maternal hair was omitted, BMDs ranged from 13 to 21 mg/kg and the corresponding BMDLs ranged from 7.4 to 10 mg/kg (Crump et al., 1998). The corresponding ingestion rates for the average BMDLs, with and without results for this child, are about 1.4 and 1.2 µg/kg bw per day, respectively.

Table 4. Benchmark levels of mercury in maternal hair (mg/kg), New Zealand

Neurological test

All children

Omitting child with highest concentration of mercury in maternal hair

BMDa

BMDLb

BMDa

BMDLb

Abnormal outcome defined so that lowest 5% of test scores in unexposed children are abnormal (p0 = 0.05)

TOLD-SL

45

20

15

9.5

WISC-RP

73

24

15

10

WISC-RF

51

21

15

10

MCC-PP

32

17

13

7.4

MCC-MOT

55

21

21

9.8

Abnormal outcome (chio) defined as test score of two SD below the mean in standard reference population

TOLD-SL

22

9.4

7.1

4.0

WISC-RP

110

31

 

 

WISC-RF

 

 

 

 

MCC-PP

 

 

 

 

MCC-MOT

 

 

 

 

a

Average concentration of mercury (mg/kg) in maternal hair during pregnancy predicted to cause an additional 10% of children to have an abnormal test score

b

95% lower bound on BMD In order to achieve normally distributed residuals, -Ln [10] test score was used as the independent variable

 

TOLD-SL, test of language development—spoken language

 

WISC-RP, Wechsler intelligence scale for children—revised performance IQ

 

WISC-RF, Wechsler intelligence scale for children—revised full-scale IQ

 

MCC-PP, McCarthy scales of children’s abilities perceptual—performance scale

 

MCC-MOT, McCarthy scales of children’s abilities—motor scale

3. ESTIMATED DIETARY INTAKE

3.1 Environmental sources of mercury and methylmercury

Methylmercury is produced from inorganic mercury by bacteria found at the bottom of lakes and seas. It is taken up into the flesh of living organisms in the aquatic environment and is known to bioaccumulate in piscivorous species. Numerous studies have shown that older and larger predatory fish species, such as shark and swordfish, contain the highest concentrations of methylmercury, occasionally surpassing 5 mg/kg of flesh, wet weight. Consumption of fish is the predominant route of exposure to methylmercury. Mercury in foods other than seafood is predominantly found as inorganic mercury.

The estimated intake of methylmercury from all sources is 2.41 µg/day (WHO, 1990). Mercury is found in air at concentrations of <10 ng/m3, with mono- and dimethylmercury comprising approximately 20% of the total. The resulting intake from air is <0.04 µg/day, an insignificant contribution to total intake of methylmercury. In water, concentrations of total mercury range from 10 to 50 ng/l, however, any methylmercury produced in water is rapidly removed by biota; in unpolluted waters, the concentration of mercury is <1 ng/l. The intake of methylmercury resulting from drinking 2 l of water would be <0.002 ng/day, also an insignificant contribution to total intake.

3.2 Biomarkers of exposure

The two most-studied biomarkers of exposure to methylmercury are hair and blood. Numerous studies have examined the relationship between intake of mercury and methylmercury, and consequent concentrations in hair, blood, urine, and faeces over time (Pesch et al., 2002; Sanzo et al., 2001; Pellizzari et al., 1999; Batista et al., 1996). The analysis of urine is more suited to examining total exposure to mercury, especially that resulting from the use of mercury in dental amalgams (Pesch et al., 2002). Analysis of faeces has not been useful for examining exposure to methylmercury.

During the course of a survey of current hair concentrations of mercury in five districts in Japan, it was found that thioglycolate contained in the lotion frequently used for artificial waving removes mercury from hair. The concentration of mercury in unwaved hair is relatively constant as a function of distance from the hair root, while the concentration of mercury in waved hair declines steadily with distance from the root. Such treatment reduced mercury content by as much as 50%. Failure to take into account the use of such lotion and the distance of a hair segment from the root could introduce error into the dose–response/dose–effect relationship estimated in studies that rely on concentration of mercury in hair as the biomarker of exposure (Yasutake et al., 2003).

The ratio of mercury in blood : hair has been explored and modelled. Assuming steady dietary conditions, the concentration of mercury in a hair when it forms is approximately 250 times higher than that in blood (WHO, 1990). Thus, analyses of either blood or hair strands could provide an estimate of recent exposure to methylmercury. Analysis of hair is the preferred method because it is non-invasive and represents a time course of exposure to methylmercury as the strand grows approximately one centimetre per month.

Numerous investigations in populations that eat large quantities of fish, such as native populations living by subsistence in the Great Lakes region and the Pacific Northwest of the United States, Inuit women in Northern Quebec, Canada, and in Munduruku Indians of the Tapajos River basin in Brazil, have revealed that hair and blood concentrations of methylmercury are elevated in these populations relative to those in populations that eat a more varied diet (Gerstenberger et al., 1997; Marien & Patrick, 2001; Muckle et al., 2001; Oliveira Santos et al., 2002). In a study published in 2003, investigators following medical patients in an urban setting in a California clinic noted that 89% had blood concentrations of methylmercury that were above the maximum of 5.0 µg/l (corresponding to a reference dose of 0.1 µg/kg bw per day) recommended by the United States Environmental Protection Agency. Self-reported intakes of fish for this group as a whole were high compared with the average in the United States. High intakes of large predatory fish, such as swordfish, were significantly associated with elevated concentrations of mercury in blood. These findings are in agreement with those of other researchers (Hightower & Moore, 2003).

In studies of biomarkers of exposure to methylmercury, when participants are asked to stop consuming seafood, significant reductions in measured concentrations of methylmercury in hair and blood are observed. Table 5 gives an overview on reports published since 1996.

Table 5. Studies of biomarkers of exposure to mercury and methylmercury

Country

Biomarker

Population

Elevated consumption of fish?

Concentration of mercury (mean, or range of means, as noted)

Reference

Brazil

Total Hg, hair, blood

Indigenous

Yes

14.45 µg/g (aged 7–12 years)

Oliveira Santos et al. (2002)

       

15.7 µg/g (aged 14–44 years,female)

 
       

14.1 µg/g (others)

 

Canada

Total Hg, hair, blood

Indigenous

Yes

4.4 µg/g

Muckle et al. (2001)

China

Total and MeHg, hair

Representative

No

0.42 µg/g (China)

Feng et al. (1998)

Indonesia

     

0.78 µg/g (Indonesia)

 

Japan

     

2.1–4.9 µg/g (Japan)

 

Germany

Total Hg, urineb

Representative

No

0.4–2.0 mg/l

Becker et al. (2003)

Japan

Total Hg, hair

Representative

Yes

1.76–3.37 µg/g

Yasutake et al. (2003)

Portugal

Total and MeHg, blood

Children

Yes

2.7 ng/kg

Evens et al. (2001)

Spain

Total Hg, hair

Children

No

0.8 µg/g

Batista et al. (1996)

 

Total Hg, blood

Representative

Yes

11–22 ng/g

Sanzo et al. (2001)

Sweden

Total, MeHg, blood,hair

Pregnant women

Yes

0.35 µg/g (hair)
1.3 µg/l (cord blood)

Bjornberg et al. (2003)

UK

Total Hg, hair

Pregnant women

No

0.19 µg/g

Lindlow et al. (2003)

USA

Total Hg, hair

Representative

No

0.3 µg/g

Pellizzari et al. (1999)

 

Total Hg, hair, blood

Women aged 16–49 years

No

1.2 µg/l

NHANES (2001)

 

MeHg, hair

Women aged 15–45 years

No

0.4 µg/g

Smith et al. (1997)

 

Total Hg, hair, blood

Indigenous

Yes

0.83 µg/g

Gerstenberger et al. (1997)

 

Total Hg, blood

Representative

Yes

14.5 µg/l

Hightower & Moore (2003)

a

Italic typeface serves to differentiate the form(s) of mercury analysed and the biomarker reported in the table

b

Inorganic mercury

Hg, mercury

MeHg, methylmercury

3.3 Concentrations of methylmercury in seafood

Fish and shellfish consume aquatic organisms that contain methylmercury. It has been estimated that methylmercury comprises 75–100% of the total mercury in seafood. The concentration of methylmercury in a given species of fish is dependent on the animal’s diet, size, and age. Piscivorous fish are known to contain the highest concentrations of methylmercury in their flesh. Generally, concentrations are <0.4 mg/kg, but fish at the highest trophic levels may contain methylmercury at concentrations of >5 mg/kg. Reported concentrations of methylmercury in fish (WHO, 1990) are listed in Table 6.

Table 6. Estimated concentrations of mercury in fish

Fish species

Concentration of mercury (mg/kg of fish)

Mackerel

0.07–0.25

Sardine

0.02–0.3

Tuna

0.03–1.2

Swordfish

0.06–0.8

Shark

0.004–1.8

Other

0.03–0.3

From WHO, 1990

Since 1997, a number of papers have been published concerning concentrations of mercury and methylmercury in seafood caught in specific regions of the ocean or in local waters. In Japan, 360 samples from 28 species of seafood available at a Tokyo fish market were analysed as part of an overall assessment of mercury intake. Concentrations of methylmercury were not measured; for the assessment of exposure, it was assumed that 75% of the total mercury was methylmercury. Large fish, such as tuna and swordfish, were found to contain the highest concentrations of total mercury, as much as 5.2 mg/kg for tuna, and 3.01 mg/kg for swordfish. Most species contained mercury at a concentration of <1 mg/kg, with salmon, prawn, and smelt containing 1–1.5 mg/kg (based on 32, 22, and 8 samples, respectively). The authors concluded that, on the basis of comparison with data obtained in 1975, concentrations of mercury in commercially available seafood in Tokyo have risen, particularly in predatory species (Nakagawa et al., 1997).

Four papers concerning concentrations of mercury and methylmercury in fish caught off the Italian coasts of the Adriatic and Mediterranean Seas were published between 2000 and 2002. In the first paper, total mercury in megrim (112 samples), common sole (100 samples), striped mullet (312 samples), anglerfish (120 samples), and black-bellied angler (156 samples) caught in the South Adriatic Sea was measured. These species represent a mix of benthic (bottom-feeding) and pelagic (open sea-feeding) fish. The pelagic species (megrim, sole, and mullet) contained mercury at a concentration of 0.4, 0.2 and 0.3 mg/kg, respectively, with all samples containing <1 mg/kg. The benthic anglerfish species contained concentrations of mercury as high as 2.22 mg/kg, with means of 1.3 mg/kg for angler-fish and 0.7 mg/kg for black-bellied angler. The authors suggest that these results can be ascribed to feeding behaviours. A correlation between size of fish and concentration of mercury was also observed in each species (Storelli & Marcotrigiano, 2000).

The second paper was concerned with swordfish and bluefin tuna caught in the Mediterranean Sea. Higher concentrations of mercury were observed in bluefin tuna (169 samples) than in swordfish (162 samples), with a mean of 1.02 mg/kg for tuna (range, 0.07–4.26 mg/kg) as opposed to a mean of 0.49 mg/kg (range, 0.15–1.05 mg/kg) for swordfish. In a given weight range, concentrations of mercury in tuna were approximately four times higher than those in swordfish. The authors again attributed these results to feeding behaviours, as well as the fact that as tuna has a slower rate of growth than swordfish, and takes a longer time to reach a given weight. A correlation between size and mercury concentration was observed. In light of overall intake of tuna and regulatory limits in place in the European Community, the authors suggest that limits on the size of tuna might be necessary to control exposure to mercury (Storelli & Marcotrigiano, 2001).

The third paper concerned mercury and methylmercury in tuna and sharks taken in the Adriatic Sea. The average concentration of mercury in spiked dogfish (Squalus acanthias) was 6.5 mg/kg, an extremely high level. This species of shark is long-lived, slow to grow and reproduce, and additionally frequently feeds on or near the sea bottom. Bullet tuna and blue shark had much lower concentrations of mercury. Methylmercury content ranged from 69% to 100% of the total mercury (Storelli et al., 2001).

The fourth paper was concerned with mercury and methylmercury residues in bluefin and albacore tuna, two species of great commercial importance caught in the Mediterranean Sea off the coast of Italy. The average concentration of mercury in each species exceeded 1 mg/kg, ranging from 0.84 to 1.45 mg/kg for albacore (127 samples; average, 1.17 mg/kg) and from 0.16 to 2.59 mg/kg for bluefin (161 samples; average, 1.18 mg/kg). Methylmercury comprised >90% of the total mercury for each species (range, 75–100%; average, 91%). Age and body weight were again positively correlated with mercury concentrations in the fish. The authors note that the concentrations observed were higher than those reported in the literature for tuna species, but suggest that direct comparison is difficult owing to lack of information on fish age and size and insufficient sample sizes in the reports in the literature (Storelli et al., 2002).

Mercury concentrations in commonly consumed fish in Croatia were examined in 2003. Hake (three samples) was found to contain mercury at an average concentration of 0.375 mg/kg. Anchovies, sardines, bogue, mackerel, and mussels all contained mercury at an average concentration of <0.28 mg/kg (Juresa & Blanusa, 2003)

In 2001, the United States Food & Drug Administration released the results of a survey of concentrations of mercury in a large variety of fish and shellfish. Four species, (tilefish, swordfish, king mackerel, and shark) were found to contain the highest average concentrations of mercury (1.45, 1.00, 0.73, and 0.96 mg/ kg, respectively). Only three of the remaining 34 species examined contained mercury at average concentrations >0.5 mg/kg (red snapper, moonfish, and orange roughy). Tinned tuna, the most commonly consumed fish in the United States, was found to contain mercury at an average concentration of 0.17 mg/kg (248 samples), while the average concentration in fresh tuna was 0.32 mg/kg, thus in agreement with the observation cited above that concentrations of mercury in tuna caught in the Mediterranean and Adriatic Seas are higher than those in tuna taken in other waters (Food & Drug Administration, 2001)

Typical concentrations of mercury found in seafood, which have been published since 1997, are summarized in Table 7. The Committee noted that no information has been published suggesting that concentrations of methylmercury in seafood are changing systematically. When fish have been serially collected in waters where pollution has increased, elevated concentrations of methylmercury have been observed. Conversely, cleaner waters have yielded seafood with lower concentrations of methylmercury. This suggests that knowledge of environmental conditions in the waters where fish are harvested is a paramount consideration for estimating exposure to methylmercury from consumption of seafood, and subsequent development of advisories concerning the potential health impacts of such consumption.

Table 7. Reported concentrations of mercury in seafood

Country

Common name of fish (where available)

Concentration of mercury (mean, or range reported, mg/kg)

No. of samples

Reference

Bangladesh

Three species

0.013–0.101

105

Joiris et al. (2000)

Brazil

Carnivorous and detritivorous fish, filter-feeding bivalves

0.01–0.19

291

Kehrig et al. (2001)

Croatia

Hake, sardine, mussel, bogue mackerel, anchovy

0.134–0.373

61

Juresa & Blanusa (2003)

Italy

Skates, rays, bonito, mackerel

0.07–1.56

1310

Storelli et al. (1998)

Italy

Swordfish, tuna

0.15–4.26

331

Storelli & Marcotrigiano (2001)

Italy

Megrim, sole, mullet, anglerfish

0.19–0.68

600

Storelli & Marcotrigiano (2000)

Italy

Tuna

1.17

288

Storelli et al. (2002)

Italy

Tuna, shark

0.38–6.5

NA

Storelli et al. (2001)

Kuwait

Seven species

0.073–3.27

330

Al-Majed & Preston (2000)

New Zealand

Marine fish

0.45

34

GEMS/Food

 

Tuna

0.21

10

 

Slovakia

Shark

0.48

15

GEMS/Food

 

Catfish

0.25

8

 

 

Mackerel

0.29

36

 

Tunisia

Sardines

0.19–0.42

271

Joiris et al. (1999)

USA

40 species

0.04–1.45

2663

Food & Drug Administration (2001)

NA, not applicable

3.4 Intake assessment

3.4.1. National estimates of intake

The Committee received intake assessments from France, Australia, Japan and New Zealand. Additionally, an assessment of exposure to methylmercury in the United States was published in 2002.

(a) France

The French assessment included concentrations of methylmercury for 92 seafood products. Concentrations below the limit of detection (9/2648 analyses) were assumed to contain methylmercury at the limit of detection (not reported). Data on fish consumption came from the 7-day French national food consumption survey, INCA, in 1999. Only consumers of one or more of the 92 seafood products were considered. The highest concentrations of methylmercury were found in dogfish (shark) (1.557 mg/kg). The median, mean, and 97.5th percentile intakes of methylmercury were 0.38, 0.63, and 2.75 µg/kg bw per week, respectively; these were calculated using a deterministic method, combining mean concentrations of methylmercury in fish with intakes of fish. A probabilistic approach using all of the available data yielded median, mean, and 97.5th percentile weekly intakes of 0.55, 1.34, and 7.54 µg/kg bw, respectively. When fish containing mercury at concentrations greater than the maximum national limits set (0.5 mg/kg for non-predatory species and 1.0 mg/kg for predatory species) are excluded from the probabilistic model, the resulting estimated exposures are reduced to 0.44, 0.93, and 4.8 µg/kg bw, respectively. The results from the probabilistic model suggest that intake for high consumers in France is at or above the PTWI, while the deterministic model does not result in exposures above the PTWI. The report concludes that the probabilistic method, which allows high concentrations of methylmercury to be combined with high intakes of fish, is more realistic than the deterministic method (INCA, 2003).

(b) New Zealand

The assessment by New Zealand combined concentrations of total mercury in foods surveyed for the New Zealand Total Diet studies (1990/1991 and 1997/1998) with data on food consumption taken from the 24-h recall New Zealand national nutrition survey (1997). More than 100 food classes were analysed for mercury. No differentiation between concentrations of total and methylmercury was reported. Consequently, it is not possible to directly calculate intake of methylmercury from the data submitted. Nonetheless, the report notes that marine fish was the major food contributing to intake of total mercury for the population aged > 15 years (76%), and for women aged 16–44 years (75.8%). If it is assumed that all of the mercury in marine fish is methylmercury and that all of the mercury in other foods is inorganic, the estimated mean intake of methylmercury would be 4 µg/person per day, with that at the 95th percentile being 14 µg/person per day (for persons aged > 15 years). The report notes that these are overestimates owing to the use of 1-day recall data on food intakes. The 95th percentile intake would be approximately 50% of the PTWI (Food Standards Australia New Zealand, 2003).

(c) Australia

The assessment by Australia was conducted in a manner parallel to that of New Zealand. Data on food intake were taken from the 24-h recall Australian national nutrition survey, in 1995. Mercury concentrations were determined for 90 food classes from the 1992 national residue survey, the "metals in meat" survey, and the Australian government analytical laboratories. Median concentrations were used "to represent the ‘most likely’ level of contamination in a positively skewed distribution of concentration data". As with the assessment from New Zealand, it is not possible to directly calculate the exposure to methylmercury because no differentiation between total mercury and methylmercury was reported. The Australian report also notes that marine fish was the major food contributing to intake of total mercury for the population aged > 2 years (69%), as well for women aged 16–44 years (65.5%). Making similar assumptions to those used above for the New Zealand assessment (that is, assuming that all of the mercury in marine fish is methylmercury and that all of the mercury in other foods is inorganic), mean intake of methylmercury for the population aged > 2 years could be estimated to be 6 µg/person per day, with that at the 95th percentile being 17 µg/person per day. The Australian report also notes that these are overestimates owing to the use of 1-day recall data on food intakes. The 95th percentile intake is approximately 60% of the PTWI (Food Standards Australia New Zealand, 2003).

(d) Japan

Results of a Japanese total diet study were submitted. For the most recent year for which data are available, 1999, the daily intake of total mercury in the diet was 9.7 µg/person per day. The daily dietary intake averaged over the years 1990–1999 was 8.6 µg/person per day (Japan, 2003).

(e) United States

Modelled long-term intake of methylmercury from the consumption of seafood was examined and compared with hair and blood biomarkers in the United States population. Information from dietary records derived from a 2-day survey was modified to reflect consumption over a year, assuming a range of allowable meal sizes. A range of concentrations of methylmercury in fish was combined with the modified daily intakes to arrive at an estimate of intake that was compared with model-derived estimates of intake using biomarkers. The mean, "eaters-only" estimated intake was 0.04 µg/kg bw per day for all ages and 0.08 µg/kg bw per day for children aged 2–5 years. The estimates derived from biomarkers were consistent with the estimates derived from modelled seafood intake for adults. The estimates derived from biomarkers for children were lower than those derived from survey data on seafood intake. Several possible model-related explanations were proposed to explain this difference in children, but it was not possible to determine conclusively the source of the discrepancy. (Carrington & Bolger, 2002)

The national estimates of exposure are summarized in Table 8.

Table 8. National estimates of intakes of methylmercury

 

Median

Mean

95th percentile

97.5th percentile

Australia

ND

6 µg/person per day
0.7 µg/kg bw per week

17 µg/person per day
1.95 µg/kg bw per week

ND

France

0.55 µg/kg bw per week
0.38 µg/kg bw per week (deterministic)

1.34 µg/kg bw per week
0.63 µg/kg bw per week (deterministic)

ND

7.54 µg/kg bw per weeka
2.75 µg/kg bw per week (deterministic)

Japan

ND

9.7 µg/person per day

ND

ND

New Zealand

ND

4 µg/person per day
0.5 µg/kg bw per week

14 µg/person per day
1.63 µg/kg bw per week

ND

a

When fish containing mercury at higher concentrations than the maximum residues levels for the European Union (0.5 ppm for non-predatory species, 1 ppm for predatory species) are excluded from the probabilistic model, this figure is reduced to 4.8 µg/kg bw.

ND, not determined

In addition to these assessments, submitted by Member countries in response to the call for data, the Committee also considered information on exposure to methylmercury published since the re-evaluation made by the Committee at its fifty-third meeting (Annex 1, references 143 and 144). In 2003, the results of the 20th Australian Total Diet Study were published. Intake of total mercury was reported to be as high as 0.09 µg/kg bw per day for people aged > 2 years. Exposure to methylmercury was not considered separately from exposure to total mercury, but it can be seen that using assumptions similar to those discussed for the Australian national intake estimate above yields consistent estimates of exposure to methylmercury for the Australian population (Australia, 2002).

In 2001, investigators reported on exposure to methylmercury from fish consumption and dental amalgams in a population of school age, in Portugal. The children reported eating fish weekly. The concentrations of methylmercury in the fish consumed, 14–24 µg/kg, were relatively low. Measures of blood and urine were taken at baseline, after placement of dental amalgams, and at a 1-year follow-up. The author concluded that dietary intake of methylmercury was not a significant source of exposure to mercury in this population. They also noted that concentrations of methylmercury in blood and hair in these children were much lower than those for the children followed in the Seychelles child development study or the study in the Faroe Islands (Evens et al., 2001).

Exposure analyses for three native American populations and two groups of recreational anglers were published in 2001. The participants’ intakes of fish were surveyed and levels of contamination with methylmercury were measured. More than 90% of one of the populations of recreational angler surveyed (freshwater fishermen) was found to eat fewer than 104 meals containing fish per year (<2 per week). Concentrations of methylmercury in all fish species consumed by this population were <0.5 mg/kg. Estimated intakes of methylmercury were all <0.08 µg/kg bw per day. Saltwater recreational anglers consumed fish species that generally contained methylmercury at <0.03 mg/kg and were also found to be exposed to methylmercury at <0.08 µg/kg bw per day. The three native American populations ate much more fish than the recreational anglers (>100 g/person per day). Within these populations, it was reported that consumption of chinook salmon resulted in methylmercury intakes as high as 0.17 µg/kg bw per day. This intake is approximately one-third of the PTWI (Marien & Patrick, 2001).

Investigators reported estimates of mercury intake for a cohort of individuals in the Basque Country in Spain, on the basis of data collected for the European Prospective Investigation into Cancer (EPIC). This population has a high consumption of fish, with a mean intake estimated at 606 g/week (87 g/day) and a 95th percentile consumption of fish of 1252 g/week (180 g/day). The average concentration of mercury in white and blue fish (504 samples) consumed by this population was 89.5 and 81.5 µg/kg, respectively. It was assumed that 90% of the mercury would be present in fish as methylmercury. The reported average intake of methylmercury for men was 48 µg/week (7 µg/day), which is approximately one-quarter of the PTWI (Sanzo et al., 2001).

Intake of methylmercury, as reflected by concentrations of mercury in hair, was examined in an indigenous population living in the Tapajos River basin in Brazil. This area is contaminated with mercury residues from nearby gold-mining operations. Additionally, this population has a high rate of consumption of fish. The fish eaten were found to contain mercury at mean concentrations ranging from 0.042 to 0.419 mg/kg. Although no acute intoxication with mercury was noted, the mean hair concentration of mercury in children aged 7–12 years was 14.45 µg/g and 15.7 µg/g in adult women aged 14–44 years. The authors concluded that this population is at risk of toxicity caused by exposure to mercury (Oliveira Santos et al., 2002).

A study of metal intakes, including mercury, in a population in Catalonia, Spain was published in 2003. This study used a total diet study approach to estimating intake of the metals. Diets were derived from food intake data gathered in the late 1990s. Analytical information on the levels of mercury in foods were used with the diets to estimate mean intake of mercury for "average" individuals in Catalonia in five population groups; children, adolescents, male and female adults, and older people. Mercury intakes ranged from 16.57 µg/day for children to 21.22 µg/day for adult males. No attempt to differentiate methylmercury from total mercury was reported. If it is assumed that only fish contribute methylmercury, and that 100% of the mercury in fish is present as methylmercury, methylmercury intakes would range from 5 µg/day for children to 9 µg/day for adult males. The estimate of intake for adult males is approximately 30% of the PTWI. The authors note that these levels are lower than previously reported for this population (Llobet et al., 2003).

3.4.2. International estimates of intake based on WHO GEMS/Food diets

The "Estimates of dietary intake" section of the monograph prepared by the Committee at its fifty-third meeting (Annex 1, reference 144) contains numerous analyses of methylmercury intake, using both national data and international data in the form of GEMS/Food estimates of food intakes (range of weekly intake for a 60-kg person, 0.3–1.5 µg/kg bw) as well as theoretical evaluations of intake scenarios. The data analysed by the Committee at its present meeting were not sufficiently different from those used at its fifty-third meeting to warrant amending their estimates. The current Committee affirmed the results reported at the fifty-third meeting.

4. COMMENTS

4.1 Absorption, metabolism and excretion

In all experimental animal species evaluated, methylmercury was readily absorbed (up to 95%) after oral exposure. Methylmercury crossed both the blood–brain barrier and the placenta effectively, resulting in higher concentrations of mercury in the brain of the fetus than of the mother. Methylmercury is eliminated mainly via the bile and faeces, neonatal animals having a lower excretory capacity than adults.

4.2 Toxicological studies

In its previous assessment, the Committee reviewed many experimental results that indicated that the developing nervous system, particularly in non-human primates, is a sensitive target for methylmercury. Experimental evidence indicates a possible protective effect of selenium against some toxic effects of methylmercury, but the results are conflicting.

4.2.1. Observations in animals

Ataxia, paralysis, loss of coordination, and hind limb crossing are common neurological signs of exposure to methylmercury in rodents. Changes in behaviour, decreased activity, and deficiencies in learning and memory have also been observed. In rodents, neurotoxic effects attributable to methylmercury usually become evident at doses that also affect other organ systems. The neurotoxic effects observed in non-human primates were consistent with the symptoms of Minamata disease, the syndrome observed in humans poisoned with methylmercury via consumption of contaminated seafood. The nature and severity of symptoms depend on dose and duration of exposure, and developmental stage. Exposure of neuroepithelial cells to methylmercury in vitro resulted in disruption of intracellular calcium homeostasis, induction of reactive oxygen species and oxidative DNA damage, and inhibition of axonal morphogenesis and cell cycle progression.

Treatment of pregnant female rodents with methylmercury induced abortion, increased the frequency of fetal resorption and malformations, and reduced offspring viability. Methylmercury also affected the rodent immune system, reducing mast cell function and, at high oral doses, decreasing spleen and thymus cell viability.

4.2.2. Observations in humans

At its fifty-third meeting, the Committee noted that methylmercury is toxic to the nervous system, kidney, liver and reproductive organs. At its present meeting, the Committee confirmed that neurotoxicity is the most sensitive end-point. In humans, the indices of neurotoxicity include neuronal loss, ataxia, visual disturbances, impaired hearing, paralysis and death. Both the central and peripheral nervous systems show signs of damage.

Information about the neurotoxicity caused by chronic fetal exposure to low doses of methylmercury has come primarily from epidemiological studies of populations in which fish consumption is frequent. The results of the neurodevelopmental assessments of 8-year-old children in the Seychelles study cohort were consistent with those obtained in this cohort previously, and provide no evidence for an inverse association between maternal exposure to methylmercury and neurodevelopmental effects in the children. Many of the neuropsychological test instruments included in the battery were the same as those used in the study in the Faroe Islands and which had been observed to be associated with biomarkers of prenatal exposure to methylmercury in 7-year-old children. Further analyses of the results of assessments of the Seychellois children at the age of 5.5 years have been published; these present alternative statistical approaches, adjustment for additional potential confounding factors, and more detailed evaluation of specific test scores. The results of these analyses do not alter the conclusion that in these populations with frequent fish consumption, no adverse effects have been detected that are attributable to prenatal exposure to methylmercury.

No new data were available from the main Faroe Islands study. Additional analyses of the assessments of the cohort at 7 years of age were carried out to explore the possibility of age- and test-dependent variation on susceptibility to methylmercury. Analyses were also conducted to determine whether the methylmercury-associated neuropsychological deficits observed in this cohort were attributable to episodes of higher exposure to methylmercury during pregnancy (associated with consumption of whale meat), residual confounding due to concomitant exposure to PCBs, and effects on children’s visual function. The results did not support a role for any of these factors in the positive associations observed in this study.

In a second smaller cohort (182 infants) assembled in the Faroe Islands, pre-natal exposure to methylmercury was found to be inversely related to newborn neurological status and to postnatal growth at 18 months of age. The association was still present after adjustment for exposure to 28 PCB congeners and 18 organochlorine pesticides or their metabolites.

A few new epidemiological studies of neurodevelopment have been reported, although these were cross-sectional rather than prospective in design, and involved much smaller sample sizes than either the Seychelles or Faroe Islands studies, and, in most cases, exposure to higher concentrations of methylmercury. A cross-sectional study of neurotoxic effects in adults reported significant mercury-associated neurobehavioural deficits in persons whose current concentration of hair mercury was <15 mg/kg. Because of the cross-sectional design of this study and because the concentration of mercury in an adult’s hair does not accurately reflect concentrations during the critical exposure period for neurodevelopment, the Committee considered that these results could not form the basis of a dose–response assessment.

Additional epidemiological studies have addressed issues such as reproductive toxicity, immunotoxicity, cardiotoxicity, and general medical status. With regard to reproductive toxicity, a methylmercury-associated decrease in the ratio of male : female births in the area of Minamata City during the period of peak pollution was reported, but the ratio subsequently returned to control levels. In a case–control study, higher concentrations of mercury were found in blood from couples experiencing fertility problems than from controls. With respect to cardiotoxicity, in a cohort study, concentrations of mercury in hair of > 2 mg/kg were associated with a doubling of the risk of suffering an acute myocardial infarction and, over a 4-year follow-up interval, with an increased risk for atherosclerotic disease. The results of two large case–control studies of mercury exposure and coronary heart disease were conflicting, one study reporting significantly higher concentrations of mercury in the toenails of cases than of controls, whereas the other reported similar concentrations in the two groups. In the latter study, half the participants were dentists and had concentrations of toenail mercury that were twice as high as those of non-dentists, suggesting that much of their exposure had been to metallic mercury rather than to methylmercury. In another study, high fish consumption, the primary route of exposure to methylmercury, was associated with an increased risk of stroke, but no biomarkers of mercury exposure were measured. The Committee determined that the available evidence for the potential cardiotoxicity of methylmercury was not conclusive, but noted that further studies were needed. With regard to general health status, the prevalence rates of liver disease, renal disease, and diabetes mellitus were not significantly increased in persons living near Minamata Bay, although the frequencies of many neurological and neuromuscular symptoms were higher.

4.3 Dose–response assessments

The Committee concluded that neurotoxic effects resulting from exposure to methylmercury in utero were the most sensitive health outcome. A number of dose–response assessments have been conducted using the results of the three major epidemiological studies of fetal neurotoxicity, conducted in the Faroe Islands, the Seychelles, and New Zealand. These assessments were made on the basis of evaluations of children at 7 years of age in the Faroe Islands study, 5.5 years of age in the Seychelles Islands study, and 6 years of age in the New Zealand study. A comprehensive dose–response assessment on the basis of the evaluations of the children in the Seychelles study at 8 years of age has not yet been reported, but the study results were similar to those obtained at 5.5 years of age. Mercury in maternal hair and/or umbilical cord blood served as the primary bio-markers of exposure to methylmercury in utero in the studies in the Faroe Islands and the Seychelles. After consideration of numerous publications, the Committee confirmed the validity of these biomarkers for both short-term (blood) and longer-term (hair) intake of methylmercury.

The concentration of mercury in maternal hair that corresponds to a no-observed-effect level (NOEL) for neurobehavioural effects was identified for the study in the Seychelles, and a mathematical analysis of the concentration– response relationship was used to determine a benchmark-dose lower-confidence limit (BMDL) for the studies in the Faroe Islands and New Zealand. The Committee noted that the concentration of mercury in maternal hair for one child (out of 237) in the study in New Zealand was 86 mg/kg, more than four times the next highest concentration in the study sample and had a heavy impact on the BMDLs. The inclusion of this observation produced BMDLs of 17–24 mg/kg, while omitting it produced BMDLs of 7.4–10 mg/kg. Because of uncertainty about which set of BMDLs was most valid, the Committee decided to base the evaluation only on the results of the studies in the Faroe Islands and the Seychelles (see Table 9). The Committee noted, however, that the inclusion of the results of the study in New Zealand did not materially alter its evaluation.

Table 9. Estimated concentration of mercury in maternal hair at NOEL and BMDL for neurotoxicity associated with exposure to methylmercury in utero

Study

n

NOEL/BMDL (mg/kg maternal hair)

References (17–22)

Faroe Islands

917

12.0

Budtz-Jorgensen et al. (1999a, 2000, 2001); United States National Research Council (2000); Rice et al. (2003)

Seychelles

711

15.3

Agency for Toxic Substances and Disease Registry (1999)

Average for two studies

Not applicable

14.0

The Committee used the average from the two studies, 14 mg/kg, as an estimate of the concentration of methylmercury in maternal hair that reflects exposures that would have no appreciable adverse effect on the offspring in these two study populations.

Calculation of steady-state ingestion of methylmercury (µg/kg bw per day) from the concentration of mercury in maternal hair comprises two steps: conversion of the concentration of methylmercury in maternal hair to that in maternal blood, and conversion of the concentration of mercury in maternal blood into maternal intake.

The mean ratio of the concentrations of methylmercury in hair to those in blood was determined in a number of studies, using samples from various study groups and with a variety of analytical methods, and was usually in the range of 140–370. The Committee used a value of 250 to represent the overall average ratio. The concentration of methylmercury in maternal blood that would be expected to have no appreciable adverse effects on the offspring was calculated to be 0.056 mg/l, determined by dividing aconcentration of mercury in maternal hair of 14 mg/kg by the hair : blood ratio of 250.

In humans, the steady state concentration of mercury in blood can be related to average daily intake using a one-compartment model that incorporates refinements (National Academy of Sciences/National Research Council, 2000) to the original WHO (1990) formula, as follows: where

C = mercury concentration in blood (µg/l)

b = elimination rate constant (0.014 per day-1)

V = blood volume (9% of body weight for a pregnant female)

A = fraction of the dose absorbed (0.95)

f = the absorbed fraction distributed to the blood (0.05)

bw = body weight (65 kg for a pregnant female)

d = dose (µg/kg bw per day)

The Committee used values appropriate to conversion during pregnancy, as the fetal period is considered to be the most vulnerable stage of life. Despite an elimination half-life for methylmercury of approximately 2 months, the maternal body burden at term is determined largely by intakes during the second and third trimesters of pregnancy.

Using this equation, the Committee determined that a steady-state ingestion of methylmercury at 1.5 µg/kg bw per day would result in a concentration of mercury in maternal blood that would have no appreciable adverse effects on offspring in these two study populations.

4.4 Estimated dietary intake

At its fifty-third meeting, the Committee re-evaluated the safety of methylmercury-contaminated foods, and fish in particular. The re-evaluation included consideration of information on potential intake submitted by numerous national bodies. For most populations, fish is the only significant source of methylmercury in food. Generally, concentrations of methylmercury are <0.4 mg/kg, but fish at the highest trophic levels may contain concentrations >5 mg/kg. Older and larger predatory fish species and certain marine mammals contain the highest concentrations of methylmercury.

At its current meeting, the Committee updated its evaluations of national intakes, adding intake information submitted by Australia, France, Japan, New Zealand, Slovakia and the USA, and use of biomarkers of exposure for methylmercury. The Committee also evaluated information published between 1997 and 2003 on concentrations of mercury and methylmercury in various fish species, as well as analyses of methylmercury intake by populations consuming large amounts of fish (>100 g per person per day). The Committee noted that overall methylmercury concentrations in fish species were similar to those considered at the fifty-third meeting and therefore concluded that the analyses of exposure conducted at the fifty-third meeting remained current. These estimates range from 0.3–1.5 µg/kg bw per week for the five regional GEMS/Food diets and from 0.1–2.0 µg/kg bw per week for numerous national diets.

5. EVALUATION

The Committee evaluated new information that had become available since methylmercury was considered at the fifty-third meeting. This information included the results of studies performed in laboratory animals and humans, and epidemiological studies of the possible effects of prenatal exposure to methylmercury on child neurodevelopment. Neurodevelopment was considered to be the most sensitive health outcome, and life in utero the most sensitive period of exposure.

The calculations made in the dose–response assessment are based on average values for each parameter, and did not allow for interindividual variability in either the hair : blood ratio or in the elimination rate constant in the equation shown above. Potential human variability was taken into account by the application of adjustment or uncertainty factors. In choosing the factors to apply to this intake estimate, the Committee considered the following:

A steady-state intake of methylmercury of 1.5 µg/kg bw per day was estimated to represent the exposure that would be expected to have no appreciable adverse effects on children. A total uncertainty factor of 6.4 (2 × 3.2) was applied to this figure to derive a PTWI of 1.6 µg/kg bw. This PTWI is considered sufficient to protect developing fetuses, the most sensitive subgroup of the population.

Pending reduction in the uncertainty associated with various aspects of the derivation of the steady-state intake from maternal concentrations of mercury in hair, the Committee concluded that the uncertainty factor could be refined and possibly reduced. The Committee also reaffirmed its position that fish are an important part of a balanced, nutritious diet and that this should be appropriately considered in public health decisions to set limits for methylmercury concentrations in fish. The Committee considered whether a provisional tolerable monthly intake PTMI rather than a PTWI for methylmercury should be established, but deferred its decision pending the outcome of the Joint FAO/WHO Project to Update the Principles and Methods for the Risk Assessment of Chemicals in Food.

At its sixteenth meeting (Annex 1, reference 30), the Committee established a PTWI for total mercury of 300 µg/person, of which no more than 200 µg should be present as methylmercury. This PTWI of 3.3 µg/kg bw for methylmercury was confirmed at subsequent meetings. Since the PTWI for methylmercury was revised at the current meeting, the Committee recommended that the PTWI for total mercury also be revised.

6. REFERENCES

Agency for Toxic Substances and Disease Registry (1999) Toxicological profile for mercury (update) (NTIS # PB/99/142416), Atlanta, GA.

Al-majed, N. B. & Preston, M.R. (2000) An assessment of total and methylmercury content of zooplankton and fish tissue collected from Kuwait territorial waters. Mar. Poll. Bull., 40, 298–307.

Axtell, C.D., Cox, C., Myers, G.J., Davidson, P.W., Choi, A.L., Cernichiari, E., Sloane-Reeves, J., Shamlaye, C.F. & Clarkson, T.W. (2000) Association between methylmercury exposure from fish consumption and child development at five and a half years of age in the Seychelles Child Development Study: an evaluation of nonlinear relationships. Environ. Res., 84, 71–80.

Baraldi, M., Zanoli, P., Tascedda, F., Blom, J.M. & Brunello, N. (2002) Cognitive deficits and changes in gene expression of NMDA receptors after prenatal methylmercury exposure. Environ. Health Perspect., 110(Suppl 5), 855–858.

Barnes, D., Daston, G., Evans, J., Jarabek, A., Kavlock, R., Kimmel, C., Park, C. & Spitzer, H. (1995) Benchmark dose workshop: criteria for use of a benchmark dose to estimate a reference dose. Regul. Toxicol. Pharmacol., 21, 296–306.

Batista, J., Schuhmacher, M., Domingo, J.L. & Corbella, J. (1996) Mercury in hair for a child population from Tarragona Province, Spain. Sci. Total Environ., 193,143–148.

Beck, B., Conolly, R., Dourson, M., Guth, D., Hattis, D., Kimmel, C. & Lewis, S. (1993) Improvements in quantitative noncancer risk assessment. Fund. Appl. Toxicol., 20, 1–14.

Becker, K., Schulz, C., Kaus, S., Seiwert, M. & Seifert, B. (2003) German Environmental Survey 1998 (GerES III): environmental pollutants in the urine of the German population. Int. J. Hyg. Environ. Health, 206, 15–24.

Belles, M., Albina, M.L., Sanchez, D.J., Corbella, J. & Domingo, J.L. (2002) Interactions in developmental toxicology: effects of concurrent exposure to lead, organic mercury, and arsenic in pregnant mice. Arch. Environ. Contam. Toxicol., 42, 93–98.

Belletti, S., Orlandini, G., Vettori, M.V., Mutti, A., Uggeri, J., Scandroglio, R., Alinovi, R. & Gatti, R. (2002) Time course assessment of methylmercury effects on C6 glioma cells: submicromolar concentrations induce oxidative DNA damage and apoptosis. 70, 703–711.

Bjornberg, K.A., Vahter, M., Petersson-Grawe, K., Glynn, A., Cnattingius, S., Darnerud, P.O., Atuma, S., Aune, M., Becker, W. & Berglund, M. (2003) Methyl mercury and inorganic mercury in Swedish pregnant women and in cord blood: influence of fish consumption. Environ. Health Perspect., 111, 637–641.

Budtz-Jorgensen, E., Keiding, N. & Grandjean, P. (1999a) Benchmark modeling of the Faroese methylmercury data. Final report to the United States Environmental Protection Agency, pp. 1–13.

Budtz-Jorgensen, E., Keiding, N., Grandjean, P. & White, R.F. (1999b) Methylmercury neurotoxicity independent of PCB exposure. Environ. Health Perspect., 107, A236–A237.

Budtz-Jorgensen, E., Grandjean, P., Keiding, N., White, R. & Weihe, P. (2000) Benchmark dose calculations of methylmercury-associated neurobehavioural deficits. Toxicol. Let., 112113, 193–199.

Budtz-Jorgensen, E., Keiding, N. & Grandjean, P. (2001) Benchmark dose calculation from epidemiological data. Biometrics, 57, 698–706.

Caicoya, M. (2002) Fish consumption and stroke: a community case–control study in Asturias, Spain. Neuroepidemiology, 21, 107–114.

Carrington, C. & Bolger, M. (2002) An exposure assessment for methylmercury from seafood for consumers in the United States. Risk Anal., 22, 701–712.

Chakrabarti, S. K. & Bai, C. (2000) Effects of protein-deficient nutrition during rat pregnancy and development on developmental hindlimb crossing due to methylmercury intoxication. Arch. Toxicol., 74, 196–202.

Choy, C.M.Y., Lam, C.W.K., Cheung, L.T.F., Briton-Jones, C.M., Cheung, L.P. & Haines, C.J. (2002) Infertility, blood mercury concentrations and dietary seafood consumption: a case–control study. B.J.O.G., 109, 1121–1125.

Chuu, J.J., Hsu, C.J. & Lin-Shiau, S.Y. (2001a) Abnormal auditory brainstem responses for mice treated with mercurial compounds: involvement of excessive nitric oxide. Toxicology, 162, 11–22.

Chuu, J.J., Liu, S.H. & Lin-Shiau, S.Y. (2001b) Effects of methyl mercury, mercuric sulfide and cinnabar on active avoidance responses, Na+/K+-ATPase activities and tissue mercury contents in rats. Proc. Natl. Sci. Counc. Repub. China B., 25, 128–36.

Clewell, H., Gearhart, J., Gentry, P., Covington, T., VanLandingham, C., Crump, K. & Shipp, A. (1999) Evaluation of the uncertainty in an oral reference dose for methylmercury dueto interindividual variability in pharmacokinetics. Risk Anal., 19, 547–558.

Coccini, T., Randine, G., Candura, S.M., Nappi, R.E., Prockop, L.D. & Manzo, L. (2000) Low-level exposure to methylmercury modifies muscarinic cholinergic receptor binding characteristics in rat brain and lymphocytes: physiologic implications and new opportunities in biologic monitoring. Environ. Health Perspect., 108, 29–33.

Cordier, S., Garel, M., Mandereau, L., Morcel, H., Doineau, P., Gosme-Seguret, S., Josse, D., White, R. & Amiel-Tison, C. (2002) Neurodevelopmental investigations among methylmercury-exposed children in French Guiana. Environ. Res., 89, 1–11.

Counter, A.S., Buchanan, L.A., Ortega, F. & Laurell, G. (2002) Elevated blood mercury and neurootological observations in children of the Ecuadorean gold mines. J. Toxicol. Environ. Health, 65, 1490–1163.

Cox, C., Breazna, A., Davidson, P.W., Myers, G.J. & Clarkson, T.W. (1999) Prenatal and post-natal methylmercury exposure and neurodevelopmental outcomes. J.A.M.A., 282, 1333–1334.

Crompton, P., Ventura, A.M., de Souza, M., Santos, E., Strickland, G.T. & Silbergeld, E. (2002) Assessemnt of mercury exposure and malaria in a Braziliam Amazon Riverine community. Environ. Res., 90, 69–75.

Crump, K. (1984) A new method for determining allowable daily intakes. Fund. Appl. Toxicol., 4, 854–871.

Crump, K., Viren, J., Silvers, A., Clewell, H., Gearhart, J. & Shipp, A. (1995) Reanalysis of dose–response data from the Iraqi methylmercury poisoning episode. Risk Anal., 15, 523–532.

Crump, K., Kjellstrom, T., Shipp, A. & Stewart, A. (1998) Influence of prenatal mercury exposure upon scholastic and psychological test performance: benchmark analysis of a New Zealand cohort. Risk Anal., 18, 701–713.

Crump, K., Landingham, C., Shamlaye, C., Cox, C., Davidson, P., Myers, G. & Clarkson, T. (2000) Benchmark concentrations for methylmercury obtained from the Seychelles child development study. Environ. Health Perspect., 108, 257–263.

Davidson, P.W. Myers, G.J., Cox, C., Axtell, C., Shamlaye, C., Sloane-Reeves, J., Cernichiari, E., Needham, L.L., Choi, A., Wang, Y., Berlin, M. & Clarkson, T.W. (1998) Effects of pre-natal and postnatal methylmercury exposure from fish consumption on neurodevelopmental outcomes at 66 months of age in the Seychelles Child Development Study. J.A.M.A., 280, 701–707.

Davidson, P.W., Myers, G.J., Shamlaye, C., Cox, C., Gao, P., Axtell, C., Morris, D., Sloane-Reeves, J., Cernichiari, E., Choi, A., Palumbo, D. & Clarkson, T.W. (1999) Association between prenatal exposure to methylmercury and developmental outcomes in Seychellois children: effect modification by social and environmental factors. Neurotoxicology, 20, 833–841.

Davidson, P.W., Kost, J., Myers, G.J., Cox, C. & Clarkson, T.W. (2001) Methylmercury and neurodevelopment: reanalysis of the Seychelles Child Development Study outcomes at 66 months of age. J.A.M.A., 285, 1291–1293.

Davidson, P.W., Palumbo, D., Myers, G.J., Cox, C., Shamlaye, C. F., Sloane-Reeves, J., Cernichiari, E., Wilding, G.E. & Clarkson, T.W. (2000) Neurodevelopmental outcomes of Seychellois children from the pilot cohort at 108 months following prenatal exposure to methylmercury from a maternal fish diet. Environ. Res., 84, 1–11.

Dey, P.M., Gochfeld, M. & Reuhl, K.R. (1999) Developmental methylmercury administration alters cerebellar PSA-NCAM expression and Golgi sialyltransferase activity. Brain Res., 845, 139–151.

Dourson, M., Wullenweber, A. & Poirier, K. (2001) Uncertainties in the reference dose for methylmercury. Neurotoxicology, 22, 677–689.

Endo, T., Haraguchi K. & Sakata, M. (2003) Renal toxicity in rats after oral administration of mercury-contaminated boiled whale livers marketed for human consumption. Arch. Environ. Contam. Toxicol., 44, 412–416.

Eto, K., Yasutake, A., Korogi, Y., Akima, M., Shimozeki, T., Tokunaga, H., Kuwana, T. & Kaneko, Y. (2002) Methylmercury poisoning in common marmosets–MRI findings and peripheral nerve lesions. Toxicol. Pathol., 30, 723–734.

Eto, K., Yasutake, A., Kuwana, T., Korogi, Y., Akima, M., Shimozeki, T., Tokunaga, H. & Kaneko, Y. (2001) Methylmercury poisoning in common marmosets–a study of selective vulnerability within the cerebral cortex. Toxicol. Pathol., 29, 565–573.

Evens, C.C., Martin, M.D., Woods, J.S., Soares, H.L., Bernardo, M., Leitao, J., Simmonds, P.L., Liang, L. & DeRouen, T. (2001) Examination of dietary methylmercury exposure in the Casa Pia study of the health effects of dental amalgams in children. J. Toxicol. Environ. Health, Part A, 64, 521–530.

Faro, L.R., do Nascimento, J.L., Alfonso, M. & Duran, R. (2002) Mechanism of action of methylmercury on in vivo striatal dopamine release. Possible involvement of dopamine transporter. Neurochem. Int., 40, 455–465.

Faustman, E.M., Ponce, R.A., Ou, Y.C., Mendoza, M.A., Lewandowski, T. & Kavanagh, T. (2002) Investigations of methylmercury-induced alterations in neurogenesis. Environ. Health Perspect., 110(Suppl. 5), 859–864.

Food and Drug Administration (2001) Mercury levels in seafood species (http://www.cfsan.fda.gov/~frf/sea-mehg.html).

Feng, Q., Suzuki, Y. & Hisashige, A. (1998) Hair mercury levels of residents in China, Indonesia, and Japan. Arch. Environ. Health, 53, 36–43.

Fredriksson, A., Gardlund, A.T., Bergman, K., Oskarsson, A., Ohlin, B., Danielsson, B. & Archer, T. (1993) Effects of maternal dietary supplementation with selenite on the post-natal development of rat offspring exposed to methyl mercury in utero. Pharmacol. Toxicol. 72, 377–382.

Fukuda, Y., Ushijima, K., Kitano, T., Sakamoto, M. & Futatsuka, M. (1999) An analysis of subjective complaints in a population living in a methylmercury-polluted area. Environ. Res., 81, 100–107.

Futatsuka, M., Kitano, T., Shono, M., Fukuda, Y., Ushijima, K., Inaoka, T., Nagano, M., Wakamiya, J. & Miyamoto, K. (2000) Health surveillance in the population living in a methyl mercury-polluted area over a long period. Environ. Res., 83, 83–92.

Gerstenberger, S.L., Tavris, D.R., Hansen, L.K., Pratt-Shelley, J. & Dellinger, J.A. (1997) Concentrations of blood and hair mercury and serum PCB’s in an Ojibwa population that consumes Great Lakes region fish. Clin. Toxicol., 35, 377–386.

Graevskaya, E.E., Yasutake, A., Aramaki, R. & Rubin, A.B. (2003) Effect of methylmercury on histamine release from rat mast cells. Arch. Toxicol., 77, 17–21.

Grandjean, P., Budtz-Jorgensen, E., White, R.F., Jorgensen, P.J., Weihe, P., Debes, F. & Keiding, N. (1999) Methylmercury exposure biomarkers as indicators of neurotoxicity in children aged 7 years. Am. J. Epidemiol, 149, 301–305.

Grandjean, P., Weihe, P., Burse, V.W., Needham, L.L., Storr-Hansen, E., Heinzow, B., Debes, F., Murata, K., Simonsen, H., Ellefsen, P., Budtz-Jorgensen, E., Keiding, N. & White, R.F. (2001a) Neurobehavioral deficits associated with PCB in 7-year-old children prenatally exposed to seafood neurotoxicants. Neurotoxicol. Teratol., 23, 305–317.

Grandjean, P., White, R.F., Sullivan, K., Debes, F., Murata, K., Otto, D.A. & Weihe, P. (2001b) Impact of contrast sensitivity performance on visually presented neurobehavioral tests in mercury-exposed children. Neurotoxicol. Teratol., 23, 141–146.

Grandjean, P., Budtz-Jorgensen, E., Steuerwald, U., Heinzow, B., Needham, L.L., Jorgensen, P.J. & Weihe, P. (2003a) Attenuated growth of breast-fed children exposed to increased concentrations of methylmercury and polychlorinated biphenyls. FASEB J., 17, 699– 701.

Grandjean, P., White, R.F., Weihe, P. & Jorgensen, P.J. (2003b) Neurotoxicity risk by stable and variable exposure to methylmercury from seafood. Ambul. Pediatr., 3, 18–23.

Guilar, E., Sanz-Gallardo, I., Van’t Veer, P., Bode, P., Aro, A., Gomez-Aracena, J., Kark, J.D., Riemersma, R.A., Martin-Moreno, J.M. & Kok, F.J. (2002) Mercury, fish oils, and the risk of myocardial infarction. N. Engl. J. Med., 347, 1747–1754.

Hansen, J.C., Tarp, U. & Bohm, J. (1990) Prenatal exposure to methylmercury among Greenlandic polar Inuits. Arch. Environ. Health, 45, 355–358.

Heidemann, S.R., Lamoureux, P. & Atchison, W.D. (2001) Inhibition of axonal morphogenesis by nonlethal, submicromolar concentrations of methylmercury. Toxicol. Appl. Pharmacol., 174, 49–59.

Hightower, J.M. & Moore, D. (2003) Mercury levels in high-end consumers of fish. Environ. Health Perspect., 111, 604–608.

Homma-Takeda, S., Kugenuma, Y., Iwamuro, T., Kumagai, Y. & Shimojo, N. (2001) Impairment of spermatogenesis in rats by methylmercury: involvement of stage- and cell-specific germ cell apoptosis. Toxicology, 169, 25–35.

Hultman, P. & Hansson-Georgiadis, H. (1999) Methyl mercury-induced autoimmunity in mice. Toxicol Appl Pharmacol., 154, 203–211.

Johansson, N., Basun, H., Winblad, B. & Nordberg, M. (2002) Relationship between mercury concentration in blood, cognitive performance, and blood pressure in an elderly urban population. Biometals, 15, 189–195.

Joiris, C.R., Das, H.K. & Holsbeek, L. (2000) Mercury accumulation and speciation in marine fish from Bangladesh. Mar. Poll. Bull., 40,454–457.

Joiris, C.R., Holsbeek, L. & Laroussi Moatemri, N. (1999) Total and methylmercury in sardines Sardinella aurita and Sardina pilchardus from Tunisia. Mar. Poll. Bull., 38, 188–192.

Juarez, B.I., Martinez, M.L., Montante, M., Dufour, L., Garcia, E. & Jimenez-Capdeville, M.E. (2002) Methylmercury increases glutamate extracellular levels in frontal cortex of awake rats. Neurotoxicol. Teratol., 24, 767–771.

Juresa, D. & Blanusa, M., (2003) Mercury, arsenic, lead and cadmium in fish and shellfish from the Adriatic Sea. Food Addit. Contam., 20, 241–246.

Kakita, A., Wakabayashi, K., Su, M., Sakamoto, M., Ikuta, F. & Takahashi, H. (2000a) Distinct pattern of neuronal degeneration in the fetal rat brain induced by consecutive transplacental administration of methylmercury. Brain Res., 859, 233–239.

Kakita, A., Wakabayashi, K., Su, M., Yoneoka, Y., Sakamoto, M., Ikuta, F. & Takahashi, H. (2000b) Intrauterine methylmercury intoxication. Consequence of the inherent brain lesions and cognitive dysfunction in maturity. Brain Res., 877, 322–330.

Kakita, A., Wakabayashi, K., Su, M., Piao, Y.S. & Takahashi, H. (2001) Experimentally induced leptomeningeal glioneuronal heterotopia and underlying cortical dysplasia of the lateral limbic area in rats treated transplacentally with methylmercury. J. Neuropathol. Exp. Neurol., 60, 768–777.

Kakita, A., Inenaga, C., Sakamoto, M. & Takahashi, H. (2002) Neuronal migration disturbance and consequent cytoarchitecture in the cerebral cortex following transplacental administration of methylmercury. Acta. Neuropathol. (Berl). 104, 409–417.

Kehrig, H. do A., Costa, M., Moreira, I. & Malm, O. (2001) Methylmercury and total mercury in estuarine organisms from Rio de Janeiro, Brazil. Environ. Sci. Poll. Res., 8, 275–279.

Kim, C.Y., Nakai, K., Kameo, S., Kurokawa, N., Liu, Z.M. & Satoh, H. (2000a) Protective effect of melatonin on methylmercury-induced mortality in mice. Tohoku. J. Exp. Med., 191, 241–246.

Kim, C.Y., Nakai, K., Kasanuma, Y. & Satoh, H. (2000b) Comparison of neurobehavioral changes in three inbred strains of mice prenatally exposed to methylmercury. Neurotoxicol. Teratol., 22, 397–403.

King, M.D., Lindsay, D.S., Holladay, S. & Ehrich, M. (2003) Neurotoxicity and immunotoxicity assessment in CBA/J mice with chronic Toxoplasma gondii infection and single-dose exposure to methylmercury. Int. J. Toxicol., 22, 53–61.

Lewandowski, T.A., Pierce, C.H., Pingree, S.D., Hong, S. & Faustman, E.M. (2002) Methylmercury distribution in the pregnant rat and embryo during early midbrain organogenesis. Teratology, 66, 235–241.

Lindlow, S.W., Knight, R., Batty, J. & Haswell, S.J. (2003) Maternal and neonatal hair mercury concentrations: the effect of dental amalgam. B.J.O.G., 110, 287–291.

Llobet, J.M., Falco, G., Casas, C., Teixido, A. & Domingo, J.L. (2003) Concentrations of arsenic, cadmium, mercury, and lead in common foods and estimated daily intake by children, adolescents, adults, and seniors of Catalonia, Spain. J. Agric. Food Chem., 51, 838–842.

Lu, F. & Sielken, R. (1991) Assessment of safety/rsik of chemicals: inception and evolution of the ADI and dose–response modeling procedures. Toxicol. Let., 59, 5–40.

Marien, K. & Patrick, G.M. (2001) Exposure analysis of five fish-consuming populations for overexposure to methylmercury. J. Exp. Anal. Environ. Epidemiol., 11, 193–206.

Miyamoto, K., Nakanishi, H., Moriguchi, S., Fukuyama, N., Eto, K., Wakamiya, J., Murao, K., Arimura, K. & Osame, M. (2001) Involvement of enhanced sensitivity of N-methyl-D-aspartate receptors in vulnerability of developing cortical neurons to methylmercury neurotoxicity. Brain Res., 901, 252–258.

Muckle, G., Ayotte, P., Dewailly, E., Jacobson, S.W. & Jacobson, J.L. (2001) Determinants of polychlorinated biphenyls and methylmercury exposure in Inuit women of childbearing age. Environ.Health Perspect., 109, 957–963.

Murata, K., Budtz-Jorgensen, E. & Grandjean, P. (2002) Benchmark dose calculations for methylmercury-associated delays on evoked potential latencies in two cohorts of children. Risk Anal., 22, 465–474.

Myers, G.J., Davidson, P.W., Cox, C., Shamlaye, C.F., Palumbo, D., Cernichiari, E., Sloane-Reeves, J., Wilding, G., Kost, J., Huang, L.-S. & Clarkson, T.W. (2003) Prenatal methylmercury exposure from ocean fish consumptions in the Seychelles child development study. Lancet, 361, 1686–1692.

Myers, G.J., Davidson, P.W., Palumbo, D., Shamlaye, C., Cox, C., Cernichiari, E. & Clarkson, T.W. (2000) Secondary analyses from the Seychelles Child Development Study: the Child Behavior Checklist. Environ. Res., 84, 12–19.

Nakagawa, R., Yumita, Y. & Hiromoto, M. (1997) Total mercury intake from fish and shellfish by Japanese people. Chemosphere, 35, 2909–2913.

National Academy of Sciences/National Research Council (2000) Toxicological effects of methylmercury. National Academy Press, Washington, DC.

Newland, M.C. & Rasmussen, E.B. (2000) Aging unmasks adverse effects of gestational exposure to methylmercury in rats. Neurotoxicol. Teratol., 22, 819–28.

Newland, M.C. & Reile, P.A. (1999) Blood and brain mercury levels after chronic gestational exposure to methylmercury in rats. Toxicol. Sci., 50, 106–116.

Newland, M.C., Yezhou, S., Lodgdberg, B. & Berlin, M. (1994) Prolonged behavioral effects of in utero exposure to lead and methylmercury reduced sensitivity to changes in reinforcement contingencies during behavioral transitions and steady state. Toxicol. Appl. Pharmacol., 126, 6–15.

NHANES (2001) Blood and hair mercury levels in young children and women of childbearing age—United States, 1999. M.M.W.R.Morb. Mortal. Wkly. Rep., 50, 140–143.

Ninomiya, T., Imamura, K., Kuwahata, M., Susa, M. & Ekino, S. (submitted) Chronic methylmercury poisoning damages somatosensory cortex.

Ochi, T. (2002) Methylmercury, but not inorganic mercury, causes abnormality of centrosome integrity (multiple foci of gamma-tubulin), multipolar spindles and multinucleated cells without microtubule disruption in cultured Chinese hamster V79 cells. Toxicology., 175, 111–121.

O’Hara, M.F., Charlap, J.H., Craig, R.C. & Knudsen, T.B. (2002) Mitochondrial transduction of ocular teratogenesis during methylmercury exposure. Teratology., 65, 131–44.

Oka, T., Matsukura, M., Okamoto, M., Harada, N., Kitano, T., Miike, T. & Futatsuka, M. (2003) Autonomic nervous functions in fetal type Minamata Disease patients: assessment of heart rate variability. Tohoku J. Exp. Med., 198, 215–221.

Oliveira Santos, E.C. de, Jesus, I.M. de, M. Camara, V. de, Brabo, E., Loureiro E.C.B., Mascarenhas, A., Weirich, J., Luiz, R.R. & Cleary, D. (2002) Mercury exposure in Munduruku Indians from the community of Sai Cinza, State of Para, Brazil. Environ. Res., 90, 98–103.

Palumbo, D.R., Cox, C., Davidson, P.W., Myers, G.J., Choi, A., Shamlaye, C., Sloane-Reeves, J., Cernichiari, E. & Clarkson, T.W. (2000) Association between prenatal exposure to methylmercury and cognitive functioning in Seychellois children: a reanalysis of the McCarthy scales of children’s ability from the main cohort study. Environ. Res., 84, 81–88.

Pellizzari, E.D., Fernando, R., Cramer, G.M., Meaburn, G.M. & Bangerter, K. (1999) Analysis of mercury in hair of Region V population. J. Exp. Anal. Environ. Epidemiol., 9, 393–401.

Pesch, A., Wilhelm, M., Rostek, U., Schmitz, N., Weishoff-Houben, M., Ranft, U., & Idel, H. (2002) Mercury concentrations in urine, scalp hair, and saliva in children from Germany. J. Exp. Anal. Environ. Epidemiol., 12,252–258.

Rasmussen, E.B. & Newland, M.C. (2001) Developmental exposure to methylmercury alters behavioral sensitivity to D-amphetamine and pentobarbital in adult rats. Neurotoxicol. Teratol., 23, 45–55.

Rice, D.C. & Hayward, S. (1999) Comparison of visual function at adulthood and during aging in monkeys exposed to lead or methylmercury. Neurotoxicology, 20, 767–784.

Rice, D., Schoeny, R. & Mahaffey, K. (2003) Methods and rationale for the derivation of a reference dose for methylmercury by the U.S. EPA. Risk Anal., 23, 107–115.

Sakamoto, M., Kakita, A., Wakabayashi, K., Takahashi, H., Nakano, A. & Akagi, H. (2002) Evaluation of changes in methylmercury accumulation in the developing rat brain and its effects: a study with consecutive and moderate dose exposure throughout gestation and lactation periods. Brain Res., 949, 51–59.

Sakamoto, M., Nakano, A. & Akagi, H. (2001) Declining Minamata male birth ratio associated with increased male fetal death due to heavy methylmercury pollution. Environ. Res., 87, 92–98.

Salonen, J.T., Seppanen, K., Nyyssonen, K., Korpela, H., Kauhanen, J., Kantola, M., Tuomilehto, J., Esterbauer, H., Tatzber, F. & Salonen, R. (1995) Intake of mercury from fish, lipid peroxidation, and the risk of myocardial infarction and coronary cardiovascular, and any death in eastern Finnish men. Circulation, 91, 645–655.

Salonen, J.T., Seppanen, K., Lakka, T.A., Salonen, R. & Kaplan, G.A. (2000) Mercury accumulation and accelerated progression of carotid atherosclerosis: a population-based prospective 4-year follow-up study in men in eastern Findland. Atherosclerosis, 148, 265–273.

Sanzo, J.M., Dorronsoro, M., Amiamo, P., Aguinagalde, F.X., Azpiri, M.A. & the EPIC Group of Spain (2001) Estimation and validation of mercury intake associated with fish consumption in and EPIC cohort of Spain. Public Health Nutr., 4, 981–988.

Shafer, T.J., Meacham, C.A. & Barone, S. Jr. (2002) Effects of prolonged exposure to nanomolar concentrations of methylmercury on voltage-sensitive sodium and calcium currents in PC12 cells. Dev. Brain Res., 136, 151–64.

Shanker, G., Allen, J.W., Mutkus, L.A. & Aschner, M. (2001) Methylmercury inhibits cysteine uptake in cultured primary astrocytes, but not in neurons. Brain Res., 914, 159–65.

Smith, J.C., Allen, P.V. & Von Burg, R. (1997) Hair methylmercury levels in U.S. women. Arch. Environ. Health, 52, 476–480.

Stern, S., Cox, C., Cernichiari, E., Balys, M. & Weiss, B. (2001) Perinatal and lifetime exposure to methylmercury in the mouse: blood and brain concentrations of mercury to 26 months of age. Neurotoxicology, 22, 467–477.

Steuerwald, U., Weihe, P., Jorgensen, P.J., Bjerve, K., Brock, J., Heinzow, B., Budtz-Jorgensen, E. & Grandjean, P. (2000) Maternal seafood diet, methylmercury exposure, and neonatal neurologic function. J. Pedriatr., 136, 599–605.

Stewart, P.W., Reihman, J., Lonky, E.I., Darvill, T.J. & Pagano, J. (2003) Cognitive development in preschool children prenatally exposed to PCBs and MeHg. Neurotoxicol. Teratol., 25, 11–22.

Storelli, M.M. & Marcotrigiano, G.O. (2000) Fish for human consumption: risk of contamination by mercury. Food Addit. Contam., 17, 1007–1011.

Storelli, M.M. & Marcotrigiano, G.O. (2001) Total mercury levels in muscle tissue of sword-fish (Xiphias gladius) and bluefin tuna (Thunnus thynnus) from the Mediterranean Sea (Italy). J. Food Prot., 64, 1058–1061.

Storelli, M.M., Giacominelli Stuffler, R. & Marcotrigiano, G.O. (1998) Total mercury in muscle of benthic and pelagic fish from the South Adriatic Sea (Italy). Food Addit. Contam., 15, 876–883.

Storelli, M.M., Giacominelli Stuffler, R. & Marcotrigiano, G.O. (2001) Total mercury and methylmercury in tunja fish and sharks from the South Adriatic Sea. Int. J. Food Sci., 13, 101–106.

Storelli, M.M., Giacominelli Stuffler, R. & Marcotrigiano, G.O. (2002) Total and methylmercury residues in tuna-fish from the Mediterranean Sea. Food Addit. Contam., 19, 715–720.

Szasz, A., Barna, B., Galbacs, Z., Kirsch-Volders, M. & Szente, M. (2000) Effect of chronic low-dose developmental methylmercury intoxication on epileptogenicity in rats. Cent. Eur. J. Public Health, 8(Suppl.), 75–77.

Vilagi, I., Doczi, J. & Banczerowski-Pelyhe, I. (2000) Altered electrophysiological characteristics of developing rat cortical neurones after chronic methylmercury chloride treatment. Int. J. Dev. Neurosci. 18, 493–499.

Watanabe, C. & Satoh, H. (1994) Brain selenium status and behavioral development in selenium-deficient preweanling mice. Physiol. Behav., 56, 927–932.

Watanabe, C., Yin, K., Kasanuma, Y. & Satoh, H. (1999a). In utero exposure to methylmercury and Se deficiency converge on the neurobehavioral outcome in mice. Neurotoxicol. Teratol., 21, 83–88.

Watanabe, C., Yoshida, K., Kasanuma, Y., Kun, Y. & Satoh, H. (1999b) In utero methylmercury exposure differentially affects the activities of selenoenzymes in the fetal mouse brain. Environ Res. 80, 208–214.

Weihe, P., Hansen, J.C., Murata, K., Debes, F., Jorgensen, P.J., Steuerwald, U., White, R.F. & Grandjean, P. (2002) Neurobehavioral performance of Inuit children with increased prenatal exposure to methylmercury. Int. J. Circumpolar Health, 61, 41–49.

WHO (1976) Mercury (Environmental Health Criteria No. 1), Geneva.

WHO (1987) Principles for the safety assessment of food additives and contaminants in food (Environmental Health Criteria 70), Geneva

WHO (1990) Methylmercury (Environmental Health Criteria 101). Geneva.

Yasutake, A. & Hirayama, K. (2001) Evaluation of methylmercury biotransformation using rat liver slices. Arch. Toxicol., 75, 400–406.

Yasutake, A., Matsumoto, M., Yamaguchi, M. & Hachiya, N. (2003) Current hair mercury levels in Japanese: survey in five districts. Tohoku J. Exp. Med., 199, 161–169.

Yokoo, E.M., Valente, J.G., Grattan, L., Schmidt, S.L., Platt, I. & Silbergeld, E.K. (2003) Low-level methyl mercury exposure affects neuropsychological function in adults. Environ. Health, 2, 8.

Yoshizawa, K., Rimm, E.B., Morris, J.S., Spate, V.L., Hsieh, C.-C., Spiegelman, D., Stampfer, M.J. & Willett, W.C. (2002) Mercury and the risk of coronary heart disease in men. N. Engl. J. Med., 347, 1755–1760.

Young, J.F., Wosilait, W.D. & Luecke, R.H. (2001) Analysis of methylmercury disposition in humans utilizing a PBPK model and animal pharmacokinetic data. J. Toxicol. Environ. Health, 63, 19–52.

Zanoli, P., Cannazza, G. & Baraldi, M. (2001) Prenatal exposure to methyl mercury in rats: focus on changes in kynurenine pathway. Brain Res. Bull., 55, 235–238.



    See Also:
       Toxicological Abbreviations
       Methylmercury (EHC 101, 1990)
       Methylmercury (WHO Food Additives Series 24)
       Methylmercury (WHO Food Additives Series 44)
       METHYLMERCURY (JECFA Evaluation)