Fetal Methylmercury Poisoning: Clinical and Toxicological Data on 29 Cases

David 0. Marsh, MD, Gary J. Myers, MD, Thomas W. Clarkson, PhD, Laman Amin-Zaki, MD, Sadoun Tikriti, MD, and Muhammed A. Majeed, M D

This report describes psychomotor retardation in infants caused by prenatal exposure to methylmercury. A study of 29 mother-infant pairs established a relationship between maximum maternal hair mercury concentration during pregnancy and the frequency of neurological effects in the infants. The latter included delayed achievement of developmental milestones with or without neurological signs. The infants were 4% to 5 years of age on last exam- ination. Ten infants of mothers who had maximum hair concentrations in the range of 99 to 384 parts per million (ppm) had a significantly higher frequency of abnormal findings than those in two groups having lower maternal hair mercury concentrations (12 to 85 and 0 to 11 ppm). The group sizes were too small to allow identification of a specific “threshold” maternal hair concentration above which such effects can be expected. Transient paresthesias during pregnancy occurred in 80% of the mothers in the higher concentration group (99 to 384 ppm) as compared with 30% and 22% in the lower groups. Neurological abnormalities were found in some children whose mothers had been asymptomatic during pregnancy.

Marsh DO, Myers GJ, Clarkson TW, et al: Fetal methylmercury poisoning: clinical and toxicological data on 29 cases. Ann Neurol 7~348-353, 1980

Methylmercury (MeHg) was first recognized as a hazard to industrial workers in 1863, and its effects on the central nervous system were described by Hunter, Bomford, and Russell in 1940 [12]. In the past 25 years, extensive outbreaks of MeHg poison- ing have occurred in populations that were not ex- posed occupationally, and the importance of MeHg as an environmental contaminant has become recog- nized. The largest outbreak has occurred in Iraq [3, 171 from consumption of bread prepared from MeHg-treated seed grain. Other cases of alkylmer- cury poisoning from ingestion of treated seed grain have been reported from Iraq [6, 131, Pakistan [lo], Guatemala [15], and Ghana [7]. Two outbreaks have occurred in Japan due to consumption of MeHg- contaminated fish [4, 211.

The earliest report of psychomotor retardation caused by fetal exposure to MeHg was by Engleson and Herner in 1952 [8]. A Swedish family had eaten porridge made from MeHg-treated grain. The asymptomatic mother gave birth to a daughter who appeared to be normal at birth and in the first two months of life, but later showed mental and physical retardation. The authors postulated that “mercury intoxication, perhaps during early fetal life, seems . . . a possible cause.” Her father and brother were diag- nosed as having mercury poisoning. Urinary mercury

concentrations were elevated in the mother; no blood or hair analyses were performed.

Harada [ 1 11 reported 22 children from Minamata, Japan, with severe psychomotor retardation appar- ently due to fetal MeHg poisoning. All were from families in which at least 1 other member had been diagnosed as having MeHg poisoning, with fatal re- sults in 13 families. Five of the mothers had experi- enced transient paresthesias during pregnancy but had been well otherwise. The children’s ages ranged from 13 to 73 months at the time of initial examina- tion, and it was not possible to determine the degree of antenatal exposure to MeHg. Two of the children died, and neuropathological studies disclosed evi- dence of disturbed brain development, with cerebral and cerebellar lesions similar to those found in kit- tens exposed to MeHg in utero [20]. By 1975 a total of 899 cases of MeHg poisoning had been diagnosed in and around Minamata, including 23 severe fetal cases but no mild fetal cases. By the end of 1974 there were 520 officially recognized cases in Niigata [2 11, including only 1 fetal case. The only published fetal case of MeHg toxicity in the United States [16, 181 was due to the mother eating MeHg-con- taminated pork.

Fifteen Iraqi mother-infant pairs who had been ex- posed to MeHg during pregnancy were studied by

From the Department of Neurology, Pediatrics and Environment Health Sciences Center. University of Rochester School of Medi- cine, Rochester, NY, and the College of Medicine and Directorate of Preventive Medicine, Baghdad, Iraq.

Accepted for publication Sept 5 , 1979. Address reprint requests Dr 601 Elmwood Ave, Rochester, NY 14642,

D~~~~~~~~ of ~ ~ ~ ~ ~ l ~ ~ ,

348 0364-5134/80/040348-06$01.25 @ 1979 by David 0. Marsh

Amin-Zaki et a1 [l]. Clinical features of MeHg in- toxication were found in 6 mothers and 6 infants; in only 1 mother-infant pair was the infant clinically af- fected and the mother asymptomatic. Four of the in- fants were blind and had abnormal limb tone.

The major outbreak of MeHg poisoning in Iraq re- sulted when more than 90,000 tons of MeHg-treated seed grain were imported into the country during the winter of 1971-72 and delivered to every province. The grain was dyed red and the grain sacks carried a warning, written in Spanish, against human or animal consumption. In many rural areas the seed grain was washed and ground and the flour used to bake bread, which was consumed over a period of three to four months. The resultant outbreak of MeHg poisoning caused 6,530 hospital admissions and 459 hospital deaths [3] plus many cases not taken to hospitals. Women of child-bearing age were exposed to a wide range of doses of MeHg. Thus it was decided to sam- ple residents of rural villages to test for exposure to MeHg and to begin a clinical follow-up study. The emphasis was placed on seeking minimal clinical ef- fects in the infants and relating such effects to mater- nal exposure to MeHg during pregnancy.

Method An extensive program of hair sampling from the head was carried out in many rural areas of Iraq between 1972 and 1974. Thousands of specimens were collected from adults and children. These included hair from women who had been pregnant during the period of MeHg exposure in late 1971 and early 1972 and that was long enough to reflect their prior exposure. Head hair grows at approximately 1 cm per month and accumulates mercury. The mercury con- centration in an active hair follicle is proportional to the blood mercury concentration. Analysis of consecutive segments of a hair specimen provided a peak level of hair mercury concentration for each subject. Thus an estimate of peak maternal blood mercury concentration could be obtained retrospectively for each mother, and this was used as an indicator of the dose of MeHg received by the fetus. Fortunately, the head hair from the adult women was fre- quently 36 to 48 cm in length, which allowed recapitulation over the previous three to four years. Hair sampling was performed on more than 90% of the inhabitants of the vil- lages selected, and identifying data and photographs were recorded.

The hair samples were obtained in a standard manner. A group of 50 to 100 adjacent hairs was grasped firmly be- tween thumb and finger, cut as close as possible to the scalp while being held at 90 degrees to the head, and maintained in its original alignment by clamping the proximal ends of the hairs with a stapler. Each sample was placed in a plastic bag and identified. Some women insisted that the samples be cut in privacy by their husbands; the husbands were given careful instructions but the precise cutting technique was unknown. The hair specimens were analyzed in con- secutive 1 cm segments for total and inorganic mercury by

flameless atomic absorption and in certain cases for methylmercury by gas chromatography [91.

The clinical investigators were provided with names that included controls and mother-infant pairs who had been exposed to MeHg during pregnancy; they had no knowl- edge of the hair analyses. These individuals were located in their rural homes, and standardized questionnaires were used to record the histories and results of physical exam- ination. There was initial difficulty in establishing the pre- cise ages of the children. The time of birth was ascertained by careful questioning, and after reference to important annual events such as times of planting and harvesting or religious festivals, it was concluded that the best dating point was the well-recalled event of the mercury poisoning. Twenty-nine children, 1 5 boys and 14 girls, were found who had been in utero during the epidemic. The mothers were questioned about pregnancy, labor, and delivery, about early milestones of development, their opinion con- cerning the child’s mental or physical retardation, any manifestations of seizure disorder, and any other medical history. The neurological examination was performed in the home, outdoors, or in the village dispensary. It began with a period of observation followed by routine neuro- logical examination, including inspection of pupils and fundi, testing for any major degree of deafness, and obser- vation of gait and posture. When last examined, the chil- dren were between 4$$ and 5 years of age. All the exam- inations were performed by the same two clinicians, who each examined every child. The following criteria for ab- normalities were adopted: motor retardation if the child was not walking at 18 months, speech retardation if not talking by 24 months, mental retardation or seizures (or convulsive-like attacks) according to the history provided by the mother, and neurological signs by agreement of the two examiners. No standards are available for head cir- cumference or height of Iraqi children, so these factors were evaluated in terms of standard deviations below the mean for the group.

Illustrative Case Reports PATIENT 1. A boy was born in March, 1972, in a village near Samawa, Iraq, after an uneventful pregnancy, labor, and delivery without antenatal care. No physician o r mid- wife was present at delivery, and no professional care was given during infancy. The mother considered that her in- fant appeared normal at birth but gradually recognized the presence of severe retardation. At age 4 years 9 months the child was blind and deaf and was unable to stand, walk, or talk. Tonic neck responses were present. All limbs showed an increase in tone and deep tendon reflexes with extensor plantar responses and abnormal posture of the wrist. Mi- crocephaly was present, with a head circumference of 43 cm. The boy’s height was 98 cm. Peak mercury concentra- tion in the mother’s hair was 165 parts per million (ppm). The mother had been asymptomatic during the pregnancy.

PATIENT 2. A girl was born in March, 1972, near Samawa. The mother experienced paresthesias over the extremities during pregnancy. Labor and delivery were normal. The infant was unable to stand or talk until 3 years of age. At 4 years 9 months her speech was dysarthric, pos-

Marsh et al: Fetal Methylmercury Poisoning 349

Table I. Primary Data for 29 Mother-lnfant Pairs

Mother Child

Peak Hair Hg Walked Talked Mental (ppm) Symptoms (mo) (mo) Retard.

384 217 209 204 165 164 154 144 112 99

85 74 5 5 52 25 2 3 20 16 13 12

11 10 9 7 7 5 4 4 2

0 :: p+ :: p : : p 0 : : p : : p : : p : :p+ : :p

. . *

0 : : p : : p t 0 : : p 0 0 0 0

: :p 0 0 : : p 0 0 0 0 0

: :24 :: 24 :: >36 : :24 : : >57 :: 30 12 :: >36 : :24 :: 24

12 12 18 :: 24 18 12 :: 24 1 5 < 12 12

12 18 1 1 : :24 :: 24 < 12 12 : :24 12

:: >36 :: 36 : : 36 :: >24 :: >57 :: 30 12 :: 36 :: 36 : : 36

12 15 : : 36 : : 36 24 12 24 24 18 18

12 24 18 24 12 24 12 : :42 12

No No ::Yes ::Yes ::Yes No No ::Yes No No

No No ::Yes No No No No No No No

No No No No No No No : : Yes No

Seizures

No ::Yes No ::Yes ::Yes No No ::Yes ::Yes No

No No ::Yes ::Yes No No No No No No

No No No No No No No No No

Other Head Plantars CNS Circum. Height

- Extensor Signs (cm)

::Yes ::Yes ::Yes ( 1 ) No ::Yes (1) ::Yes : : Yes : : Yes ::Yes : : Yes No No No No No No No No ::Yes (1) No

No No No ::Yes No No No No ::Yes No No No No No No No No No ::Yes ( 1 ) No

No No ::Yes ( 1 ) No No No No No No No No No ::Yes ( 1 ) No ::Yes ::Yes ::Yes ( 1 ) No

48.5 48.0 48.5 52.0 43.0 48.0 49.0 50.0 50.0 48.5

48.5 51.0 47.0 48.0 51.0 49.5 50.0 48.5 51.5 50.0

49.0 48.5 49.5 48.2 48.2 49.0 49.0 48.5 50.5

(cm)

101.0 94.0 91.0

114.0 87.0 92.0 94.0 96.0 96.0 90.0

92.0 99.5 91.0 94.0

103.0 103.0 97.0 91.0

104.0 104.0

101.0 95.0

102.0 102.0 101.0 100.0 99.0

104.0 . . .

Sex

F F F M M F M M M F

F F F M M F M M M F

F F M F F M M M M

-

0 = no symptoms; P = paresthesias; + = weakness; :: = abnormality; * = weakness, impaired vision, and ataxia; + = weakness and walking difficulty during pregnancy; yes (1) = 1 plantar extensor.

sibly due in part to severe deafness; vision was intact. Other neurological manifestations included hypotonicity of limbs and trunk, ataxia of limbs and gait, and a right extensor plantar response. The head circumference was 48.5 cm and the height was 91 cm. Peak mercury concentration in ma- ternal hair was 209 ppm.

Results The 29 peak mercury concentrations in maternal hair ranged from 2 to 384 ppm with a median concentration of 25 ppm. Two children had signs of severe neuro- logical abnormality, 22 had less marked neurological symptoms and signs of varying degree, and 5 were apparently normal. The mothers of the 5 normal children had hair mercury concentrations of 25 ppm

or less, whereas the mothers of the 2 severely af- fected children had levels of 165 and 209 ppm, re- spectively.

Table 1 summarizes the primary data for the entire group and shows that both maternal symptoms dur- ing pregnancy and abnormalities in the children oc- curred more frequently with higher maternal mer- cury concentrations. These data are depicted in the histograms in the Figure; the 29 pairs are divided into three cohorts of approximately equal size according to the maternal hair mercury levels (0 to 11, 12 to 8 5 , and 99 to 384 pprn). Visually, it appeared that the symptoms and signs were more frequent in the group with the highest mercury level in maternal hair (99 to 384 ppm).

350 Annals of Neurology Vol 7 No 4 April 1980

Table 2. Results of Chi-square Comparisons of Abnormalities According to Maternal Hair Mercury Concentration

Significance Levels from Chi-square Tests

3 x 2 2 x 2 2 x 2 on the 3 Groups on 2 Groups on 2 Groups

Abnormality (0-11, 12-85, 99-384) (0-1 1, 12-85) (0-85,99-384)

Motor retardation p < 0.01 NS p < 0.01 Speech retardation p < 0.001 NS p < 0.001 Mental retardation NS NS p = 0.10 Seizures p < 0.05 NS p < 0.03 Plantars extensor NS NS NS Other CNS signs NS NS p = 0.10 Maternal paresthesias p < 0.05 NS p < 0.01

Maternal Hair Mercury

0 O-llppm(N=9) W 12-85ppm (N=10)

99-384 (N=10)

Historical Data 10 1

Motor Speech Mental Seizures

c-- Retardation - Clinical Data

lo 1

None "

Extensor Other CNS Small Short Plantars Signs Head Height

I- 2 S D below mean 4

Infants' symptoms and signs related t o maternal hair mercury concentration.

Using the data in Table 1, chi-square tests were performed on the histograms in the Figure. Three tests were carried out: a chi-square 3 x 2 test on the three hair mercury groups (0 to 11, 12 to 85, and 99 to 384); a chi-square 2 x 2 test on the two lower groups (0 to 1 1 and 12 to 85); and a chi-square 2 x 2 test on the combined lower groups (0 to 85) versus the highest group (99 to 384). The results are sum- marized in Table 2. In some cases, exact tests were used instead of the chi-square because of small num- bers in certain categories. The outcome indicates that

Table 3. Frequency of Infants' Symptoms and Signs Related t o Maternal Hair Merculy Concentration

No. of Infants

Abnormalities per Infant Maternal Total Hg (ppm) 0 1 2 3 4 5 6 Infants

0-11 3 5 0 0 0 1 0 9 12-85 4 4 0 2 0 0 0 1 0 99-384 1 0 1 2 3 1 2 10

there is no significant difference between the fre- quency of abnormalities in children of mothers with hair mercury levels of 0 to 11 ppm versus those with 12 to 85 ppm. The difference between 0 to 85 and 99 to 384 ppm is significant (for p values see Table 2).

The children's head circumferences ranged from 43 to 52 cm with a mean of 4 9 cm and standard de- viation of 1.65 cm. Only 1 child (Patient 1) had a circumference 2 standard deviations below the mean. The mean head circumference for boys was 49.3 cm and for girls 48.7 cm; these are at approximately the fiftieth percentile for children in the United States aged 21/2 years. However, there was no significant difference in head circumference between the three hair mercury concentration groups.

The mean height for boys was 98.6 cm and for girls 96.9 cm, which are average heights for children in the United States between 3 and 395 years. No child was 2 standard deviations below the mean height.

The high mercury group also differed from the other two groups in the number of infants having multiple symptoms and signs (Table 3). Six of the 10 infants in the highest exposure group had four or more adverse effects, whereas the two groups with lower levels of exposure contained only 1 such infant (p < 0.005, Fisher exact test).

The most common maternal symptom during pregnancy was transient paresthesias, which occurred in 80% of the group having a high mercury level

Marsh et al: Fetal Methylmercury Poisoning 35 1

compared with 30% and 22% in the two lower expo- sure groups.

Discussion Evidence has been presented indicating adverse ef- fects of MeHg o n the developing fetal nervous sys- tem. In general, these effects are less severe than those reported previously from Iraq [l] and Japan 11 11, probably because the mother’s exposure during pregnancy was less severe.

Clinical examination alone could not have iden- tified a single case with confidence since the signs are not sufficiently specific. Clinical examination to- gether with hair mercury analysis could not have proved a causal relationship in isolated cases, but the combined clinical and analytical study of children subjected to a range of doses of MeHg does suggest an association between dose and response.

Apart from MeHg, the possible causes of the ob- served abnormalities may include a variety of factors. The effects of MeHg are seen against this back- ground. Nevertheless, because of the nonspecificity of the clinical picture, we cannot exclude the possible presence of a covariable that might have contributed to the increased frequency of adverse effects in the group with a high mercury concentration (99 to 384; see Table 3 ) . Some possible covariables are as fol- lows. First, the possibility was considered that low- income families ate more of the contaminated bread, but marked differences in socioeconomic status and nutrition were not apparent. Height and head cir- cumference were not significantly different in the high and low mercury groups. Furthermore, the rea- sons for consumption of the contaminated bread were many [ 3 ] , and none would indicate a covariable responsible for the clinical effects. Second, although the examination team conducted the study “blind,” being unaware of mercury levels, the mothers were probably aware that they had eaten the contaminated bread. This knowledge may have influenced their re- plies to questions on developmental milestones and seizures. Third, the observed developmental and neurological effects on the infant, assuming that MeHg was the cause, may not have been entirely due to prenatal exposure of the fetus. Continued intake of MeHg in breast milk may have contributed. The infants were breast-fed for at least one year and usu- ally for two years. Amin-Zaki et al [2] reported that breast-feeding tended to slow the rate of decline of blood mercury concentrations in infants with high blood MeHg concentrations at birth compared to their mothers. However, blood concentrations were not elevated by breast-feeding to values higher than those at birth. Fourth, exposure of the mother to MeHg may have affected her dietary habits during pregnancy and her care of the infant after birth.

None of these “indirect” effects of MeHg can be completely discounted, but, as will be discussed shortly, the effects on the mother were slight and transient. Finally, none of the mothers drank al- coholic beverages, so the fetal alcohol syndrome was not a consideration.

Reports from Japan 111, 211, Sweden [8], and the United States [16, 181 indicate that MeHg can cross the placenta and severely damage the fetal brain while causing no or slight transient symptoms in the mother. Observations on animals are in agreement [19]. The findings in this report also support the con- clusion that effects on the fetus are more severe. Transient paresthesias during pregnancy were the most frequent health complaint in the mothers. More serious effects were observed in the infants, includ- ing severe disability in 2. The increased frequency of symptoms in both mothers and infants occurred in the same concentration range (99 to 384 ppm in ma- ternal hair). The important difference lies in the se- verity of the effects.

Minor congenital anomalies were reported in 2 cases of fetal MeHg poisoning by Harada 1111, and congenital cleft palate and harelip have been noted in animals exposed to MeHg prenatally [19]. N o such congenital abnormalities were observed in this group of patients.

The subjects of this study were exposed during the mothers’ pregnancy to MeHg that resulted from in- gestion over a period of a few months. It is not clear that level of peak exposure is the only important factor, and duration of exposure may be relevant. The dose required to produce a fetal effect may be smaller if maintained at a constant level throughout pregnancy than if applied during part of the preg- nancy. Difficulties in determining exact birth dates precluded any correlation of fetal exposure with gestational stage in this study, but it may be specu- lated that the fetus is more susceptible at some stages than at others.

Neuropathological reports on the effects of MeHg poisoning on human fetal brain have been restricted to 2 cases from Japan [14] and the 2 Iraqi infants re- ported by Choi et al [ 5 ] . The latter were exposed during the third and fourth months of gestation, which is a critical period for neuronal migration; the chief findings were abnormal migration of neurons to the cerebral and cerebellar cortices together with de- ranged cortical organization. The location of this im- portant effect on the development and maturation of fetal brain contrasts with the areas of focal atrophy present in postnatal MeHg poisoning.

The small number of infant-mother pairs in this study does not allow us to identify a specific “threshold” maternal hair concentration below which there will be no adverse effects in either mother or

352 Annals of Neurology Vol 7 No 4 April 1980

infant. A high risk of adverse effects appears to exist at maternal hair concentrations in the range of 99 to 384 ppm. However, in the next lower concentration range (12 to 85 pprn), the frequencies fall dramati- cally and do not differ significantly from those seen in the lowest range (0 to 11 ppm). Thus, the adverse effects seen at levels up to 85 ppm may have been due to causes other than the MeHg exposure. Un- fortunately, only 4 infant-mother pairs were available for study in the range of 50 to 99 ppm maximum maternal hair concentration, and none was available at levels between 2 5 and 50 ppm. It is clear that more data are needed in this concentration range.

Psychomotor retardation, ranging from mild ef- fects to severe cerebral palsy, can be caused by fetal exposure to MeHg; the effects are dose related.

Supported by Grant ES01248 from the National Institutes of Health.

We wish to thank the Deans of the Medical College of Baghdad and of the University of Rochester School of Medicine and Den- tistry as well as the Minister of Health of Iraq for facilitating this study; the following physicians for their assistance in the field: Drs Saeed Ali, S. N. Ali-Kawak, Jakub Gawo, Samir Habbana, Amer Al-Khalisi, Towfik Muhawish, Amin Nassir, Zeki Shabana, and Ali Sudani; and Dr Christopher Cox for statistical consultation.

References 1 . Amin-Zaki L, Elhassani S, Majeed MA, et al: Intra-uterine

methylmercury poisoning in Iraq. Pediatrics 54:587-595, 1974

2. Amin-Zaki L, Elhassani S, Majeed MA, et al: Studies of in- fants post-natally exposed to methylmercury. J Pediatr 85:

3. Bakir F, Damluji SF, Amin-Zaki L: Methylmercury poisoning

4. Berglund F, Berlin M, Birke G, et al: Methylmercury in fish.

81-84, 1974

in Iraq. Science 181:230-241, 1973

Nord Hyg Tid Suppl4, 1971

5. Choi BH, Lapham LW, Amin-Zaki L, et al: Abnormal neu- ronal migration, deranged cerebral cortical organization and diffuse white macter astrocytosis of human fetal brain: a major effect of methylmercury in utero. J Neuropathol Exp Neurol

6. Damluji S: Mercurial poisoning from fungicide Granosan. Med J Fac Med Baghdad 4:83-103, 1962

7. Derlan LKA: Outbreak of food poisoning due to alkylmercury fungicide on Southern Ghana State Farm. Arch Environ Health 28:49-53, 1974

8. Engleson G, Herner T: Alkylrnercury poisoning. Acta Paediatr 41:289-294, 1952

9. Giovanoli-Jacubczak T, Greenwood MR, Smith JC, et al: Determination of total and inorganic mercury in hair by flameless atomic absorption and of methylmercury by gas chromatography. Clin Chem 30:222-225, 1974

10. Haq FU: Agrosan poisoning in man. Br Med J 1:1579-1582, 1963

11. Harada Y: Congenital (or fetal) Minamata disease, in Study Group of Kumamoto University, Minamata Disease. Shuhan, 1968

12. Hunter J, Bomford RR, Russell DS: Poisoning by methylmer- cury compounds. Q J Med 33:193-213, 1940

3. Jalili MA, Abbasi AH: Poisoning by ethyl mercury toluene sulphonanilide. Br J Indust Med 18:303-308, 1961

4. Matsumoto H, Koya G, Takeuchi T: Fetal Minamata disease. J Neuropathol Exp Neurol 24:563-574, 1965

5. Ordonez JV, Carrillo JA, Miranada M: Estudio epidemio- logic0 de una enfermedad considerada como encefalitis en la region de Los Altos de Guatemala. Bol Of Sanit Panam

16. Pierce PE, Thompson JF, Likosky WH, et al: Alkylmercury poisoning in humans; report of an outbreak. JAMA

17. Rustam H, Hamdi T: Methylmercury poisoning. Brain

18. Snyder RD: Congenital mercury poisoning. N Engl J Med

19. Spyker MJ, Smithberg M: Effects of methylmercury on pre- natal development in mice. Teratology 5:181-186, 1972

20. Takeuchi T: Pathology of Minamata disease, in Study Group of Kumamoto University, Minamata Disease. Shuhan, 1968

2 1. Tsubaki T, Irukayama K: Minamata Disease. Amsterdam, Elsevier, 1977

37:719-733, 1978

60~510-519, 1966

220:1439-1442, 1972

97:499-510, 1974

28411014-1016, 1971

Marsh et al: Fetal Methylmercury Poisoning 353