Methylmercury Exposure and Neurotoxicity

Exposure to high levels of methylmercury has produced fatalities and devastating neurological damage among adult survivors.¹ Severe cases of in utero methylmercury poisoning strongly resemble cerebral palsy and are often accompanied by deafness and mental retardation.^{1 - 2} In addition, clinically evident cases of neurological damage have occurred among infants born to mothers whose own symptoms were often only transient paresthesias. Consequently, the fetus is considered much more sensitive to methylmercury exposure than is the adult.

Although recognizing clinically evident symptoms is relatively straightforward, determining levels of in utero methylmercury exposures associated with the onset of toxic responses in infants and children is a much more complex problem. In this issue of THE JOURNAL, Davidson et al³ present findings from a cohort in the Republic of Seychelles that indicate consumption of methylmercury from fish did not produce developmental deficits among children exposed both in utero and postnatally to methylmercury. Developmental status among the Seychellois was assessed by traditional indices of child development (among others, the Woodcock-Johnson Tests of Achievement and the General Cognitive Index of the McCarthy Scales of Children's Abilities). In the Seychellois cohort, mean maternal hair mercury concentrations during pregnancy were 6.8 ppm (SD, 4.5; range, <3-26.7 ppm). Davidson et al have responded to the need to evaluate effects of long-term, low-level methylmercury exposures and suggest that the Seychellois data can be used to support efforts to define an upper tolerable level for methylmercury in hair.

The most direct comparison with the Seychellois data is the longitudinal assessment of children from a seafood (mostly pilot whales)–consuming population in the Faroe Islands.⁴ The Faroese cohort had maternal hair mercury concentrations comparable with the Seychellois children (median, 4.5 ppm and 13% >10 ppm⁴ ). However, in the Faroese cohort maternal hair mercury concentrations in the range of 3 to 10 ppm were associated with neuropsychological dysfunctions among the children that were most pronounced in the domains of language, attention, and memory and to a lesser extent in visual-spatial and motor functions.⁴ In addition, this research group used tests that are better able to discern subtle cognitive and neuromotor performance disturbances (eg, the Neurobehavioral Evaluation System tests for motor speed, hand-eye coordination, and attention⁵ ).

Defining the level of exposure for neurodevelopmental toxic effects is influenced by the types of tests used as the end points for neurological assessment. Traditional measurements of child development were the basis of the conclusion by Davidson et al³ that no effects were observed following in utero methylmercury exposures in their cohort. However, use of additional tests that can identify subclinical neurocognitive function may provide an expanded understanding of methylmercury effects. Recognizing the importance of evaluating subtle indicators of neurodevelopmental status, Davidson et al plan to assess the children in the Seychellois cohort further by using tests that can detect more subtle deficits in neurobehavioral function. Until these data are available, current findings from the Seychellois cohort must be regarded as interim.

Davidson et al³ suggest that developmental effects shown by other groups⁴ may reflect their exposures to other neurotoxicants, including polychlorinated biphenyls (PCBs) from seafood. The statistical analyses of data from the Faroese study directly assessed this question and found that in utero methylmercury exposure was associated with dysfunction in broad cognitive domains and the mercury-associated effects persisted after statistical control for PCB exposure. Only 1 test of language showed combined PCB and methylmercury effects in the Faroese cohort.⁴ The role of other contaminants remains to be elucidated.

The advanced developmental status of the Seychellois population, particularly when tested at younger ages, raises concern about the difficulty in generalizing results from this cohort to other populations.⁶ Earlier reports illustrate the robust developmental status of this population.⁷ In the Seychellois cohort, only 3 (0.4%) of 737 individual examinations on the Denver Developmental Screening Test–Revised were rated as abnormal, indicating a highly skewed distribution when compared with samples of children in the United States.⁸ In the article in this issue of THE JOURNAL on postnatal exposures to methylmercury at age 66 months, children within the highest quartile of exposures had statistically significant increases in test scores on several developmental outcomes compared with the other 3 quartiles. Four of 6 measures of development showed better scores in the highest methylmercury exposure groups compared with lower-exposure groups for both prenatal and postnatal methylmercury exposures. This seemingly paradoxical response to additional methylmercury exposure raises further questions about interpretation of the study's results.

One possibility is that the Seychellois population is in some way buffered from the adverse consequences of in utero exposure to low levels of methylmercury and benefits from a high level of fish consumption. Davidson et al³ suggest that high levels of fish consumption may be beneficial to development. Generalizing this possible benefit of a diet high in fish to populations that consume much smaller amounts of fish that contain higher levels of mercury contamination is problematic. For people whose mercury exposures come from eating smaller amounts of fish that are more highly contaminated with methylmercury (ie, fewer benefits of eating fish but higher exposures to methylmercury) than the Seychellois cohort, the buffering effects of fish consumption experienced by the Seychellois may not exist. Such people would receive fewer benefits from fish consumption but would be at risk for all the potential adverse effects from methylmercury exposure.

Potential differences in timing, magnitude, and duration of methylmercury exposure may be influenced by methylmercury concentrations in fish and the frequency and amount of fish consumption. Methylmercury from fish is readily absorbed by humans, with approximately 95% of an ingested dose transferred from the gastrointestinal tract.^{9 - 10} Outbreaks of developmental neurotoxicity caused by methylmercury in Japan amply document that methylmercury poisoning is produced by heavily contaminated fish.^{1 - 2} In addition, experimental studies with laboratory animals have provided evidence that fish intake is not protective against methylmercury poisoning.¹¹

Davidson et al³ observe that the Seychellois population is exposed to methylmercury through consuming substantial amounts of fish with mercury concentrations of about 0.05 to 0.2 ppm. On average, the mixes of fish species consumed in the United States have mercury concentrations of about 0.13 ppm.¹² However, some local freshwater sources of fish in the United States have much higher mercury concentrations than those found in typical commercial sources. Fishing advisories based on mercury contamination are issued by 41 states and 11 states advise limitations on fish consumption from all water bodies. Data gathered in the mid-1990s from the northeast states show average mercury concentrations more than 0.5 ppm in 20% to 100% of samples and more than 1 ppm in 2% to 25% of samples.¹³ In Wisconsin, the mercury concentration of walleye, the most commonly sought-after game species, averages approximately 0.5 ppm with some individual values greater than 3 ppm (James Amrhein, written communication, 1998).

With 35 states having PCB-based fishing advisories, contamination with PCBs is also a problem in many states that have advisories for mercury. For example, Maine has a mercury-based advisory on all water bodies and advises pregnant and nursing women, as well as children younger than 8 years, to limit or avoid consumption of fish from many rivers because of PCBs and dioxin.¹⁴ Likewise, Minnesota recommends limited consumption of fish because of both PCBs and mercury (Patricia McCann, MS, written communication, 1998).

Patterns of exposure to contaminants from fish differ between the United States and the Seychelles. Much smaller amounts of fish are usually consumed in the United States,¹² although some US subpopulations consume far larger amounts of fish (eg, people of Asian or Pacific Island heritage and some Native American tribes).¹² The amounts of fish consumed in the United States are far more varied than in the Seychelles. Approximately 1% to 2% of the US population reports eating fish daily, whereas about 10% rarely consumes fish.¹² Many of the US water bodies that are high in mercury are also high in PCBs, a recognized developmental neurotoxin.¹⁵ By contrast, fish ingested by the Seychellois are thought to have low levels of PCB exposure.³

The findings from the study by Davidson et al³ on methylmercury exposure and neurotoxicity in the Seychellois cohort represent interim results based on testing with traditional measurements of child development. The results from additional evaluations using more sensitive tests of development are needed before firm conclusions can be drawn. Until then, women of childbearing age who eat fish nearly daily need to be aware of the level of mercury contamination in the fish they consume. Forty-one US states provide advice to different age groups about types and amounts of fish that should be consumed. Advisories that recommend restrictions of fish consumption for women of childbearing age because of mercury contamination should be followed carefully.