Myers, GJ, PW Davidson, C Cox, CF Shamlaye, D Palumbo, E Cernichiari, J Sloane-Reeves, GE Wilding, J Kost, L-S Huang, and TW Clarkson. 2003. Prenatal methylmercury exposure from ocean fish consumption in the Seychelles child development study. Lancet 361: 1686-92

This study challenges previous findings that fetal exposure to relatively low levels of methyl mercury resulting from maternal consumption of contaminated fish can harm neurological development.

The research team, led by Dr Gary Myers from the University of Rochester Medical Center, reports that methyl mercury in the womb does not harm development of a baby's brain, at least at levels experienced because of ocean fish consumption. The weight of the evidence on mercury's impacts, however, still warrants strong protections, and justifies strengthening current US FDA standards to be consistent with the US EPA and National Academy of Sciences.

What did they do? Myers et al. carried out a series of mental and motor tests on 9-yr old children born on the Seychelles Islands. Women living on the Seychelles eat ocean-going fish carrying methyl mercury levels comparable to that found in comparable fish in the US (roughly 0.3 µg/g), but the fish consumption rate in the Seychelles is significantly higher. Hence babies in the womb in the Seychelles are exposed to more methyl mercury than is typical for US babies.

The tested children had been enrolled in this ongoing research project at birth, when samples of their mother's hair were taken to determine mercury levels experienced in pregnancy. During infancy and childhood, the research team conducted a series of psychomotor tests on the subjects, at ages 6, 19, 29 and 66 months of age, none of which had revealed developmental effects of mercury.

The current measurements, conducted when the enrolled children were 9 years old, tested a series of psychomotor endpoints, including neurocognitive, memory, motor, perceptual-motor and behavioral functions. Other variables known to affect cognitive development, for example the home environment, maternal IQ and socioeconomic status were also assessed and treated as potentially confounding variables in the statistical analyses the team conducted.

What did they find? Myers's group obtained measurements from 779 infants whose mothers ate an average of 12 fish meals per week.

No pattern linking higher methyl mercury exposures to adverse neurocognitive scores emerged in the analysis. Two measured variables at first appeared to be related to mercury levels, but statistical considerations indicated that the relationships were most likely due to chance.

The research team did confirm, however, consistent associations between some of the confounding variables (e.g., maternal IQ) and neurocognitive development, thus indicating their testing methodologies could have detected mercury effects, were they to have occurred.

What does it mean? Viewed in isolation, these results might be interpreted to ease concerns about low level mercury exposures associated with ocean-going fish consumption and the possible impact on fetal development, at least beneath current US FDA exposure standards. According to the lead scientist, Dr. Gary Myers, current FDA standards "are reasonable." Existing standards, for example, recommend that pregnant women avoid certain highly contaminated fish like swordfish altogether, and limit consumption of certain other species to no more than 12 ounces per week.

This is good news if it means that ongoing mercury contamination has not yet been sufficient to force all ocean-going fish, an important source of nutrition, out of diets.

Indeed, one possible interpretation of the results is that the positive effects of eating fish oils may be protective against adverse effects of mercury. The research group led by Myers is now attempting to test this hypothesis.

But two other extensive studies of mercury impacts on fetal development have shown mercury to have deleterious effects. One of these studies is of children born on the Faroe Islands whose mothers ate whale meat and blubber. The other is from New Zealand, where the exposure is via consumption of contaminated shark meat used in fish and chips.

At least four possible explanations for the conflicting results are now being explored.

• One focuses on the fact that the shark muscle and whale meat/blubber on the Faroe Islands and in New Zealand carry much higher methyl mercury concentrations than does the fish eaten on the Seychelles. Thus even if the Seychelle Islanders ate cumulatively as much mercury as the other study populations, because fish formed a larger and more consistent portion of the diet, mercury would have reached fetal tissue in a relatively low-level continuous stream on the Seychelles, as compared to a more burst like pattern with brief but higher exposure levels in the Faroes and New Zealand. These higher burst exposures may have different effects than more continuous but lower exposures. Recent analysis has suggested, however, that the Faroe Island exposures are relatively continuous instead of burst-like.
• A second builds from the observation that the whale meat consumed on the Faroe Islands not only contains mercury but also a collection of persistent chlorinated compounds, including PCBs, and that these may either cause the effect or interact with mercury to influence neurological development. This interpretation, however, would not explain the New Zealand results, because the shark meat consumed there does not contain significant PCBS. Moreover, the Faroe Island study took great care in separating the effects of mercury from the effects of PCBs. Their mercury effect is most likely real, and independent of simultaneous PCB exposures. This interpretation is supported by independent analysis by the U.S. National Academy of Sciences.
• The Faroe Island study measured mercury in umbilical cord blood and hair whereas the Seychelles study used hair samples. Other research has suggested that hair samples may be less accurate. Hence it may be that had the Seychelles work measured mercury in cord blood, they would have been able to find effects at lower levels. This interpretation, however, is not universally accepted.
• For a combination of reasons related to study design and execution, the Seychelles study does not have sufficient statistical power to detect small differences.

Thus at the moment these studies collectively do not provide clear signals for how parents or public health officials should make decisions about consuming fish contaminated with low levels of mercury. The current study suggests existing standards are adequate; the others suggest they should be tightened.

In issues of public health, particularly where irreversible damage may be at stake, when two well-designed and well-executed studies conflict in their conclusions, the prudent course is to structure public health regulations based on the study that shows adverse effects rather than those showing no effect. The Faroe Island and New Zealand studies both indicate that current FDA standards remain too lax.

Parents wishing to exercise caution might reasonably:

• continue to keep ocean fish such as Pacific salmon, flounder and haddock in family diets because of the nutritional benefits;
• limit consumption to no more than one meal a month of moderately contaminated fish like canned tuna, mahi mahi, blue mussel, cod, pollock, salmon from the Great Lakes; and
• forego species known to be relatively high in mercury, including swordfish, tilefish, king mackeral and tuna steaks.

Note that data from studies by the Mobile Register in Alabama suggest that special circumstances in the Gulf of Mexico related to disposal of oil drilling muds have lead to unexpectedly high methyl mercury contamination in predatory sport fish in these waters. For access to these studies see (1) the introductory exposé published in the Mobile Register on 22 July 2001; and (2) a list of all articles in the series.