Faroe Islands Studies
Drs. Richard Clapp and Philippe Grandjean
Boston University School of Public Health

Health Effects of Seafood Contamination with Methylmercury in the Faroes
Dr. Philippe Grandjean
University of Southern Denmark and Boston University

Neurotoxicity due to methylmercury is well documented from unfortunate poisoning incidents. However, the dose-effect relationships have been poorly documented, and the impact of imprecision and potential bias is unclear from past epidemiological studies. We first generated a cohort of 1022 consecutive singleton births during 1986-1987 in the Faroe Islands, where increased methylmercury exposure is mainly due to consumption of pilot whale meat. We measured total mercury concentrations in cord blood and maternal hair collected at parturition. Because prenatal neurotoxic effects would be permanent, assessment of neurobehavioral functions was postponed to the age of 7 years, where the children could undergo detailed testing. A total of 917 of the cohort children underwent thorough examinations. Significant exposure-related dysfunctions were seen in most neuropsychological tests and were most pronounced in the domains of language, attention, and memory. Mercury-associated effects were also seen in delayed latencies for evoked potentials and in blood pressure regulation. The associations remained significant after adjustment for covariates and also after exclusion of children with high maternal hair-mercury concentrations (corresponding to the benchmark dose) or high PCB exposures. Results from examinations conducted of the same children at age 14 years are currently underway. As expected, the cord-blood mercury concentration was the best risk indicator. However, statistical analyses suggested that even this parameter was associated with an error variance that substantially exceeded the analytical imprecision, although it was much less than that associated with the maternal hair-mercury concentration. Such imprecision leads to an underestimation of the true mercury effects.

A second Faroese cohort of 182 singleton term births was generated in 1994 with more detailed exposure documentation. In this cohort, we have documented mercury-associated decreases in the neonatal Neurological Optimality Score and in postnatal growth. Detailed statistical analyses of both prospective cohort studies have failed to identify any covariates that could account for the mercury-associated effects, which remained robust when using different analytical strategies. Although exposure misclassification may be more likely in cross-sectional studies, we have also seen adverse effects in studies of mercury-exposed children from Brazil and Madeira, where developmental exposure levels were determined from current hair-mercury concentrations. In conclusion, we have obtained evidence of subtle adverse effects on neurobehavioral functions, blood pressure, and growth. At age 7 years, a doubling of the mercury exposure corresponds to a developmental delay of up to 2 months. Although IQ tests were not done, such delays would be comparable to a loss of about 1.5 IQ points. These dysfunctions are detectable at exposure levels prevalent in many parts of the world where contaminated seafood or freshwater fish constitutes an important part of the diet.

Complete Abstract List (suitable for printing)