Title: Fish Consumption, Dental Amalgams, and Other Aspects of Mercury Toxicity

    While fish and seafood contain methylmercury, they also provide important nutrients, which may benefit nervous system development. Impact on the same health outcomes by two exposures originating from the same food source provides a classical example of confounding. To explore the extent of bias, structural equation modeling was applied to the data, with adjustment for the benefits conferred by maternal fish intake during pregnancy. This adjustment resulted in an increased effect of the prenatal methylmercury exposure, as compared to the unadjusted results, and the change was statistically significant for motor functions. Because a dietary questionnaire response is likely to be an imprecise proxy for the transfer of seafood nutrients to the fetus, the impact of imprecision was also modeled. At a realistic imprecision level, the mercury-associated deficit on the motor function increase by two-thirds, as compared to the unadjusted effect. These results suggest that uncontrolled confounding, the imprecision of the confounder, caused a substantial bias toward the null. The adverse effects of methylmercury exposure from fish and seafood are therefore likely to be underestimated by observational studies. Dr. Emily Oken continues her previous work on the competing effects of fish consumption and essential fatty acids and methylmercury on neurodevelopment. This work in many ways complements that of Dr. Grandjean and has led to numerous public health insights on Hg toxicity.
Research on Hg toxicity has been conducted by Drs. Philippe Grandjean, Emily Oken and David Bellinger. A birth cohort established at the Faroe Islands has now been followed through age 16 years, and most recently updated with data from the 9th grade school exams. This project is supported by NIEHS grant ES 09797 (PI: Philippe Grandjean). His ongoing cohort was recently refunded for further long term follow up. Statistical analyses first focused on the neurophysiological results. In identifying potential confounders, covariates were examined for their association with the exposure variables. The following parameters were associated with an increased prenatal mercury exposure: Residence in fishing village; Tested at an older age; Living with both parents; Tested in Faroese; Examined in the Faroes; Higher maternal parity; Greater number of amalgam fillings; and Attending a small secondary school.
In addition to organic Hg, center investigators also tackled inorganic Hg. Dr. David Bellinger recently published the definitive study of mercury containing dental amalgam and neurodevelopment. In a landmark randomized controlled trial he demonstrated that Hg containing dental fillings are not associated with neurocognitive deficits. A total of 534 6-to-10 year old children with substantial unmet dental needs were recruited, from urban (Boston/Cambridge, MA) and rural (Farmington, ME) areas of New England. Children were randomly assigned to have all existing and incident caries restored using either dental amalgam, which contains approximately 50% elemental mercury by weight, or composite resin, which is mercury-free. The children received free dental care over a 5-year follow-up period and underwent yearly neuropsychological evaluations. The primary neuropsychological endpoint was Full-Scale IQ on the WISC-III. Approximately 60 other neuropsychological endpoints were assessed. The primary renal endpoint was urinary albumin concentration (creatinine-corrected). Although children in the amalgam group had significantly higher levels of mercury in urine than the children in the composite group, intention-to-treat analyses did not reveal any statistically significant differences between treatment groups on the primary endpoints (Bellinger et al., 2006). Analyses of the secondary endpoints also did not reveal a consistent pattern of group differences. Use of two continuously-distributed indices of exposure to mercury (surface-years of amalgam, urinary mercury excretion) provided similar results (Bellinger et al., in press; Bellinger et al., submitted). This trial provides no evidence that, under the conditions in which dental amalgam was used in this trial, exposure to mercury released from amalgam restorations produced any detectable adverse effects on school-age children over the 5-year follow-up interval.