New Seychelle study: Mercury and autism not linked

Mercury, entering our environment both from natural and anthropogenic sources, can be transformed by microbes to methylmercury, which can be accumulated in aquatic organisms, leading to relatively high concentrations in bigger predatory fish (see picture to the left). Although vaccines have been suggested as a source of mercury exposure, the main source of exposure to organic mercury in humans is through methyl mercury found in fish.

Since Methylmercury is a strong neurotoxin, exposure through the diet is of great concern, and health organizations recommend to restrict the consumption of fish depending on the type of fish and the vulnerability of the consumer. Despite the great concerns, scientists have great difficulties to prove any problems related to this intake. Therefore the "sentinel" population of the Rebublic of Seychelles, an island chain lying northeast of Madagascar has long attracted scientific interest. As islanders, the people of Seychelles consume a lot of fish - 10 times more than the average North American or European - and much of what they eat contains methylmercury. Although concentrations are low, the total amounts of methylmercury taken up are greater than much of the rest of the world eats, resulting in an intake about 20 times higher than that of an American.

This special situation of 87,000 people naturally consuming a fish-rich diet attracted the attention of epidemiologists already in the 1980s, when the Seychelles Child Development Study was launched to examine the developmental effects of this low-level dietary mercury exposure.

To examine whether consuming too much fish during pregnancy is associated with autism, Davidson and colleagues evaluated 1,784 young people and their mothers, enrolled in the study from 1986 to 2001. For the evaluation, the researchers used the Social Communication Questionnaire, a 40-item questionnaire widely used to screen people for ASD symptoms. Although not a diagnostic test, a score of 15 or higher warrants further evaluation.

Maternal hair samples collected around the time of birth were used to assess prenatal exposure to methyl mercury. A subset of 532 children (mean age 10 years) were also administered the Social Responsiveness Scale, a 65-item scale designed to recognize ASD symptoms in social settings. A score of 70 or greater in males and 65 or greater in females indicates a possible ASD. The primary endpoint was Social Communication Questionnaire scores in relation to mothers' prenatal mercury levels.  

Study participants had an average prenatal methyl mercury level of 8.4 parts per million, and average score on the Social Communication Questionnaire was 8.0. Approximately 8% of participants scored 15 or higher on the Social Communication Questionnaire.

Average prenatal methyl mercury exposure for the 532 participants who were administered the Social Responsiveness Scale was 6.7 parts per million. The mean score on that test was 57.6, with 34% of participants scoring above the cut-off  for possible autism.

Using linear and nonlinear regression analyses, the researchers found no consistent correlation between prenatal exposure to methyl mercury and scores on ASD screening instruments.

"This study shows no consistent association in children with mothers with mercury levels that were six to 10 times higher than those found in the U.S. and Europe," study investigator Philip Davidson, PhD, an emeritus professor at University of Rochester Medical Center said in a press release. "This is a sentinel population and if it does not exist here than it probably does not exist."

As with similar studies, also this one comes with several caveats. Primarily because of the decades-long period of the study and variations in data collection over time, the researchers couldn’t adjust for all of the possible variables that might have skewed the analysis. One thing they note is that the study population comes from an area with almost 100% vaccination uptake and where vaccines contain thimerosal, so exposure to thimerosal–which contains a different form of mercury was considered to be consistent across the group.

A review of earlier studies on the developmental effects of prenatal mercury exposure found inconsistencies (see J.L. Jacobson 2001) , which were ascribed to differences in study design. Because prospective epidemiological studies are often hampered by limited control over confounding and other factors, including unmeasured between cohort differences in genetic vulnerability and nutritional adequacy, interferences about toxicity often depend heavily on a qualitative assessment of the weight of the evidence from multiple studies.