Schober, SE, TH Sinks, RL Jones, PM Bolger, M McDowell, J Osterloh, ES Garrett, RA Canady, CF Dillon, Y Sun, CB Joseph and KR Mahaffey. 2003. Blood mercury levels in US children and women of childbearing age, 1999-2000. Journal of the American Medical Association 289:1667-1674.

Written by senior scientists from the US Food and Drug Administration, Environmental Protection Agency and the Centers for Disease Control, this study reports that almost 8% of women of childbearing age in the US have blood mercury levels at levels of concern for fetal development.

Methyl mercury is a known neurotoxicant. It easily crosses the placenta and enters into the fetus, where its effects on the developing brain can lead to psychomotor retardation, seizures, developmental delays, and mental retardation.

The main exposures of people to mercury occur when they eat predatory fish like swordfish and tuna that are high on the food chain. Mercury contaminating these fish originates in a combination of human and natural releases of mercury into the environment. The principal human-created emissions occur when coal is burned to generate electricity and from improper disposal of mercury-containing equipment, especially when it winds up in incinerators.

The special significance of this report goes beyond the details of the results. Authored by a team of government scientists, it reflects the convergence of different government agencies toward agreement about what level of mercury is high enough to trigger concerns and thus to require food safety warnings.

The US EPA had already begun to follow a standard recommended by the US National Research Counci in 2000: that food with levels of mercury at or above 5.8 µg/L should trigger alarms about risks to a developing fetus. This level is called the "reference dose" by the EPA.

The Food and Drug Administration, which is the federal agency responsible for establishing these standards in food, however, had resisted lowering their standard to this level. This paper signals that FDA is lowering its standard to be in agreement with the NRC and the EPA.

If the FDA follows through with this new standard, it will be forced to implement new fish advisories. More fish will be added to the list, and the allowable consumption levels will be lowered. For example based on this new standard, as little as half a can of tuna per week will push a child over the acceptable limit. The limit for a 130-pound woman will be one can per week. Scientists familiar with past FDA policies describe the new approach as "a sea change."

What did they do? Schober et al. analyzed data obtained during 1999 and 2000 in the 1999-2000 National Health and Nutrition Examination Survey (NHANES), a carefully structured health survey conducted by the Centers for Disease Control to get a statistically-valid snapshot of the health of Americans. During the survey, blood was obtained and analyzed for a range of contaminants, including mercury. Schober et al. carried out a detailed statistical analysis of the data on mercury. An overview of the results had been released in January.

In addition to taking blood samples, participants were questioned about their diet, yielding information on links between consumption patterns and mercury exposure.

The survey was structured to obtain data from a statistically representative sample of Americans. The sample size was too small to allow geographic comparisons across the United States. It also would have been too small to allow ethnic and age comparisions, but extra samples (a process called "over-sampling) were obtained from several groups (ages 12-19 and over 60; Mexican Americans, Black Americans, women who said they were pregnant) that would otherwise have had too few representatives in the sample.

What did they find? Eighty-one percent of children and 94% of women sampled contained detectable levels of mercury.

• Levels were higher for women than for children, with women averaging 1.02 µg/L and children 0.34 µg/L.
• Levels were higher for people who ate fish more frequently.
• Among women between the ages of 20 and 49, mercury levels were higher in women with higher education.
• Levels were higher in Black and Mexican American children than white children. There is some uncertainty about this conclusion because the samples taken from people of different ethnic origins were from overlapping but not identical distributions. Hence the ethnic differences may also reflect geographic differences.

The key finding is that mercury levels in almost 8% of women of child-bearing age exceeded the target level of concern established by the National Research Council, 5.8 µg/L. This value had been established first by determining what is called a "benchmark dose" and then by building in an safety factor of 10 because of uncertainties in the calculation.

The benchmark dose (BMD) recommended by the NRC was 58 µg/L. Their BMD recommendation was determined from studies of the developmental effects of mercury on children living (and studied) on Faroe Island, in the North Sea north of England. They calculated the BMD as the lower 95% confidence limit of the estimated level of mercury in maternal blood that was associated with a doubling of the proportion of children with an abnormal score on developmental test used in the study.

What that means, in essence, is that beneath the reference dose there is less than a 5% chance that a given level of exposure could cause a doubling in the abnormal score.

It represents an attempt to acknowledge there are uncertainties in the calculations but that simultaneously, there is a need to recommend a threshold level. The choice of "doubling" is somewhat arbitrary... it could have been "increase by 50%," in which case the threshold would have been lower. It's not clear that parents, for example, would be satisfied with the focus on "doubling" the likelihood of abnormal scores.

Adding a safety factor of 10 to the BMD of 58 µg/L yields a level of concern of 5.8 µg/L. The US EPA accepted this recommendation.

What does it mean? Perhaps at first blush, the finding that over 90% of American women of child-bearing age have mercury levels beneath the EPA's reference dose of 5.4 µg/L is positive. But the reality is that children exposed to excessive mercury in the womb enter life with neurodevelopmental disabilities that put them at a permanent cognitive disadvantage. Thus with some 4 million live births each year, the fact that approximately 8 percent exceed the reference dose means that up to 320,000 babies are placed at risk each year. That number is unacceptable.

Two considerations heighten this concern.

First, the history of study of the two best known developmental neurotoxicants of environmental origins, lead and mercury, show that as studies become more sophisticated, the reference dose is repeatedly lowered. The graph on the right depicts how estimates have fallen over time of the threshold at which harmful effects of lead take place. Both figures adapted from "In Harms Way."

The graph to the left shows how the daily intake levels associated with harmful effects of mercury exposure (yellow squares) have changed over the past 3 decades, and how agency regulatory standards (red triangles) have gradually come down in response.

There is no a priori reason to think that 5.4 µg/L in a woman's serum is the end of adjustments for mercury.

Second, a lower FDA reference dose will significantly broaden the range of species of fish with health warnings, increase the number of places from which caught fish should not be eaten, and decrease the amount of each type of fish that children and women of child bearing age should consume. Tragically, this contamination is forcing people to remove from their diets an important source of nutrients that for many is an affordable source of protein and oils. It hits especially hard upon people whose culture ties them closely to native diets rich in fish and other top predators in the marine environment, or whose economic circumstances force them to depend upon fishing as a source of protein. But as other research has shown, and as is indicated by the link between higher education and higher mercury levels in the current study (above) high-end diets also create mercury exposures that undermine health.