Environmental Engineering Reference
In-Depth Information
643 children followed up to 17 years of age did not iden-
tify neurodevelopmental defi cits associated with prenatal
exposure (Davidson et al., 1998, 2010; Myers et al., 2003).
In the Seychelles, exposure occurred primarily through
maternal consumption of ocean fi sh. The arithmetic mean
total mercury concentration in maternal hair at delivery
was 6.9 ppm (standard deviation, 4.5). Proposed explana-
tions for the discrepancies between the two studies include
hazardous levels of co-occurring compounds, such as poly-
chlorinated biphenyls (PCBs), in pilot whale; co-occurring
protective substances in ocean fi sh; different exposure
patterns (intermittent, high doses from pilot whales, as
compared with steadier and lower doses from ocean fi sh);
population-specifi c toxicokinetic differences in response
to exposure; choice of biomarker (mercury in hair or cord
blood); and the sensitivity of the different neurodevelop-
mental tests used in each study (NRC, 2000).
One of the diffi culties of studying the effect of expo-
sure to methylmercury through fi sh consumption is that
fi sh also contain nutrients that are benefi cial to neurode-
velopment and overall health. An observed mercury effect
may be offset, or confounded, by co-occurring, benefi cial
nutrients such as long-chain omega-3 fatty acids. More
recently, studies of prenatal exposure have attempted to
separate mercury risks from nutrient benefi ts by estimat-
ing a mercury effect while controlling for the nutrients in
fi sh—or frequency of fi sh meals (Oken et al., 2005, 2008b;
Budtz-Jorgensen et al., 2007; Lederman et al., 2008; Strain
et al., 2008). A second cohort study of prenatally exposed
children from the Seychelles was designed specifi cally to
separate risks from benefi ts (Davidson et al., 2008). Results
from these studies support the hypothesis that nutrients in
fi sh can confound observed associations between prenatal
exposure to methylmercury and child development. Both
benefi cial and adverse effects appear to be strengthened
when the effect of each exposure is assessed while simulta-
neously controlling for the other.
The concern that an observed association between expo-
sure to methylmercury and adverse neurodevelopmental
outcomes in fi sh-eating populations may be attributed to
co-occurring PCBs in fi sh has also been raised. However,
results from analyses that have considered both sets of
compounds simultaneously lend support to the hypoth-
esis that methylmercury has an independent effect (Budtz-
Jorgensen et al., 1999; Suzuki et al., 2010).
Mercury has been measured in breast milk, and mater-
nal fi sh or seafood consumption can raise levels in a breast-
feeding child (Grandjean et al., 1994). However, infants
from the Faroe Island cohort who were breast-fed—and
therefore had higher hair mercury concentrations (geomet-
ric mean, 1.8 ppm)—achieved developmental milestones
earlier than children who were not breast-fed (Grandjean
et al., 1995). Methylmercury passes into breast milk less
readily than inorganic forms of mercury (Skerfving, 1988;
Bjornberg et al., 2005). Furthermore, breast milk con-
fers many health benefi ts to a child, including improved
neurodevelopmental outcomes (Kramer et al., 2008; Oken
et al., 2008a). Most, but not all (Davidson et al., 2010),
analyses from the Faroe Islands and the Seychelles cohorts
that have looked at effects of postnatal mercury exposure—
independent of prenatal exposure—have not reported
poorer developmental outcomes, underscoring the relative
importance of earlier exposure (Debes et al., 2006; Myers
et al., 2009).
Several studies in the Amazon and other regions of
Brazil have examined effects of methylmercury exposure
in adults consuming contaminated fi sh. Mercury has been
used in these areas for mining gold, which has contrib-
uted to its accumulation in local fi sh. Deforestation prac-
tices have also contributed to the release of mercury into
the environment in this region (Passos and Mergler, 2008).
Results from one study of 91 adults reported subtle defi cits
in visual and motor function associated with an average
hair mercury concentration of about 13 ppm (Lebel et al.,
1998). Another study of 129 adults with average hair con-
centrations of about 4 ppm reported subtle decreases in fi ne
motor function and verbal memory (Yokoo et al., 2003).
Both studies were cross sectional and could not rule out
confounding by exposure at an earlier age. In a case series
of high-end fi sh consumers in the United States, adults and
children with total blood mercury levels ranging from 2.0
to 89.5 µg/L reported a variety of nonspecifi c symptoms
including fatigue, headache, decreased memory, decreased
concentration, and muscle or joint pain
(Hightower and
Moore, 2003). However, the extent to which the mercury
was causally associated with these symptoms remains
uncertain, since there was no control group for compari-
son. In an urban US population of older adults, total blood
mercury was not associated with worse neurobehavioral
performance, but average levels were substantially lower
than those measured in the studies from Brazil (Weil et al.,
2005).
There is some evidence that exposure to methylmercury
through fi sh consumption increases the risk of cardiovas-
cular disease in adults. One of the most methodologically
sound studies followed 1871 Finnish men with an aver-
age hair mercury concentration of 1.9 ppm, and reported
increased incidence and mortality for cardiovascular end
points (Virtanen et al., 2007). These associations have
been controversial (Mozaffarian and Rimm, 2006; Stern,
2007), since the long-chain omega-3 fatty acids in fi sh are
widely accepted to offer protection from cardiovascular
disease. Even more provocative are the relatively low expo-
sure levels at which associations were observed. In another
large European study, toenail mercury was positively asso-
ciated with incident myocardial infarction, whereas doco-
sahexaenoic acid (a cardioprotective omega-3 fatty acid) in
adipose tissue was negatively associated with the end point
when the two exposures were modeled simultaneously
(Guallar et al., 2002). This pattern is consistent with the
notion that there may be negative confounding between
mercury risks and nutrient benefi ts associated with fi sh
