revised final - Agency for Toxic Substances and Disease Registry ...

MERCURY 236
2. HEALTH EFFECTS
Nakagawa (1995) analyzed total mercury in hair samples from 365 volunteers in Tokyo, and reported
higher mercury levels in those who preferred fish in their diet, compared to those who preferred other foods
(preference choices were fish, fish and meat, meat, and vegetables). The mean hair mercury levels were
4 ppm in men who preferred fish and 2.7 ppm in women who preferred fish. The lowest hair mercury levels
were seen in men and women who preferred vegetables, 2.27 and 1.31 ppm, respectively. The mean hair
level for the whole group was 2.23 ppm (median 1.98).
Drasch et al. (1997) assayed tissue samples of 150 human cadavers (75 males, and 75 females) from a
“normal” European (German) population, i.e., there were no occupational or higher than average exposures
to metals found in any of the biographies of the deceased. The objective was to evaluate the validity of
blood, urine, hair, and muscle as biomarkers for internal burdens of mercury, lead, and cadmium in the
general population. All individuals died suddenly and not as a result of chronic ailments. Age ranged from
16 to 93 years, and every decade was represented by approximately 10 males and 10 females. Tissues
sampled included kidney cortex, liver, cerebral cortex, cerebellum, petrous portion of the temporal bone,
(pars petrosis ossis temporalis), pelvic bone (spina iliaca anterior-superior), muscle (musculus gluteus),
blood (heart blood), urine, and hair (scalp-hair). Statistically significant rank correlations between
biomarker levels and tissues were observed, but with large confidence intervals for the regressions. The
authors conclude that specific biomarkers relative to each metal are useful in estimating body burdens and
trends in groups, but are not useful for determining the body burden (and therefore the health risks) in
individuals. A notable exception was for correlation to brain mercury. By comparison to a generally poor
correlation of cadmium, lead, and mercury between hair and tissue, there was a strong correlation between
mercury in hair and mercury in brain (cerebrum and cerebellum). The authors state that this may be due to
the high lipophilicity of elemental and short-chain alkyl mercury compounds. As seen in other studies
comparing European to Japanese hair mercury levels, the mercury hair levels reported by Nakagawa (1995)
of 2–4 ppm for a Japanese population are 10–20 times higher than total mercury levels observed in the
Drasch et al. (1997) study (median, 0.247 µg/g in hair; range, 0.43–2.5 µg/g).
Other studies have confirmed a good correlation between hair mercury and brain mercury levels. In a study
on the Seychelles Islands cohort, Cernichiari et al. (1995b) compared maternal hair levels, maternal blood
levels, fetal blood levels, and fetal brain levels. Autopsy brains were obtained from infants dying from a
variety of causes. The concentrations of total mercury in six major regions of the brain were highly
correlated with maternal hair levels. This correlation was confirmed by a sequence of comparisons among
the four measurements. Maternal hair levels correlated to maternal blood levels (r=0.82) and infant brain