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

MERCURY 315
2. HEALTH EFFECTS
When the inorganic mercury value was tested, significant correlations remained, with weaker coefficients
in all the organs but the spleen. Stepwise multiple regression analysis indicated that hair organic mercury
value was the major correlating variable for the organ total mercury or organ methyl mercury value in all
the organs. With respect to the organ inorganic mercury value, the hair organic mercury value was the
major correlate for the cerebrum and kidney (both cortex and medulla), the hair inorganic mercury value
was the major variable for the cerebellum and heart, and the hair phosphorous and hair organic mercury
were the major variables for the liver. No explanatory variable existed for the spleen. Auxiliary
correlating variables accounted for the organ total mercury and inorganic mercury levels, among which
the hair selenium value was conspicuous and with negative regression coefficients.
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, 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, that in 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