Good afternoon. I am Barry Johnson, PhD, Assistant Administrator of the
Agency for Toxic Substances and Disease Registry (ATSDR). The
subcommittee invited ATSDR to testify on regional haze and mercury
pollution. Our agency has had no involvement with regional haze but has
worked on various mercury issues. Much of our work on the scientific
issues of mercury pollution is reflected in the ATSDR draft Toxicological
Profile for Mercury.
ATSDR has developed a toxicological profile on mercury in compliance
with a mandate in the Comprehensive Environmental Response,
Compensation, and Liability Act, as amended (CERCLA, or Superfund).
Section 104(i)(3) directs us to develop toxicological profiles for priority
hazardous substances released from Superfund sites. Our priority list of
hazardous substances is developed jointly with EPA and updated every
2 years. Mercury is number three on the 1997 list. ATSDR is also
required by CERCLA (.104(i)(3)) to revise and republish individual
toxicological profiles as necessary, but no less often than once every 3
CERCLA also requires that ATSDR's toxicological profiles contain "An
examination, summary, and interpretation of available toxicological
information and epidemiologic evaluations on a hazardous substance in
order to ascertain the levels of significant human exposure for the
substance and the associated acute, subacute, and chronic health effects."
This language directs ATSDR to develop numerical estimates of health
risks posed by hazardous substances. Health assessors and risk managers
use numerical values to characterize the toxicities of hazardous
substances. Risk assessment methods are most often used for
carcinogenic substances. This results in point estimates or ranges of
health risk that are based on various exposure scenarios. ATSDR uses the
numerical minimal risk levels (MRLs) for noncarcinogenic toxicities of
An MRL is an estimate of what level of daily human exposure to a
hazardous substance is likely to be without appreciable risk of adverse
noncancer health effects over a specified duration and route of exposure.
These substance-specific estimates are intended to serve as screening
levels. Public health assessors use MRLs to identify contaminants and
potential health effects that may be of concern at hazardous waste sites.
It is important to note that MRLs are not intended to define clean-up or
action levels for EPA or other agencies.
ATSDR develops MRLs during the preparation of toxicological profiles.
They are derived when ATSDR determines that reliable and sufficient
data exist to identify the target organs of effect or the most sensitive
health effects for a specific duration of exposure. MRLs are based only
on noncancer health effects and not on a consideration of cancer effects.
ATSDR first published a toxicological profile on mercury in 1989. The
mercury profile was then updated in 1994, and a second update was
released in a draft version in October 1997.
ATSDR's 1994 mercury profile contained a chronic duration, oral
exposure MRL that was based primarily on a 1989 analysis of data
collected on persons in Iraq who had been accidentally exposed to
methylmercury in grain during the early 1970s. That MRL was 0.1
micrograms of mercury per kilogram body weight per day ( g/kg/day).
It was numerically equal to EPA's reference dose (RfD), and was based
on the same neurodevelopmental endpoint (children's delayed walking
and talking) that EPA used to derive their RfD for methylmercury.
The CERCLA mandate to update the mercury profile, coupled with the
need to increase our knowledge of the health effects associated with
mercury, led us to convene a series of meetings in Atlanta in 1994 and
1995. We invited peer scientists to join us in panel meetings to review the
direction we should take in our continuing assessment of the health
impact of methylmercury. At that time, we were cognizant of a number
of ongoing studies, including studies in the Faeroe Islands and the
In 1995 our panel of experts recommended that we await the development
of the Seychelles data and use them as a starting point in our mercury
reassessment efforts. So we waited until 1996, when published data from
the Seychelles study began appearing in the scientific literature, to begin
updating our mercury toxicological profile. In October 1997, ATSDR
released for public review and comment our current draft profile. The
document remains in draft, pending further discussions with EPA, other
federal agencies, and the public. An upcoming interagency workshop in
November to evaluate the major scientific studies on methylmercury and
its developmental effects in children will be a key forum for resolving
some remaining points of science and public health.
ATSDR's MRL for chronic, oral exposure to methylmercury in the
October 1997 draft toxicological profile is derived from a study
conducted by University of Rochester investigators in the Seychelles
Islands that reflects multiple generations of human exposure to organic
mercury through fish as a primary route of exposure. Because of the long-term
nature of this exposure, the large sample size, and the rigorous study
design, this data set was used as the primary basis of ATSDR's
evaluation. The Seychelles study overcomes several of the limitations in
the Iraqi study. For example, there is a rather large sample size of 779
mother-infant pairs before the application of the exclusion criteria. This
was a prospective study, with the goals and objectives stated before data
In the Seychelles study, children were evaluated at 6.5, 19, 29, and 66
months of age. Through the age of 29 months, no effects attributable to
methylmercury exposure were found using a battery of neurobehavioral
and neurophysiological tests. The only endpoint that was correlated in
any way with mercury exposure was the subjective observation by several
examiners that some boys, but not girls, showed a decreased activity level
during the testing period. This decrease in activity associated with an
increase in maternal hair mercury levels was, however, not considered by
the Rochester team to be attributable to mercury exposure, and was
observed only in boys whose mothers had hair mercury levels above the
median value (5.9 ppm) used as a no-observed-adverse-effect level
(NOAEL). This value is 2 4 times less than the level at which the
transient depressed activity levels in young boys were noted (i.e., >12 26
ppm). Similarly, the Rochester researchers found no statistical association
between prenatal exposure to mercury and the age at which the children
in the Seychelle Islands study walked or talked.
Since October 1997, there have been additional scientific publications on
the human health effects of mercury. For example, results of the 66-month testing of children in
the Seychelle Islands are now available.
Further, a study of the Faeroe Islands population published in December
of 1997 will need to be examined by ATSDR in the context of our draft
A scientific evaluation of studies like those conducted in Iraq, the Faeroe
Islands, and the Seychelles is complicated. All studies like these have
particular strengths and limitations. One must evaluate each study for its
statistical power, the adequacy of data collection and analysis, the
relevance of exposure data, biological plausibility, and relevance of health
findings. Examination of all currently available scientific information
must be concluded and throughly debated by peer scientists before final
pronouncements on made MRLs and similar health guidance values. The
results from studies published since October 1997 will be carefully
reviewed by ATSDR and incorporated in our final version of the
toxicological profile on mercury.
As previously noted, our toxicological profiles require developing MRLs,
the derivation of which is a rather straightforward algorithm. It is
deliberately analogous to what the EPA and the Food and Drug
Administration (FDA) do for RfD and ADI derivations, respectively. You
have a benchmark of toxicity, no-observed-adverse-effect level
(NOAEL), or some other surrogate for that, and an uncertainty factor. The
higher the uncertainty factor, the lower the overall quality of the data set.
The higher the quality of the data set, the lower the uncertainty factor.
Although the operational derivation of an MRL is straightforward, any
derivation involves a substantial amount of professional judgment. In the
case of mercury, we must consider the fact that mercury is ingested by
mothers, yet we know the fetus is the target of concern. The mercury
concentrations are measured in the mothers' hair, but we are really
concerned about the concentration of mercury in the blood reaching the
fetus. So we have to convert the maternal hair mercury concentration to
a blood mercury concentration, and subsequently convert that blood
mercury concentration to an oral daily intake.
In addition, we have the issue of uncertainty. ATSDR considers four
general areas of uncertainty in the derivation of MRLs: cross-species
extrapolation, 1 to 10; intra-species variability, 1 to 10; the use of an
adverse effect level as opposed to a non-adverse effect level, 1 to 10; and
a factor to account for the quality of sufficiency of the overall database,
1 to 10, sometimes referred to as a modifying factor by ATSDR.
In our draft 1997 mercury profile, we looked at the uncertainty and the
available mercury data. We have the most sensitive population, the
developing fetus. We have a known and relevant route of exposure,
consumption of mercury-contaminated fish. We have identified a NOAEL
in the most sensitive subpopulation, and we see an absence of any
neurodevelopmental deficit at similar exposure levels in other
populations. That led us to select an uncertainty factor of 1. This
uncertainty factor is consistent with the standard application of those
factors in the derivation of health guidance values. This is not to suggest
that there is no uncertainty remaining about any threshold for the health
hazards of methylmercury, or that ATSDR's proposed MRL does not
reflect a margin of safety consistent with adequate precautionary
approaches and good public health practice.
From the foregoing considerations, ATSDR derived an MRL for chronic,
oral exposure to methylmercury of 0.5 g/kg/day. MRLs for both
elemental mercury and inorganic mercury (salts) are also presented in the
draft mercury toxicological profile. The significant neurotoxic (nervous
system) and nephrotoxic (kidney) health hazards posed by these forms of
mercury are often addressed by ATSDR in response to spills or other
unplanned releases of mercury in schools, hospitals, and homes. In fact,
mercury has been the single most frequently encountered chemical in our
emergency response program for the last 8 years. For this reason, ATSDR
and EPA jointly developed and released a health alert in the summer of
1997 that has been widely distributed to schools and other potential
targets of mercury spills.
In July of 1998, ATSDR assembled a panel of 18 experts from the
scientific and medical communities to assist the agency in reviewing key
issues in the areas of toxicology and human heath risk assessment as they
relate to metallic, inorganic, and organic mercury compounds. The
purpose of the meeting was to review all relevant science on mercury and
its compounds, including methylmercury, and to react to ATSDR's
proposed response to comments we received during the public review and
comment period. The meeting of the panel was held as an announced
public meeting. It consisted of scientists from the EPA, FDA, the Centers
for Disease Control and Prevention (CDC), academia, and the private
sector, including the teams of scientists who conducted the human
epidemiological studies in Iraq, the Seychelles, and the Faeroe Islands.
The discussions from this expert panel are being used by ATSDR and
other federal agencies to help form part of the agenda for the November
1998 interagency workshop on mercury.
ATSDR continues to work with other federal agencies to reach a
consensus on mercury issues. The Committee on Environment and
Natural Resources (CENR), Subcommittee on Toxics and Risks,
established a Mercury Working Group in September of 1997 to look at a
number of issues concerning mercury. This workgroup, which includes
representation from ATSDR, EPA, FDA, CDC, NIH, NOAA, USDA, and
the Office of Science and Technology Policy, has already come to a
consensus on a range of issues, and continues to work towards agreement
on all public health issues concerning mercury.
The CENR Working Group is sponsoring the November 18 20
interagency workshop in Research Triangle Park, North Carolina, to
further elucidate issues regarding methylmercury exposure and public
health. The research teams responsible for the Seychelles and Faeroes
studies, as well as a team conducting similar studies along the Amazon
River basin, will present and discuss their findings at this meeting.
ATSDR considers this meeting to be an important step toward resolving
remaining scientific issues. The meeting will be used by ATSDR to help
bring closure to its toxicological profile on mercury.
A challenge for health officials is to balance the known public health
benefit of consuming more fish in the diet, and the known dangers of
excess mercury exposure. To mitigate adverse health effects of excessive
exposure to mercury, ATSDR supports efforts to reduce or eliminate
exposure to mercury in the environment. Such efforts must be pursued
through pollution prevention strategies, including health education for
both health care providers and the citizens who may be at risk due to high
levels of exposure to not only organic, but also inorganic and metallic
mercury. Throughout the profile revision period, ATSDR has advised the
public that it would be premature to predicate any risk management
decisions on information in a draft document. That needs to await the
sorting out of all relevant issues and the finalization of the profile.
I thank you for this opportunity to present to you some of the conclusions
in ATSDR's Toxicological Profile for Mercury and to discuss our
ongoing efforts to enhance the accuracy and utility of our mercury MRLs,
including the chronic oral MRL for methylmercury.
Mr. Chairman, we would be pleased to answer any questions that you or
subcommittee members may have.