State of Science on the Safety of Amalgam and
Other Restorative Materials

One of the implicit commitments made in the 1993 USPHS report on dental amalgam was to keep abreast of scientific advances with a clear or potential bearing on national policy regarding the continued use of amalgam in contemporary dental practice. USPHS agencies have made good on this pledge to keep an open mind on the science relating to amalgam safety in particular, and have done so in a number of ways, as described below.

USPHS-Sponsored Workshops on Mercury and Amalgam Safety

The USPHS has long recognized that a substantial amount of basic and applied research with direct and peripheral relevance to the public health risks of dental amalgam is ongoing in both the private and public sectors. As a means of monitoring progress in both research arenas, the USPHS has on two occasions sponsored workshops to facilitate cross-sharing of information as well as to provide direction for Federally-funded mercury-related research.

The first of these occurred on September 16, 1994. Sponsored by the National Institute of Environmental Health Sciences, National Institute of Dental Research and the NIH Office of Minority Research, the workshop focused chiefly on methods for detection of health effects of low-level exposure to different forms of mercury. The 1-day event was organized around five areas of interest:

  1. methods for assessing mercury exposure;
  2. developmental effects of mercury exposure on the central nervous system and target organs;
  3. methods for evaluating health effects;
  4. modulating effects of mercury exposure; and
  5. need and directions for clinical research.

Bringing together researchers from around the United States, Canada, and Sweden, the workshop addressed the following issues.

Because mercury health issues are particularly complex given the involvement of organic or methyl mercury from consumption of fish and mercury vapor released from dental amalgam, the workshop focused on a number of scientific questions regarding the nature of the health effects that result from exposure. In addition, discussions centered on methods to detect these effects on susceptible populations and on the degree to which effects from exposure to different forms of mercury overlap.

Most of the research regarding mercury exposure and health effects reported at the workshop concerned effects of methyl mercury from ingestion of fish and occupational exposure to mercury vapor. Similarities and differences in metabolism for these two species were described. However, there was no specific information reported that indicated a health risk from mercury from dental amalgam.

In light of this information and the continuing need to confirm the presence or absence of a low-level health hazard, the workshop recommended the following:

Research activities currently being supported by NIEHS include mechanisms of effects of mercury vapor on the nervous system in rodents and non-human primates, and epidemiological studies on effects of methyl mercury on cognitive and behavioral development in children, immunotoxicity of mercury and chelation of mercury.

The second workshop was held on June 28, 1996, under the auspices of the Environmental Health Policy Committee's dental amalgam working group. Its focus was three-fold:

  1. appropriate biomarkers of exposure to mercury vapor;
  2. development of physiologically-based pharmacokinetic models for mercury vapor exposure to predict tissue levels and resultant toxicity; and
  3. appropriate neurobehavioral endpoints to identify neurotoxic effects of mercury.

As with the first workshop, invited experts were drawn from the international scientific community. The following is a synopsis of the salient discussions for each of the three targeted issues.


Biological monitoring of mercury has generally focused either on the determination of mercury concentrations in urine, blood, or other tissues, or on clinical assessments of toxic effects to the renal and neurologic systems. Each endpoint appears to have disadvantages. Tissue mercury measurements, for example, may reflect only the most recent mercury exposures and provide little information on possible human biological effects. In contrast, clinical measures of mercury toxicity, while reflecting bioeffects, are not sufficiently sensitive or specific to serve in the predictive evaluation of preclinical events associated with low-level mercury exposure.

A leading U.S. researcher described investigations of metal-induced changes in urinary porphyrin profiles as biomarkers of exposure and effect. (Porphyrins are intermediates in the biosynthesis of heme. In renal tissues, porphyrin precursors are produced in excess of what is required for heme biosynthesis and excreted in the urine as porphyrins.) The researcher has identified a highly distinctive change in urinary porphyrin excretion patterns based on mercury-induced alterations in renal heme biosynthesis following exposure of experimental animals and humans to various forms of mercury. In animal studies, methyl mercury administered chronically in drinking water produced unique changes in urinary porphyrin excretion patterns and were highly correlated with the dose and time course of the mercury exposure and with the renal mercury content during and following cessation of the exposure.

The changes observed in animals were similarly observed in dentists who were occupationally exposed to high levels of elemental mercury vapor and were significantly correlated with specific indices of neurobehavioral function. Also, studies performed on dental workers in Mexico demonstrated that urinary porphyrins were significantly associated with mercury body burden. Currently available data are not adequate to determine whether urinary porphyrin levels are specific and sensitive measures of mercury exposed and bioeffect. Although this area is promising, more research is needed, particularly in determining functional and/or clinical changes associated with specific levels.

Physiologically-Based Pharmacokinetic (PBPK) Modeling

Significant data gaps limit our ability to determine the amount of mercury that reaches sensitive target tissues following human exposure to mercury vapor. Specifically, most exposure studies involve levels of mercury vapor experienced by patients receiving mercury fillings. As a result, it is not clear whether the exposure-response curve below the occupational exposure level is continuous and it is not correct to assume that it is. This lack of knowledge hinders the assessment of risk in mercury vapor exposures.

Physiologically-based pharmacokinetic (PBPK) models utilize a series of mathematical equations that model the pharmacokinetic behavior of compounds in various organs and tissues. PBPK models have the potential to predict target tissue doses of toxicants, and thus reduce the uncertainties in risk assessment, by incorporating existing data and providing a means to conduct high-to-low doses extrapolation, as well as interspecies, intra-species and route-to-route extrapolations of dose. These types of models may also guide future research by identifying data gaps. Unlike traditional compartmental pharmacokinetic models, PBPK models utilize physiological and biochemical parameters and data to predict target tissue doses under varying conditions of exposure. A PBPK model for mercury vapor may reveal specific non-linearities between inhalation exposures to mercury and subsequent mercury levels in target tissues.

The current PBPK model for mercury vapor behavior presented at the workshop holds promise in that it has been shown to accurately predict many of the kinetic properties of mercury vapor, such as exhaled breath concentration, time course of relative amounts of total mercury in red blood cells and plasma, and the percent of total absorbed mercury vapor excreted as divalent mercury in urine and feces over time. However, the model did not predict the pharmacokinetic behavior of all mercury vapor metabolites. With further refinements, the PBPK model may become an important tool for risk assessment in order to fill in current gaps in the exposure-response curve for patients and other individuals who experience very low levels of exposure from dental amalgam fillings.

Neurobehavorial Endpoints

The most commonly observed effect following occupational mercury vapor exposure is altered neurologic function. A remarkable diversity of observational experimental and psychometric tests have been used to assess various neurobehavorial endpoints in humans and laboratory animals. At the workshop, two university-based researchers outlined a systematized grouping of endpoints (i.e., by type of function) for which valid test methods exist. One involved a battery of tests to detect deficit patterns relating to intellectual function, memory, coordination, and hand tremor. Although the test battery has been validated and shown to be reliable, questions remain as to whether it is sufficiently sensitive to determine a Lowest Observed Adverse Effect Level (LOAEL) in humans. Moreover, the neurologic endpoints are not necessarily specific for mercury vapor exposure nor have they been validated for low level exposures to mercury vapor from dental amalgams.

A second academic researcher described the use of conditioned behavior to investigate effects on learning (i.e., cognitive effects) and motor behavior. More research, on the animal level, is needed to determine the applicability of such testing in humans.

Workshop Conclusions

Both workshops yielded interesting and thought-provoking "leads" insofar as future research directions are concerned. However, no definitive answers to the central question of whether mercury vapor exposure, at typical levels associated with dental amalgam restorations, pose an unacceptable health risk to the general population.

Review of Scientific Literature In Support of Citizen Petitions

Earlier in this decade, a consumer advocacy group filed a lawsuit which petitioned the court for a writ of mandamus alleging that the Food and Drug Administration had ignored safety issues relating to the use of mercury in dental fillings and failed to take appropriate regulatory action. On January 29, 1993, the U.S. Court of Appeals for the District of Columbia dismissed the lawsuit because the petitioner had not exhausted all available administrative remedies. Subsequently, the same group, followed by two others, filed administrative petitions with FDA seeking various regulatory actions relating to dental amalgam, ranging from a total product ban to specific use restrictions for certain patient sub-populations.

In support of their requests, the petitioners collectively cited 175 literature articles. The Working Group on Dental Amalgam formed an ad hoc review group to evaluate the articles for their scientific merit in two contexts. In the first instance, the FDA was interested in obtaining an independent and interdisciplinary review of the material to determine if the scientific basis for the petitions, as purported by the submitters, was credible and valid. Secondly, independent of the petitions, the Working Group was interested in determining whether the science cited by the petitioners, in whole or part, shed any new light on the safety of amalgam and past risk assessments performed by USPHS and others.

The referenced articles were an assortment of publications from peer reviewed journals, non-refereed publications, untranslated foreign sources and articles culled from the print media. Therefore, the review group first performed a triage in order to focus its evaluation on those studies that met criteria established by the review group, which were based on whether the studies were of sufficient quality to review, whether they had been peer reviewed, and whether they were referenced in the 1993 USPHS report. This winnowing process reduced the number of articles to 57. The surviving publications were then clustered by study type (i.e., general toxicology, neurotoxicology, immunotoxicology, epidemiology, dental/clinical materials) and disseminated to 12 scientific specialists and dental professionals recruited from various USPHS agencies.

The results of the individual reviews, which were aggregated in 1997 into a single synopsis, are highlighted in the following paragraphs. The reviews centered on five major areas of concern: adequate controls; methodological flaws; mercury exposure measurements; relevance of the article to dental amalgam safety assessment; and fetal mercury exposure.

None of the reviewers suggested that any study under review would indicate that individuals with dental amalgam restorations would experience adverse health effects. In general, the reviewers found that many of the studies involving human subjects were found to be not well-controlled; thus, the conclusions of those studies were of limited use. Control groups were characterized as inappropriate, too small or poorly described. In some cases, the reviewers felt that the most appropriate control groups were not included. Many variables in the human studies which could have affected the results, e.g., diet, health status, atopy and mercury exposure, were not accounted for.

There was also a lack of data describing dose-response relation-ships between mercury exposure via dental amalgam fillings and adverse health outcomes, including neurological and renal disease. Doses in some experimental studies were of such magnitude that the reported adverse effects would not be unexpected. In some cases, experimental and control groups were often too small to permit conclusions to be drawn to the general population. In a few cases, the studies were also found to lack adequate descriptions of methodology and were inadequately designed.

With respect to assessment of mercury exposure, the reviewers found much of the information to be ill-defined or poorly quantified, thus making the results difficult to relate to the magnitude of exposures. Many studies, including those with a high degree of scientific rigor, were conducted for hazard identification purposes, and therefore employed exposures to high concentrations of mercury. Mercury levels determined from spot urine collections were also determined to be less reliable measures of exposure than multiple or cumulative urine collections. Cases of non-discrimination between exposures to mercury vapor versus methyl mercury in the diet were evident. Room air sampling devices were used when personal mercury monitors are a more appropriate method for assessing mercury exposure in individuals.

On the issue of fetal exposure, the few relevant studies evidenced that mercury can pass from a mother to her fetus, but the relationships of maternal burden and fetal burden were often inconsistent and, in any event, no adverse effects were associated with exposures that mimicked those from dental amalgam.

A number of the research papers were considered inappropriate by the reviewers, e.g., they were not related to amalgam exposure or health effects. Some of the papers which described in vitro studies demonstrating that mercury is toxic to cell cultures did not provide adequate rationale to extrapolate the in vitro data to risks of human health effects.

In summary, the reviewers' analyses, like the 1993 USPHS report, supported the conclusions that mercury is a well-known toxicant, that its toxicity is dependent on dose, and that mercury from amalgam fillings can accumulate in tissues. The analyses also supported the report's conclusion that mercury is an allergen sensitizer in susceptible strains of experimental animals and in some humans; however, the data did not support claims that dental amalgam has a causative role in oral disease, e.g., oral lichen planus. The reviewers' analyses did not support claims that individuals with dental amalgam restorations will experience adverse health effects, including neurologic, renal or developmental effects. Finally, it should be noted that no single study had all of the short-comings discussed above, but when considered in the aggregate, the data reviewed did not demonstrate to the reviewers that adverse human health effects would occur as a result of exposure to dental amalgam.

International Symposia on Mercury Health Effects

The Working Group on Dental Amalgam has maintained a world-wide watch on scientific developments with potential relevance to the assessment of amalgam safety. A case in point is a meeting in July 1997, sponsored by the International Programme for Chemical Safety of the World Health Organization. International experts, including the Working Group's NIEHS representative, assembled in Montreal, Canada to discuss the strengths and weaknesses of various methodologies to assess the neurobehavorial development of children exposed to methyl mercury in utero. The emphasis of the meeting was to review the comparability of the results from various models, the domains measured and the power of the tests to predict latent adverse effects when performed at various time points in child development. While the meeting focused on tests currently being used to evaluate effects of in utero exposure to methyl mercury from maternal consumption of fish, the results of the meeting have relevance to neurodevelopmental effects of elemental mercury or mercury vapor exposures people experience from dental amalgam fillings. The proceedings and information from the meeting will be shared with PHS researchers as a means to ensure consistency in ongoing and planned Federal research.

Risk Information on Non-Amalgam Alternative Materials

In comparative terms, the amount of scientific and risk assessment information for non-amalgam replacement materials vis--vis dental amalgam is quite sparse. The issue, quite obviously, is one of major significance given the potential for major shifts in dental practice away from amalgam and toward alternative materials whose toxicological properties are far less well understood and could prove far riskier than amalgam itself. For this reason, the EHPC Working Group on Dental Amalgam has continued to consult researchers, health professionals and government policymakers around the world in search of information on the relative safety of non-amalgam filling materials.

Return to Table of Contents

Return to Committee Reports Page