Amalgam Risk Assessment finds 120 million Americans over daily safe dose of mercury from amalgam fillings.


Mark Richardson P.h.D.
(previously of Health Canada)
explains the amalgam risk assessment he presented to the FDA in 2010

On December 14 and 15, 2010, the FDA convened a scientific panel to re-examine the issue of mercury exposure from amalgam dental fillings. Two private foundations, assisted by IAOMT, have commissioned G. Mark Richardson, PhD, of SNC Lavallin, Ottawa, Canada, formerly of Health Canada, to provide the scientific panel and FDA regulators with a formal risk assessment using the latest information from the scientific literature.

Science of the Total Environment 409 (2011) 4257–4268

Mercury exposure and risks from dental amalgam in the US population, post-2000

G.M. Richardson, R. Wilson, D. Allard, C. Purtill, S. Douma, J. Gravière


Dental amalgam is 50% metallic mercury (Hg) by weight and Hg vapour continuously evolves from in-place dental amalgam, causing increased Hg content with increasing amalgam load in urine, faeces, exhaled breath, saliva, blood, and various organs and tissues including the kidney, pituitary gland, liver, and brain. The Hg content also increases with maternal amalgam load in amniotic fluid, placenta, cord blood, meconium, various foetal tissues including liver, kidney and brain, in colostrum and breast milk. 

Based on 2001 to 2004 population statistics, 181.1 million Americans carry a grand total of 1.46 billion restored teeth. Children as young as 26 months were recorded as having restored teeth. Past dental practice and recently available data indicate that the majority of these restorations are composed of dental amalgam. Employing recent US population-based statistics on body weight and the frequency of dentally restored tooth surfaces, and recent research on the incremental increase in urinary Hg concentration per amalgam-filled tooth surface, estimates of Hg exposure fromamalgamfillingswere determined for 5 age groups of the US population. Three specific exposure scenarios were considered, each scenario incrementally reducing the number of tooth surfaces assumed to be restored with amalgam. Based on the least conservative of the scenarios evaluated, it was estimated that some 67.2million Americans would exceed the Hg dose associated with the reference exposure level (REL) of 0.3 μg/m3 established by the US Environmental Protection Agency; and 122.3million Americans would exceed the dose associated with the REL of 0.03μg/m3 established by the California Environmental Protection Agency.

Exposure estimates are consistent with previous estimates presented by Health Canada in 1995, and amount to 0.2 to 0.4 μg/day per amalgam-filled tooth surface, or 0.5 to 1 μg/day/amalgam-filled tooth, depending on age and other factors.


Mercury (Hg) is globally recognized as a toxic substance with numerous national and international efforts to phase out its use, the most recent being the initiative of the United Nations Environment Programme on a global phase out strategy (UNEP, 2009), for which negotiations began in June 2010. The one lingering exception to this phase out is dental amalgam.

Although now banned in Sweden and Norway (Norway Ministry of Environment, 2007; Sweden Ministry of Environment, 2009), dental amalgam is still a restorative material of choice for the majority of US general dentists for repair of dental caries (cavities) (ADA, 2008a).

It is a solid emulsion composed of a mixture of metals comprising approximately 50% metallicmercury (Hg0) byweight. Formulations vary in their Hg content, ranging from 43 to 50.5% Hg by weight, mixed with a powder of other metals typically containing silver (40 to 70%), tin (12 to 30%), copper (12 to 30%), indium (0 to 4%), palladium (0.5%) and zinc (0 to 1%) (Berry et al., 1994).

Dental amalgam has been used in North American dentistry for perhaps 150 years or more (Clarkson and Magos, 2006) and during that time has been the subject of repeated controversy, often referred to as the Amalgam Wars (Clarkson and Magos, 2006). A brief historical account of its introduction, use and controversy is provided by Molin (1992). Scientific articles regardingamalgam’s potential toxicity date back at least to 1885 (Talbot, 1885). These wars  or debates have been due to the recurring concern for the potential health risks posed by exposure to the Hg used in the manufacture of dental amalgam.

It is now accepted that dental amalgam continuously releases Hg0 which results in exposure in those persons possessing fillings composed of this material (USFDA, 2009).

The quantity of Hg0 released from amalgamis often referred to as ‘minute’ (ADA, 2008b; CDA, 2005) or ‘very small’ (AGD, 2007). However, it is not the dose itself that determines safety, it is how that dose compares to levels considered ‘safe’ or without anticipated harm that determines whether or not the dose is significant with respect to health concern. Irrespective of quantity, aminute dose can present a risk if the substance is sufficiently toxic and received in sufficient dose to exceed a reference level considered ‘safe’.

Dental amalgam has been identified as the largest single source of continuous Hg exposure for members of the general population who possess amalgam fillings (WHO, 1991; Heath Canada, 1996). Previous assessments of dental amalgam have demonstrated that the dose of Hg received as a result of this dental material exceeds what is considered to be a safe or reference dose (see Health Canada, 1995; Richardson and Allan, 1996).

A series of recent studies have reported urinary Hg concentrations  (variably corrected or uncorrected for urine creatinine content) as a function of amalgam filling load (Barregard et al., 2008; Dunn et al., 2008; Melchart et al., 2008; Woods et al., 2007; Bellinger et al., 2006; Dye et al., 2005; Factor-Litvak et al., 2003; Pesch et al., 2002; Kingman et al., 1998). In these and in earlier studies (reviewed by Richardson and Allan, 1996; Health Canada, 1995), the average urine Hg content is consistently greater in groups with amalgam fillings than in those without, and urine Hg content consistently increases as amalgam load increases.

Numerous other studies have also demonstrated that the Hg exposure or concentration increases with increasing amalgam load in the following tissues and situations:

• Due to chewing, brushing and bruxism (Hansen et al., 2004; Ganss et al., 2000; Isacsson et al., 1997; Sallsten et al., 1996; Berdouses et al., 1995; Bjorkman and Lind, 1992; Forsten, 1989; Vimy and Lorscheider, 1985a, b; Berglund, 1990; Svare et al., 1981; Gay et al., 1979);

• In exhaled or intra-oral air of persons with amalgam fillings (Halbach andWelzl, 2004; Skare and Engqvist, 1994; Gay et al., 1979; Svare et al., 1981; Patterson et al., 1985; Vimy and Lorscheider, 1985a,b; Berglund et al., 1988; Jokstad et al., 1992);

• In saliva of persons with amalgam fillings (Fakour et al., 2010; Melchart et al., 2008; Zimmer et al., 2002; Ganss et al., 2000; Pizzichini et al., 2000; Bjorkman et al., 1997; Berglund, 1990);

• In blood of persons with amalgam fillings (Gerhardsson and Lundh, 2010; Halbach et al., 2008;Melchart et al., 2008; Lindberg et al., 2004; Pizzichini et al., 2003; Ganss et al., 2000; Vahter et al., 2000; Kingman et al., 1998; Oskarsson et al., 1996; Skare and Engqvist, 1994; Akesson et al., 1991; Abrahamet al., 1984; Snapp et al., 1989;Molin et al., 1990; Jokstad et al., 1992; Svensson et al., 1992; Herrstrom et al., 1994);

• In various organs and tissues of amalgam bearers, including the kidney, pituitary gland, liver, and brain or parts thereof, (Barregard et al., 1999, 2010; Björkman et al., 2007; Guzzi et al., 2006; Weiner and Nylander, 1993; Nylander et al., 1987, 1989; Eggleston and Nylander, 1987);

• In faeces of amalgam bearers (Engqvist et al., 1998; Bjorkman et al., 1997; Skare and Engqvist, 1994);

• In amniotic fluid, cord blood, placenta, and various foetal tissues including liver, kidney and brain, in association with maternal amalgam load (Palkovicova et al., 2008; Ursinyova et al., 2006; Luglie et al., 2005; Ask-Björnberg et al., 2003; Lindow et al., 2003; Ask et al., 2002; Vahter et al., 2000; Lutz et al., 1996; Drasch et al., 1994);

• In colostrum and breast milk in association with maternal amalgam load (Ursinyova et al., 2006; Ask-Bjornberg et al., 2005; Da Costa et al., 2005; Drexler and Schaller, 1998; Drasch et al., 1998; Oskarsson et al., 1996).

Amalgamfillings are sufficiently significant to personal Hg exposure that the influence of amalgamload on blood and urine Hg concentration can be detected despite moderate occupational Hg exposure, occupational exposure that results in up to about 10 μg Hg/L urine (Skare et al., 1990; Martin et al., 1995; Soleo et al., 1998a; Jokstad, 1990).

To date, at least 14 assessments quantifying Hg exposure from dental amalgam have been published (see Richardson, 2003), estimating Hg dose rather than simply reporting Hg concentrations in urine or other bodily fluids or tissues. However, the last such publication was in 1996 (Richardson and Allan, 1996). To date, no population-based assessment of Hg exposure from dental amalgam specific to the US general population has been undertaken. Dye et al. (2005) provided a statistical analysis of the association between numbers of amalgam filled tooth surfaces and urinary Hg concentrations for US women aged 16 to 49 years, but no dose conversions/ calculations were provided to permit comparison to regulatory reference exposure levels. The quantification of Hg dose associated with dental amalgam is required to complete a proper risk assessment. Determining the amalgam-associated dose can be directly compared to the dose associated with regulatory reference exposure levels (RELs) for Hg0, prescribed for the protection of the health of the general population, such as the reference air concentration (RfC; analogous to REL) published by the California Environmental Protection Agency (CalEPA, 2008), the RfC published by the US EPA (1995), and others.

The most recent assessment of Hg exposure from amalgam for a North American population was prepared for the Canadian Federal Department of Health (Health Canada, 1995; also published as Richardson and Allan, 1996). That assessment combined populationbased data on the frequency of filled teeth in the Canadian population, and specific Canadian data on body weight and other required information to quantify Hg exposure from dental amalgam in Canadians ranging in age from 3 years to N90 years. However, due to its relative age (now some 15 years old) and other factors that may distinguish exposure in the US from that in Canada (availability of social service dental treatment, proportion of population having dental health insurance coverage, etc.), a population-based assessment specific to the US was considered warranted. 

The availability of recent research on the link between dental amalgam and levels of Hg in various bodily fluids and tissues, as well as post-2000 population-based statistical data on the oral/dental health of the US population (from the National Health and Nutrition Examination Survey, or NHANES) provide an unique opportunity to update and improve the quantification of Hg exposure from dental amalgam on a population basis. Table 1 summarizes the data and information of note that are recently available to facilitate an improved exposure and risk assessment.

This paper does not reconsider every aspect of Hg exposure, toxicity, or pharmacokinetics. These topics have been addressed in detail elsewhere (USATSDR, 1999; WHO, 2000, 2003; Health Canada, 1995; Richardson et al., 2009). This paper does not attempt to quantify exposure to mercuric ions (Hg2+) associated with amalgam corrosion, wear and subsequent ingestion. Health Canada (1995; see also Richardson and Allan, 1996) demonstrated that inclusion or exclusion of this ingestion exposure resulted in essentially the same estimates of exposure, indicating that ingestion of amalgam particles and Hg2+ ions is insignificant compared to exposure to Hg0 alone. This paper does not address exposure to methyl Hg that may result from the methylation of amalgam-related Hg in the oral cavity or gastro-intestinal tract (Leistevuo et al., 2001; Heintze et al., 1983; Rowland et al., 1975). 

Finally, this paper does not evaluate nor assess the association of amalgamfillings or Hg exposure to specific diseases or disorders such as Alzheimer’s disease, autism, multiple sclerosis, amyotrophic lateral sclerosis (ALS), or Parkinson’s disease.

This paper is based on a more detailed report submitted to and used by the US Food and Drug Administration (USFDA) as a centerpiece of its December 2010 Expert Panel review of the safety of dental amalgam.


Based on the foregoing report, we have formulated a number of recommendations for further work and research that we believe would benefit the ongoing debate regarding the presence or absence of health effects associated with the Hg0 exposure arising from dental amalgam. 

These recommendations are:

  1. As part of a future NHANES survey, compile data on the specific restorative materials used to fill tooth surfaces within the US population. At the very least, recording whether the material used was amalgam versus some other material should be relatively simple. This distinction is relatively easy as it can be based solely on restoration color (silver versus other).
  2. The USEPA and USATSDR should immediately initiate the review of Hg0 toxicology, including all studies conducted in the past 2 decades, towards updating and revising their RELs for Hg0. This review and update should include consideration of heme synthesis enzyme inhibition as one of the toxic endpoints.
  3. A post-hoc analysis should be undertaken of the statistical power offered by the Casa Pia and New England children’s amalgam trials to quantify precisely the degree of difference in incidence of neurological impairments that can be statistically differentiated between higher exposure subgroups and lower exposure subgroups within the amalgam cohorts of each study.
  4. Quantitatively determine the impact of urinary Hg concentrations in the CAT referent groups (those that received composite resin fillings) relative to the amalgam groups to determine if non-amalgam sources and levels of Hg0 exposure in the referent groups negate any ability to rely on these studies as a means of demonstrating the absence of health effects due to Hg exposure from amalgam. This could include a post-hoc rescreening of referent group members to re-examine inter-group differences employing those referents with a urine Hg concentration ≤ 0.5 μg Hg/g creatinine.
  5. Combine the New England and Casa Pia studies in a meta-analysis, thereby providing increased statistical power for detecting differences in incidence of neurological effects between higher dose and lower dose members of the combined amalgam cohorts.
  6. Conduct a dose-response analysis of both (and combined) amalgam trials data on neurological and other outcomes that appropriately controls for confounders and employs a dose metric that reflects both exposure level and exposure duration, analogous to methods employed to assess porphyrin profiles conducted by Geier et al (in press). Dose-response data must be presented and analyzed with respect to individual CAT participants, and not simply as overall averages for exposed and referent cohorts.
  7. Consider future follow up of both cohorts to increase the data available on duration of exposure, thereby extending the exposures to more effectively represent true chronic exposure, particularly given Hg’s accumulation in the brain and other tissues over time (i.e., to exceed 5 and 7 years for the New England and Casa Pia amalgam trials, respectively).
  8. Clarify the average numbers of amalgam filled tooth surfaces possessed by the different cohort groups that should be considered as in-place for the full duration of the CAT studies. It is apparent that members of these cohorts had varying numbers of amalgam fillings throughout the duration of these studies. A more detailed dose response analysis of these data, as described in point 4, could make this unnecessary, however.
  9. Explicit publication of the urine Hg concentration data from the Casa Pia study, with an analysis of the association of urine Hg concentration with amalgam load.
  10. Efforts should be expended to find an appropriate reference group for future CAT studies that are free of mercury exposure, not just free of amalgam.

Why was this Report Prepared?

To date, no population-based assessment of Hg exposure from dental amalgam specific to the US general population has been undertaken. The quantification of Hg dose associated with dental amalgam is required to complete a proper risk assessment. Determining the amalgam associated dose can be directly compared to the dose associated with regulatory reference exposure levels (RELs) prescribed for the protection of the health of the general population. Such RELs are published by the USEPA (1995), the USATDSR (1999), the California EPA  (2008), and others; these RELs are discussed in greater detail later in this report. 

The final work is presented here in two parts. 

Part 1 is titled UPDATING EXPOSURE, REEXAMINING REFERENCE EXPOSURE LEVELS, AND CRITICALLY EVALUATING RECENT STUDIES.  “…it was determined that some 67.2 million Americans would exceed the Hg dose associated with the REL of 0.3 ug/m3 established by the US Environmental Protection Agency in 1995, whereas 122.3 million Americans would exceed the dose associated with the REL of 0.03 ug/m3 established by the California Environmental Protection Agency in 2008.”

Published Estimates of Hg Exposure in Adults With Dental Amalgam (Mercury Fillings)


UPDATE 2/14/2011 From Mark Richardson:

I indicated during our interview that ATSDR considers amalgam to be ‘safe’. This is based on my overall interpretation of ATSDR’s failure to update their toxicological profile since 1999, and their failure to consider mercury vapour in any of their Interaction (mixtures) profiles. Here is what the ATSDR Interaction Profile for Chlorpyrofos, Lead, Mercury and Methylmercury states:

Some absorption of metallic mercury occurs from dental amalgam fillings, probably following volatilization from the fillings. Clear evidence of adverse effects from this pathway of exposure is lacking, as are joint action studies with the other components of this mixture, so this form of mercury is not considered further in the interaction profile. The critical effect of inorganic mercury is on the kidney, which is not a sensitive target organ for the other components of the mixture.

It is apparent that ATSDR does not recognize any potential for adverse effects from exposure to Hg0, whether from amalgam or any other source. Further, ATSDR appears to suggest that the main target organ is the kidney rather than the central nervous system (CNS), which is incorrect. The NHANES data provide direct and unambiguous (clear) evidence of simultaneous exposure to Hg0 and Pb and MethylHg in 122 million Americans. Therefore, it is evident to me that the omission of Hg0 from the ATSDR’s Interaction Profile for Chlorpyrofos, Lead, Mercury and Methylmercury was based on factors other than data and evidence.
Mark Richardson

{slide=How the U.S. amalgam risk assessment came about}

mark-richardson-01Late 2009, while Mark Richardson was still at Health Canada he was contacted by Michael Adjodha of the Food and Drug Administration (FDA) who was seeking some clarifications about his 1995 Risk Assessment for Health Canada and about his recently (at that time) published article on mercury vapour and setting the reference exposure levels (REL) for mercury. During those phone conversations with Adjodha, Mark answered various questions about his research and publications; the details of which he admits are vague at this point.

In April 2010 Mark Richardson had moved on from Health Canada to SNC-Lavalin. Not long after the move Adjodha contacted Mark again, following up with him about the same issues. Adjodha then pitched the idea of Mark becoming a ‘special employee’ (basically a temporary fulltime staffer) with FDA to assist them with preparations for the planned expert panel and perhaps some follow up. While Mark was interested, it turned out that it would require a green card to work for the US government; a green card Mark did not have, so the idea was eventually abandoned.

Mark decided to put together an unsolicited proposal to complete an up-to-date risk assessment of mercury exposure from dental amalgam for the US population, as it was apparent that this would be relevant and useful (perhaps critical) to an effective evaluation by the Expert Panel. Mark ultimately sent two proposals to Adjodha via email, dated June 29, 2010.

  • Proposal 1 was for part 1 of the report (Hg exposure in US population);
  • Proposal 2 was for the review of concomitant exposure to Hg0, MeHg and Pb (part 2 of the report).

While Adjodha responded favorably to the idea and the need for this work, there were high costs involved ($116K in total), and the fact that Mark was Canadian and not a US citizen. Additionally, Adjodha could not have FDA fund the risk assessment via sole source (no competition). If FDA were to try and fund the risk assessment, Adjodha (FDA) would have to post an open request for proposals and could award it to Mark only if he were the lowest bidder. Since, this RFP process would take months it would miss the deadline for submission to and consideration by the expert panel. Adjohda indicated that if Mark could find an alternate funding agency or organization, and could get the reports to him no later than November 14, 2010, Adjohda would submit them as part of the information package for consideration by the expert panel.

Mark then sent the proposals to the International Academy of Oral Medicine and Toxicology* (IAOMT) who opted to fund them (this is now late August). As a result, Mark and his team got started in early September and got the work done and delivered to the FDA by the November 14th deadline. True to his word, Adjodha included the reports in the package of studies and information to be considered by the expert panel. As part of the contract with IAOMT, Mark flew to Washington DC, in Dec 2010 to briefly present the results to the FDA dental products panel.

* IAOMT was actually the contract manager; funding came from two charitable / not-for-profit foundations. IAOMT merely managed the disbursement of funds once the work was completed.{/slide}

ATSDR mixtures assessment report
Amalgam Risk Assessment Part 1

Mark Richardson discusses the Amalgam Risk Assessment he presented to the FDA

[iframe title=”YouTube video player” width=”480″ height=”340″ src=”″ frameborder=”0″ allowfullscreen]

Mercury Exposure

About The Author

Leave a Reply