Although primates are considered the gold standard laboratory animal model for humans, the relevance for human health of Summers' findings was questioned by the American Dental Association. Fortunately, two Scandanavian groups undertook similar work with human study populations and more recently important additions to this area have been made by a British group. The actual observations in these studies largely replicated those of Summers et al. in laboratory animals but some of the human data were sufficiently noisey that it was not possible to decide whether their results agreed with the laboratory primate observations. The following brief critiques of each article show the areas of agreement and the dangers of overlooking the realities of commensal microbiology in designing and interpreting such human experiments.

A study from this group (Österblad et al., 1995) was the first to address explicitly the influence of amalgam Hg exposure on antibiotic resistance in humans and is significant for that reason. Österblad et al. examined the occurrence of Hg and antibiotic multi-resistance in persons with fillings, persons who had had fillings removed, and persons who had never had fillings. Most of their data were consistent with those of Summers et al. Despite this, the authors concluded that the earlier observations with monkeys did not apply in humans. 

Their conclusion arises from the following flaws in their methods and interpretations of their data.:

Missing amalgam status data and superficial statistical analysis.

Österblad et al. provided no data on the average, standard deviation, and range in the numbers or the types of amalgam fillings in their study cohorts, although such data were at the time well recognized as essential to evaluate Hg exposure from amalgams. They also did not report how many amalgams were removed from their "amalgam removed" subjects and when they had been removed.

Consequently, the standard deviations in their three cohorts are larger than the averages for the fecal Hg concentrations by 2.7-fold and 1.5-fold, respectively, for those with amalgams and for those who have had them removed (p. 2500, first paragraph of the Results section). Such large standard deviations strongly suggest  clustering in the data; i.e. the existence of a sub-group with unusually high fecal Hgconcentrations. Failure to do cluster analysis or even multivariate analysis renders unconvincing Österblad et al.'s conclusion that there is no correlation between fecal Hg concentration and either Hg or antibiotic resistances. It is very surprising that the reviewers missed these points.

Certainly no paper concerned with the effect of antibiotic consumption on the commensal microbiota would have been published if the authors failed to distinguish those who had only consumed one tablet of ampicillin from those who took the full course of the drug. 

The stated conclusions are contradicted by the data.

In the abstract (lines 8-9) and also on p. 2500 of the paper (first sentence of the 2nd paragraph of the Results section), the authors state broadly that there were no correlations between antibiotic resistance and Hg resistance in their subjects. However, in Fig. 1, both ampicillin resistance and nalidixic acid resistance are higher in both amalgam-exposed groups than in those who had never had amalgams at a level which is called "marginally significant" (<0.05). This 95% level is typically regarded as solidly significant in most laboratory and epidemiology work.

Overlooking the significance of antibiotic multiresistance.

In Fig. 2 Houvinen compares the number of resistances per subject in those with Hg resistance and those without Hg resistance and states that there was no difference among the three groups of subjects. However, in every category of multiple resistance (from 1 to 6 additional resistances), the fraction which is also Hg resistant is higher than the fraction which is Hg sensitive by as much as 2-fold or more (e.g. cases with 4-6 additional resistances). In the legend he also notes that 74% of the subjects having Hg resistances have two or more other resistances. So, although Houvinen cannot correlate multiresistances with his different amalgam-exposed groups (not surprising for the reasons noted above), there is a very strong correlation of Hg resistance with antibiotic multiresistance for all of these humans.

Accepting an erroneous assumption.

The largest error in this work is a conceptual one: that you can learn about the cause of a phenomenon from community prevalence studies. This error has plagued the subject of non-hospital antibiotic resistance since its recognition in the mid-1970's. Hospital-based research on nosocomial antibiotic resistance involves rigorous documentation of individual patients' medical histories, continuous, exacting environmental monitoring (including hospital staff), and often continuous draconian measures to control the spread of an antibiotic resistant infectious agent. In other words, such work is based on longitudinal data on individual humans and their environment. Using such protocols, cause and effect relationships can actually be established.

Houvinen's study is a good example of the limitations of prevalence studies for discerning causal relationships in "free-range" humans. If the eight monkeys which Summers et al. examined were given a random number of fillings (both the number and type unknown to the investigator), had been free to roam among the other animals and humans in the vivarium, sharing food, and interacting ad lib, with uncontrolled exposure to antibiotics over a period of years, and if only a single fecal sample had been taken from each monkey at some random point in its life, it is unlikely that any correlation between fecal Hg concentration, Hg resistance, and single or multiple antibiotic resistances could have been seen.

Prevalence studies can tell us about where we are now; they cannot tell us how we got here, i.e. which of many possible factors caused the present state. Longitudinal studies can make such distinctions, for example, Mark Richmond and colleagues established in their landmark 1973 longitudinal studies (Anderson et al., 1973) that even brief antibiotic therapy spreads antibiotic resistances among multiple bacteria in the human gut flora. Thus, well- ontrolled longitudinal studies on individual humans are necessary to resolve whether there is an effect of acute and/or chronic Hg exposure on antibiotic resistance in their normal flora. And in choosing the control population for such human experiments it will be essential to keep in mind the following realities of commensal microbiology.

Exchange of commensal microbiota.

We share our microbes with our family members; Caufield and colleagues had documented even before the Österblad et al. study that infants are colonized with their mothers' strain of Streptococcus mutans (Caufield et al., 1993; Li and Caufield, 1995). More recently Griffin and colleagues demonstrated the transfer of the bacterial agent of periodontal disease from mother to child (Tuite-McDonnell et al., 1997). We also observed several years ago that children without fillings, or even without teeth, have Hg and antibiotic multiresistant fecal bacteria if their parents have amalgam fillings (Wireman and Summers, unpublished observations). Thus, before deciding that a person is a suitable "amalgam-free" control one would need information on the amalgam status of immediate family members; those with amalgambearing kin might well have been colonized by incoming Hg and antibiotic resistant bacteria, perhaps fostering persistance of these acquired microbes each time they took an antibiotic themselves.

In a later paper (Leistevuo et al., 2000) the Finnish group showed that for any given antibiotic the degree of resistance as measured by the minimal inhibitory concentration (MIC) for that antibiotic did not differ for bacteria obtained from persons with amalgams or those without them. Indeed, this is the expected result, since as shown in myriad other studies involving exchangeable genes, a given resistance gene is the same regardless of the bacterial source and the MIC measure only reports on the characteristics of the gene.

Unfortunately, the authors did not use any bona fide susceptible and resistant control strains in their survey, so for Hg at least, they had no working definition of Hg resistance and susceptibility against which to compare their measurements. Moreover, it is not the degree of resistance to any one antibiotic that differs in bacteria derived from high Hg environments; it is the number of different kinds of antibiotic resistances that each bacterial strain carries that increases.

The authors only surveyed four antibiotics and did not report individual strain phenotypes, so we do not have an answer to the critical question: do the amalgam-exposed bacteria have more antibiotic resistance genes or not?

This thorough, well controlled, and well executed longitudinal study (Edlund et al., 1996) was the second to address explicitly the effect of amalgam restorations on antibiotic resistance in the human normal flora. Unlike the earlier study (Österblad et al., 1995) Edlund et al. employed a longitudinal model with 10 subjects with amalgams and 10 subjects without them. Edlund's experimental design involved the removal of pre-existing fillings whereas Summers' work involved the installation of fillings in naive animals.

Edlund et al. also examined specific bacterial populations from the feces; their human study agrees with the work on monkeys for the following points: 

Fecal Hg concentrations People with fillings have high levels of fecal Hg

(the average number of fillings was 19) (Fig. 1 in Edlund paper). The pre-removal fecal Hg levels are in the 0.1 to 1 μg/g range which is close to the "steady state" post-installation Hg observed in the monkeys (which had only 16 fillings) (Fig. 5 in (Summers et al., 1993).). 

Hg dynamics on amalgam removal

Removal of the fillings results in a large pulse of Hg going through the feces. One week after the procedure the fecal Hg had increased approximately 3-4 fold (Fig. 1, Edlund paper). The range of concentrations was very much higher, and even reached the 10 μg/gm level in some persons, just as seen with the monkeys (Fig 5, (Summers et al., 1993).) 

Bacterial dynamics on amalgam removal.

For the group with fillings, the percent of Hg resistant bacteria increased immediately after the fillings were removed and then began to decline (Fig. 2, Edlund paper). This was also seen in the monkeys (Figs. 3,4 (Summers et al., 1993)). 

Increase in antibiotic resistance In the amalgam group (compared to the controls)

Minimal inhibitory concentrations (MIC) for several antibiotics increased in Bacteroides and E.coli. Although Summers' group did not use the MIC method to assess resistance nor did they monitor Bacteroides, they did monitor Gram- egative facultative bacteria (including E. coli) and similarly found an increase in the numbers of antibiotic resistant isolates during the 4 weeks following amalgam installation. 

Increase in multiresistance Antibiotic multiresistance (having 2 or more antibiotic resistances) is very strongly associated with being Hg resistant (Tables 2-4) for all three bacterial genera examined (Bacteroides, E.coli, and Enterococci).

These data exactly corroborate Summers original findings in humans which provoked their animal studies ((Summers et al., 1993), Fig. 1) which revealed the same Hg-multiresistance phenomenon ((Wireman et al., 1997a), Fig. 4).

The Edlund group also made one finding which Summers et al. did not make.

Specifically, Edlund et al. found more Hg resistant Bacteroides in the amalgam subjects than in the controls (Table 1.) Summers et al. did not examine Bacteroides whose cultivation involves techniques not available in their lab. Edlund et al. did not find more Hg resistant E.coli or Enterococcus in the amalgam group compared to the controls, but Summers et al. did find such a difference in the monkeys for these two types of bacteria. It is very interesting that Edlund et al. found increased Hg resistance in Bacteroides, a significant component of the intestinal flora. It is of less concern that Edlund et al. did not find Hg resistance in the minor components Escherichia and Enterococcus. The variability in composition of the gut communities of individual humans and other primates will affect the detectability of the minor components more than in the major ones. Moreover, as noted above, the possible confounding factor of relatives' amalgam fillings might have increased the numbers of Hg resistant bacteria in the controls.

Curiously, despite the extensive concordance with the animal observations, Edlund et al. concluded that amalgams have no effect on the occurrence of Hg and antibiotic resistance in humans because their amalgam-free control group could not be distinguished statistically from the subjects with amalgams.

Actually, the reason they could not distinguish the controls from the amalgam-bearers was because of very high, randomly fluctuating levels of Hg resistant bacteria in the control cohort. It would have been correct to state "no conclusion can be drawn" rather than to state as they did that their data show there is no difference between the two groups.

This latter conclusion simply cannot be drawn from data with as much variability as theirs. In addition, they used only Wilcoxon, rank (non-parametric) statistics in their data analysis. Such a statistical tool specifically minimizes the ability of very high or very low observations to influence the mean and thus obscures the occurrence of "outliers". At the very least, they should have done both parametric and non-parametric analyses; the latter could have unmasked individual differences in the response to Hg. And, as noted above people share their normal flora with their families.

If Edlund's "no amalgam" controls had family members with amalgams, there is some likelihood that they would be colonized with Hg resistant strains. That Edlund et al. did not consider this possible confounder suggests that they were also unaware of the earlier work of Caufield noted above (Caufield et al., 1993; Li and Caufield, 1995).

Beginning in 2002 the group of Peter Mullany has carried out several epidemiological and molecular studies on the relationship of Hg and antibiotic resistances to amalgam exposure on oral bacteria in children. 

Prevalence and longitudinal studies.

Their initial case-control prevalence study found that Hg resistance is common in streptococci of the oral cavity of children regardless of whether they have amalgam restorations, demonstrating again the limitations of prevalence studies on "free range" humans for such studies (Pike et al., 2002).

Interestingly, they observed strongly positive skewing in the data, suggestive of large individual variation in carriage of resistant bacteria among children in both groups. A subsequent assessment of the prevalence of Hg and antibiotic resistance in dental placque bacteria in a larger cohort of children lacking amalgams reached the same conclusion (Ready et al., 2003). Mullany and colleagues have also undertaken one longitudinal study (Pike et al., 2003), monitoring changes in Hg and antibiotic resistances in oral bacteria upon placement of amalgam restorations in children.

They were not able to find significant differences in the pre- and post-installation oral bacteria, possibly owing to the already high prevalence of both characters in the oral microbiota of children in the general population.

Novel Hg resistance genes

Moving to molecular analysis, Mullany's group has sequenced the hallmark merA gene of their oral bacterial isolates and found that they carry either the classic Gram positive merA gene originally defined in Bacillus cereus RC607 or a completely novel version of merA that they have also found in a coagulasenegative Staphylococcus strain (Stapleton et al., 2004) located on a novel transposable element. It is clear that Hg resistance genes travel very widely among the Gram positive bacteria, just as they have been found to do among the Gram negative bacteria. 

Genetic linkage of Hg and antibiotic resistance in Gram positive bacteria.

In their two most recent papers, Mullany's group break some especially exciting ground in this area. Firstly, in work following immediately upon their sequencing of the merA genes of oral bacteria, they have found in a strain of Enterococcus collected from the feces of one of the monkeys used in the Summers' studies that the mer operon lies within a new transposon and is genetically linked to streptomycin resistance as both are carried on a conjugative plasmid (Davis et al., 2005b).

This is the first clear demonstration that mobile plasmids of Gram positive bacteria have also evolved to carry both Hg- and antibiotic resistances.

Genetic linkage of silver and antibiotic resistances in oral bacteria.

In a second paper, Mullany's team makes a very important observation concerning silver resistance. Since silver (Ag) has been used as a topical antiseptic in the prevention of neonatal ophthalmic gonorrhea and a genetic locus conferring resistance to silver sulfadiazine has been identified in plasmids recovered from bacteria infecting burn patients, Mullany's group asked whether Ag resistant bacteria might also be recovered from the mouths of persons with amalgam restorations.

They found Ag- and antibioticmultiresistant Gram negative Enterobacter cloacae could be isolated from infected teeth with dental restorations (Davis et al., 2005a). They confirmed the presence of the previously defined sil operon by PCR amplification of the hallmark silE gene and think that the sil locus is plasmid-carried in these strains.

This finding provides an important clue to the possible origins of the plasmid-encoded Ag resistance (sil) locus which has been the bane of burn centers since the mid-1970's (McHugh et al., 1975). All known examples of the sil locus are in bacteria isolated from severely burned patients, many of whom died from infection with the silver-sulfadiazine resistant bacteria. Surprisingly, when the original Ag resistance locus was sequenced, it was found to be a very large and complex 9-gene operon for constructing two multi-protein membrane-bound efflux pumps, a two component regulatory system, and a periplasmic silver binding protein (Silver, 2003), not something that could evolve in a few weeks of growing on a burn patient being treated with silver sulfadiazine, nor for that matter in response to the single dose of ophthalmic silver nitrate given to newborns. However, amalgam fillings can in principle leach silver by abrasion and galvanic corrosion and they remain in the mouth, a rich bacterial ecosystem, for decades, certainly sufficient time for bacteria to evolve and optimize the silver defense system we have today.

It is noteworthy that the first characterized Ag resistant strain carried both the sil and mer loci along with 5 antibiotic resistant genes genetically linked on the same large, mobile plasmid, pMG101 (McHugh et al., 1975). 

In summary, both prevalence and longitudinal studies in "free range" humans are consistent with and extend the findings made with the more controllable laboratory animal longitudinal experiments. Thus, it is reasonable to conclude that Hg and possibly Ag released from amalgam restorations select for antibiotic multiresistant bacteria as effectively as a course of antibiotic does. However, unlike antibiotics which are typically given for only a few days, amalgams are meant to stay for many years and, as demonstrated elsewhere, leak Hg for their entire lifetimes, continuously selecting for multiresistant bacteria, both benign and pathogenic, in the oral and gastrointestinal microbial communities.

Amalgams and antibiotic resistance.

The spectre of a world where antibiotics no longer work is upon us. Thousands of people die each year in the developed world alone from infections that are no longer treatable by any antibiotic (McGowan, 1991; Witte et al., 2002; Office, 2004).The problem is especially acute in the very young and in the elderly as well as for the immunocompromised. Physicians everywhere are working to minimize the use of antibiotics, especially those of "last resort" and the general public has been educated not to demand antibiotics for  colds and flu's for which they are not effective.

The EU has taken draconian measures to eliminate the use of antibiotics in the food animal industry  (http://www.engormix.com/e_news_view.asp?news=8303&AREA=GDL) and equally severe measures are under consideration by the US Congress (http://www.tufts.edu/med/apua/News/AnimalFeed.html).

These efforts directed exclusively at the use of antibiotics themselves were begun almost two decades ago and have grown steadily in all sectors of antibiotic use.

However, antibiotic resistance continues to spread, so, there must be a missing factor that is overlooked despite our best efforts. A reasonable candidate for this missing factor is amalgam Hg. Amalgam is still the most widely used restorative material in developed countries and the only material available in poorer countries and in the poorest segments of richer nations (Lorscheider et al., 1995).

The EU has recently prohibited use of antibiotics as growth promoters in food animal production because it fosters the increase of antibiotic resistance. Perhaps it is time to recognize that amalgam is fostering the spread of antibiotic resistance directly in humans and should be phased out for that reason as well as for its direct toxic effects.

Anderson, J.D., Gillispie, W.A., and Richmond, M.H. (1973) Chemotherapy and antibiotic resistance transfer between enterobacteria in the human gastro-intestinal tract. J.Med.Microbiol. 6: 461-473.

Axtell, C.D., Myers, G.J., Davidson, P.W., Choi, A.L., Cernichiari, E., Sloane_reeves, J. et al. (1998) Semiparametric modelling of age at achieving developmental milestones after prenatal exposure to methylmercury in the Seychelles Child Development

Study. Environ. Health. Perspect. 106: 559-563. Barkay, T. (1987) Adaptation of aquatic microbial communities to Hg2+ stress. Appl.Environ.Microbiol. 53: 2725-2732.

Barkay, T., and Olson, B.H. (1986) Phenotypic and genotypic adaptation of aerobic heterotrophic sediment bacterial communities to mercury stress. Appl.Environ.Microbiol. 52: 403-406.

Barkay, T., Miller, S.M., and Summers, A.O. (2003) Bacterial mercury resistance from atoms to ecosystems. FEMS Microbiol Rev 27: 355-384. 

Berntssen, M.H., Hylland, K., Lundebye, A.K., and Julshamn, K. (2004) Higher faecal excretion and lower tissue accumulation of mercury in Wistar rats from contaminated fish than from methylmercury chloride added to fish. Food Chem Toxicol 42: 1359-1366.

Caufield, P.W., Cutter, G.R., and Dasanayake, A.P. (1993) Initial acquisition of mutans streptococci by infants: evidence for a discrete windowof infectivity. J. Dent. Res. 72: 37-45.

Clarkson, T.W. (1997) The toxicology of mercury. Crit Rev Clin Lab Sci 34: 369-403. 

Compeau, G.C., and Bartha, R. (1985) Sulfate-reducing bacteria: principle methylators of mercury in anoxic estuarine sediment. Applied and Environmental Microbiology 50: 498-502.

Conway, T., Krogfelt, K.A., and Cohen, P.S. (2004). The life of commensal Escherichia coli in the mammalian intestine [web book] 

Davis, I.J., Richards, H., and Mullany, P. (2005a) Isolation of silver- and antibioticresistant Enterobacter cloacae from teeth. Oral Microbiol Immunol 20: 191-194. 

Davis, I.J., Roberts, A.P., Ready, D., Richards, H., Wilson, M., and Mullany, P. (2005b) Linkage of a novel mercury resistance operon with streptomycin resistance on a conjugative plasmid in Enterococcus faecium. Plasmid 54: 26- 8.

Edlund, C., Bjorkman, L., Ekstrand, J., Sandborgh-Englund, G., and Nord, C.E. (1996) Resistance of the normal human microflora to mercury and antimicrobials after exposure to mercury from dental amalgam fillings. Clinical Infectious Diseases 22: 944-950.

Edwards, T., and McBride, B.C. (1975) Biosynthesis and degradation of methylmercury in human faeces. Nature 253: 463-464.

Gibson, G.R., Macfarlane, S., and Macfarlane, G.T. (1993a) Metabolic interactions involving sulphate-reducing and methanogenic bacteria in the human large intestine. FEMS Microbiol. Ecol. 12: 117-125.

Gibson, G.R., MacFarlane, G.T., and Cummings, J.H. (1993b) Sulphate reducing bacteria and hydrogen metabolism in the human large intestine. Gut 34: 437-439.

Grandjean, P., Budtz-Jorgensen, E., Keiding, N., and Weihe, P. (2004) Underestimation of risk due to exposure misclassification. Int J Occup Med Environ Health 17: 131-136.

Grandjean, P., Weihe, P., White, R.F., Debes, F., Araki, S., Yokoyama, K. et al. (1997) Cognitive deficit in 7-year old children with prenatal exposure to methylmercury. Neurotoxicol. Teratol. 19: 417-428.

Harris, H.H., Pickering, I.J., and George, G.N. (2003) The chemical form of mercury in fish. Science 301: 1203.

Heintze, U., Edwardsson, S., Derand, T., and Birkhed, D. (1983) Methylation of mercury from dental amalgram and mercuric chloride by oral streptococci in vitro. Scandinavian Journal of Dental Research 91: 150-152.

Hooper, L.V., and Gordon, J.L. (2001) Commensal host-bacterial relationships in the gut. Science 292: 1115-1118.

Hursh, J., Sichak, S., and Clarkson, T.W. (1988) In vitro oxidation of mercury by the blood. Pharmacol. and Toxicol. 63: 266-273.

Kanauchi, O., Matsumoto, Y., Matsumura, M., Fukuoka, M., and Bamba, T. (2005) The beneficial effects of microflora, especially obligate anaerobes, and their products on the colonic environment in inflammatory bowel disease. Curr Pharm Des 11: 1047-1053.

King, J.K., Saunders, F.M., Lee, R.F., and Jahnke, R.A. (1999) Coupling mercury methylation rates to sulfate reduction rates in marine sediments. Environ. Toxicol. Chem. 18: 1362-1369.

King, J.K., Kostka, J.E., Frischer, M.E., and Saunders, F.M. (2000) Sulfate-reducing bacteria methylate mercury at variable rates in pure culture and in marine sediments. Appl. Environ. Microbiol. 66: 2430-2437.

Leistevuo, J., Jarvinen, H., Österblad, M., Leistevuo, T., Huovinen, P., and Tenovuo, J. (2000) Resistance to mercury and antimicrobial agents in Streptococcus mutans isolates from human subjects in relation to exposure to dental amalgam fillings. Antimicrob Agents Chemother 44: 456-457.

Leistevuo, J., Leistevuo, T., Helenius, H., Pyy, L., Österblad, M., Huovinen, P., and Tenovuo, J. (2001) Dental amalgam fillings and the amount of organic mercury in human saliva. 

Li, Y., and Caufield, P.W. (1995) The fidelity of initial acquisition of mutans streptococci by infants from their mothers. J. Dent. Res. 74: 681-685.

Liang, L., and Brooks, R.J. (1995) Mercury reactions in the mouth with dental amalgams. Water, Air, and Soil Pollution 80: 103-107.

Liebert, C.A., Wireman, J., Smith, T., and Summers, A.O. (1997) Phylogeny of mercury resistance (mer) operons of gram-negative bacteria isolated from the fecal flora of primates. Applied and Environmental Microbiology 63: 1066-1076.

Lorscheider, F.L., Vimy, M.J., and Summers, A.O. (1995) Mercury exposure from "silver" tooth fillings: Emerging evidence questions a traditional dental paradigm. FASEB J. 9: 504-508.

Lundberg, J.O., Weitzberg, E., Cole, J.A., and Benjamin, N. (2004) Opinion: Nitrate, bacteria and human health. Nat Rev Microbiol 2: 593-602. 

Macdonald, T.T., and Monteleone, G. (2005) Immunity, inflammation, and allergy in the gut. Science 307: 1920-1925.

Magos, L., and Clarkson, T.W. (1978) Role of catalase in the oxidation of mercury vapor. Biochem.Pharmacol. 27: 1373-1377.

McGowan, J.E., Jr. (1991) Antibiotic resistance in hospital bacteria: current patterns, modes for appearance or spread, and economic impact. Reviews in Medical Microbiology 2: 161-169.

McHugh, G.L., Moellering, R.C., Hopkins, C.C., and Swartz, M.N. (1975) Salmonella typhimurium resistant to silver nitrate, chloramphenicol, and ampicillin. Lancet 1: 235-240.

Merk, K., Borelli, C., and Korting, H.C. (2005) Lactobacilli - bacteria-host interactions with special regard to the urogenital tract. Int J Med Microbiol 295: 9-18. 

Myers, G.J., and Davidson, P.W. (1998) Prenatal methylmercury exposure and children: neurologic, developmental, and behavioral research. Environ. Health Perspect. 106:841-847.

Office, U.S.G.A. (2004) Antibiotic Resistance: Federal agencies need to better focus efforts to address risk to humans from antibiotic use in animals. In. Washington, D.C., p. 95.

Olson, B.H., Barkay, T., and Colwell, R.R. (1979) Role of plasmids in mercury transformation by bacteria isolated from the aquatic environment. Appl.Environ.Microbiol. 38: 478-485.

Pike, R., Lucas, V., Stapleton, P., Gilthorpe, M.S., Roberts, G., Rowbury, R. et al. (2002) Prevalence and antibiotic resistance profile of mercury-resistant oral bacteria from children with and without mercury amalgam fillings. J Antimicrob Chemother 49:777-783. 

Pike, R., Lucas, V., Petrie, A., Roberts, G., Stapleton, P., Rowbury, R. et al. (2003) Effect of restoration of children's teeth with mercury amalgam on the prevalence of mercury- and antibiotic-resistant oral bacteria. Microb Drug Resist 9: 93-97.

Porter, F.D., Ong, C., Silver, S., and Nakahara, H. (1982) Selection for mercurial resistance in hospital settings. Antimicrob.Agents Chemother. 22: 852-858.

Ready, D., Qureshi, F., Bedi, R., Mullany, P., and Wilson, M. (2003) Oral bacteria resistant to mercury and to antibiotics are present in children with no previous exposure to amalgam restorative materials. FEMS Microbiol Lett 223: 107-111.

Risher, J.F., Murray, H.E., and Prince, G.R. (2002) Organic mercury compounds: human exposure and its relevance to public health. Toxicol Ind Health 18: 109-160.

Rowland, I.R., Grasso, P., and Davies, M.J. (1975) The methylation of mercuric chloride by human intestinal bacteria. Experientia 31: 1064-1065.

Rudd, J.W., Furutani, A., and Turner, M.A. (1980) Mercury methylation by fish intestinal

contents. Appl Environ Microbiol 40: 777-782.

Silver, S. (2003) Bacterial silver resistance: molecular biology and uses and misuses of silver compounds. FEMS Microbiol Rev 27: 341-353.

Smits, H.H., Van Beelen, A.J., Hessle, C., Westland, R., De Jong, E., Soeteman, E. et al. (2004) Commensal Gram-negative bacteria prime human dendritic cells for enhanced IL-23 and IL-27 expression and enhanced Th1 development. Eur J Immunol 34: 1371-1380.

Stapleton, P., Pike, R., Mullany, P., Lucas, V., Roberts, G., Rowbury, R. et al. (2004) Mercuric resistance genes in gram-positive oral bacteria. FEMS Microbiol Lett 236: 213-220.

Summers, A.O., Wireman, J., Vimy, M.J., Lorscheider, F.L., Marshall, B., Levy, S.B. et al. (1993) Mercury released from dental "silver" fillings provokes an increase in mercury- and antibiotic-resistant bacteria in oral and intestinal floras of primates. Antimicrob Agents Chemother 37: 825-834.

Tannock, G.W. (1995) Normal Microflora: An introduction to the microbes inhabiting the human body. London: Chapman & Hall.

Tuite-McDonnell, M., Griffen, A.L., Moeschberger, M.L., Dalton, R.E., Fuerst, P.A., and Leys, E.J. (1997) Concordance of Porphyromonas gingivalis colonization in families. J. Clinical Microbiology 35: 455-461.

Vonk, J.W., and Sijpesteijn, A.K. (1973) Studies on the methylation of mercuric chloride by pure cultures of bacteria and fungi. Antonie van Leeuwenhoek J.Microbiol.Serol. 39: 505-513.

Wireman, J., Liebert, C.A., Smith, T., and Summers, A.O. (1997a) Association of mercury resistance with antibiotic resistance in the gram-negative fecal bacteria of primates. Appl Environ Microbiol 63: 4494-4503.

Wireman, J., Liebert, C.A., Smith, C.T., and Summers, A.O. (1997b) Association of mercury resistance and antibiotic resistance in the Gram negative fecal bacteria of primates. Appl. Environ. Microbiol. 63: 4494-4503.

Witte, W., Klare, I., and Werner, G. (2002) Molecular ecological studies on spread of antibiotic resistance genes. Anim Biotechnol 13: 57-70.

Yan, F., and Polk, D.B. (2004) Commensal bacteria in the gut: learning who our friends are. Curr Opin Gastroenterol 20: 565-571.

Österblad, M., Leistevuo, J., Leistevuo, T., Jarvinen, H., Pyy, L., Tenovuo, J., and Huovinen, P. (1995) Antimicrobial and mercury resistance in aerobic gramnegative bacilli in fecal flora among persons with and without dental amalgam fillings. Antimicrob Agents Chemother 39: 2499-2502.