4.1. Influence of dental amalgam on Hg burden

This study is the first in Saudi Arabia to evaluate the levels of Hg in children measured in various matrices and to assess Hg's relation to dental amalgam fillings. Our results have clearly demonstrated the association between dental amalgam fillings and the levels of Hg in the urine and hair of children. These observations support previous studies (Levy et al., 2004; Woods et al., 2007). In general, our overall means of urinary Hg measured as UHg (3.749 μg/L) and UHg-C (3.763 μg/g creatinine) in children with dental amalgam fillings were much higher than those reported in other studies, as shown in Table 7. Although, the World Health Organization established an upper limit of normal UHg for unexposed adults of 20 μg/L, respective levels for  children have not yet been established (WHO, 2003). Background levels of UHg and UHg-C in an unexposed population are generally expected to be b10 μg/L (ATSDR, 1999) and 5 μg/g creatinine respectively (IPCS, 2003). In this study, the levels of UHg in children without amalgam were in the range of general background levels (b10 μg/L) in unexposed populations. UHg-C levels were over 5 μg/g creatinine, with an average of 9.163 μg/g creatinine (range: 5.355– 19.242 μg/g), in 19.7% of children without amalgam. In children with dental amalgam, 2.8% had UHgN10 μg/L, and 22.6% had UHg-CN5 μg/g creatinine. Our results should be interpreted with caution because WHO and ATSDR estimates are only applied to healthy adults and not to children. In general, very limited information is available on background Hg exposure in children. Many published data indicate that the vast majority of unexposed children should have urinary Hg levels b5 μg/L (Ozuah et al., 2003; Bose-O'Reilly et al., 2010).

However, comparing our results to the recently defined reference value for UHg by the German Commission of Human Biomonitoring for UHg in children (3–14 year olds), 90.8% of children's urinary Hg concentrations without amalgam were above the reference value of 0.4 μg/L (Schulz et al., 2009). In contrast, the authors reported that UHg for children with more than two dental amalgams was 3.1 μg/L at  the 95th percentile. In this study, children with dental amalgam had a considerably higher 95th percentile of 8.538 μg/L, and more than 47% had UHg levels over 3.1 μg/L, with a range of 3.129 to 15.575 μg/L. 

Surprisingly, 51.3% of children without amalgam also had UHg levels over 3.1 μg/L, with a range of 3.113 to 7.946 μg/L and a 95th percentile of 5.73 μg/L. Few authors have proposed background levels of UHg-C for children. Batáriová et al. (2006) considered 0.37 μg/g creatinine as a background level of UHg-C which corresponds to a level of UHg of 0.42 μg/L. Using this reference value from Batáriová et al. (2006), 99.1% and 88.2% of children in our study with and without dental amalgams, respectively, had UHg-C levels above 0.37 μg/g creatinine. Although, the results presented here generally support the contribution of dental amalgam filling to urinary Hg levels, we also notably, found high urinary Hg among children that did not have dental amalgam. Our findings likely suggest the possible contribution of other sources. 

Hair has been widely used as a bioindicator to evaluate human exposure because of its growth rate (1 cm per month) and the tendency of Hg to accumulate in the hair follicles over long-periods of time (WHO, 1990). Many studies have used hair to assess Hg exposure from the consumption of fish (Srogi, 2007). To our knowledge, few studies have examined the relation of HHg to dental amalgam. Interestingly, our study revealed that HHg was also highly associated with dental amalgam fillings, and children with dental amalgam had HHgN2.5 times higher than those in children without. The finding was in accordance with the levels reported in adults (Babi et al., 2000; Lindow et al., 2003; Fakour et al., 2010a,b; Ryo et al., 2010) but not in children (Batista et al., 1996; Pesch et al., 2002;Dunn et al., 2008; Kruzikova et al., 2009). In children, 62.6%with dentalamalgamate fish and 37.4%did not. Multiple logistic regression modeling confirmed that the presence of amalgam fillings leads to an increased risk of high HHg in children, even after adjusting for seafood consumption (OR=5.396). As shown in Table 7, the levels of HHg found in children with amalgam fillings (0.614 μg/g dry wt.) were higher than those reported by Dunn et al. (2008) and Barregard et al. (2008) of 0.3 and 0.4 μg/g dry wt., respectively. Comparable studies are very limited, and further research is required to confirm our observations. Hair is generally the preferred matrix to determine exposure to methylmercury, but hair may also be used an alternative to urine as an indicator of long-term Hg exposure from dental amalgam fillings. Hair is a non-invasive matrix especially useful in studies involving children.

The mean HHg level for all children in this study was 0.453± 1.155 μg/g dry wt., which was notably higher than levels found in other studies: 0.18 μg/g in the Czech Republic (Puklová et al., 2010); 0.23 μg/g in Germany (Pesch et al., 2002), and 0.31 μg/g in the USA (Surkan et al., 2009). Our levels, however, were lower than some others: 0.74 μg/g Seoul, Korea (Kim et al., 2008); 0.94 μg/g in Barcelona, Spain (Diez et al., 2009); 0.938 μg/g in Italy (Montuori et al., 2006); 2.2 μg/g South China (Ip et al., 2004); and 0.96 μg/g in Granada, Spain (Freire et al., 2010). Both the US Environmental Protection Agency and the National Academy of Sciences recommend to keep hair Hg levels at b1.0 μg/g, corresponding to a reference dose (RfD) of 0.1 μg/kg body weight per day (US EPA, 1997). This reference dose was not based on a direct adverse effect of HHg but was derived from maternal HHg levels that correspond to the Hg levels in cord blood, which over a lifetime of exposure should not result in adverse effects. Recent studies by both Surkan et al. (2009) and (Freire et al., 2010) reported that HHg levels"1 μg/g (EPA reference dose) in children from consumption of fish were associated with delays in neuropsychological development. In our study, hair Hg levels  exceeded the US EPA reference dose of 1 μg/g in 10 children (5.8%), with a range of 1.003 to 14.304 μg/g dry wt. Of  these, only one child had a level of HHg higher than the tolerance limit of 10 μg/g set by the WHO (1990). Nine children  had dental amalgam fillings, and eight of these consumed fish. These children were apparently concurrently exposed to various forms of Hg. We know that inorganic Hg does not incorporate into hair in a fashion similar to organic Hg (Dunn et al., 2008; Diez et al., 2008; Kruzikova et al., 2009). A study by Magos and Clarkson (2008) revealed that hair accounts for only a small fraction, less than 10%, of the total elimination of organic Hg from the body if this form of Hg is the main source of exposure. Other species of Hg are eliminated by hair but at a lower rate. Further studies on the chemical forms of Hg (organic and inorganic) in hair samples are needed in order to identify the sources of Hg in the hair of Saudi children. 

The levels of Hg in toenails in this study are generally very small for both groups of children,with amalgam(0.163±0.339 μg/g drywt.) and without amalgam (0.222±0.212 μg/g dry wt.). Interestingly, NHg was inversely associated with dental amalgam fillings, which might reflect another chemical form of Hg. One of the possible explanations for this finding seems related to the association between recent vaccination (b1 year) and dental amalgam fillings. As shown in Table 3, 79.3% of the children with amalgam were vaccinated. After excluding children without amalgam fillings,we stratified NHg levels by the use of vaccine. Non-vaccinated children tend to have significantly higher NHg (0.27± 0.534 μg/g dry wt., N=37) than vaccinated children (0.101± 0.095 μg/g dry wt., N=64), with a P-value of 0. In contrast, when we excluded children with amalgam, we found the reverse.

Non-vaccinated children had lower NHg (0.212±0.230 μg/g dry wt., N=45) than vaccinated children (0.252±0.156 μg/g dry wt., N=16) with a moderate level of significance (P=0.065). Although we cannot rule out a chance finding, the excretion pattern of Hg appears to differ in vaccinated and nonvaccinated children with amalgam. Thimerosal, a preservative in vaccines, contains 49.6% Hg by weight and is metabolized to ethylmercury and thiosalicylateethyl (American Academy of Pediatrics, 1999). Haley (2005) reported a synergistic mechanism between thimerosal and other vaccine components that affects Hg toxicity. This might let us think that if some synergistic effect between amalgam fillings and thimerosal may be responsible for reduced Hg excretion in these children.

The overall Hg mean level in toenails found in this study was 0.185±0.298 μg/g dry wt. The data from this study is difficult to compare to data from others, only carried out in adults (Ohno et al., 2007; Rees et al., 2007; Joshi et al., 2003). A threshold of 0.33 μg/g for Hg in toenails has been suggested to be equivalent to the EPA reference dose of 1  μg/g of Hg in hair (US EPA, 1997; Choi et al., 2009). 

Of all tested children, 11.7% (N=19) had levels of NHgN0.33 μg/g, with a range of 0.332–3.337 μg/g. Of these seven had dental amalgam fillings. Guallar et al. (2002) found an increase in the risk of myocardial infarction among men with toenail HgN0.66 μg/g. In our study, four children (2.5%) had toenail Hg greater than 0.66 μg/g, in a range of 0.778 to 3.337 μg/g. Of these four, only one child had four dental amalgam fillings.

4.2. Relationship between hair and nail Hg

Like hair, nails have also been recommended as a biomarker of long-term Hg exposure. Rodushkin and Axelsson (2000), Ohno et al. (2007), and Zolfaghari et al. (2007) reported correlation coefficients between HHg and NHg of 0.9, 0.858, and 0.57, respectively. Hair and nail matrices are assumed to have unique properties of keratin compositions (Kitahara and Ogawa, 1991).

Our study has shown that the overall level of HHg (0.453 μg/g dry wt.) was more than two times higher than Hg levels in toenails (0.185 μg/g dry wt.), with a weak but significant correlation (Pearson's r=0.25). These values suggest that hair and nails may share similar sources and patterns of exposure, but a previous study revealed that the higher HHg concentration might be attributable to the differences in chemical composition, particularly the sulfur content, and in blood flow during growth (Suzuki et al., 1989). Rodushkin and Axelsson (2000) suggested that differences in mode of deposition (endogenous and exogenous) could account for higher levels of Hg in hair. Ohno et al. (2007) estimated the strength of the correlation between HHg and NHg using a reciprocal-x model of regression analysis. Applying their approach, we found that the 95% CIs of the regression line of HHg (Y-axis) on NHg (X-axis) ranged from 0.583 to 1.728 for the slope and from 0.397 to 0.806 for the intercept. 

The ratio between HHg to NHg had a mean±SD of 5.047±10.427, ranging from 0.2 to 92.37. Our ratio is higher than the ratios of 2.38 and 2.56 reported by Ritchie et al. (2002) and Ohno et al. (2007), respectively, but these differences might be due to differences between children and adults. Budtz-Jørgensen et al. (2004) suggested that strands of hair in children tend to retain more Hg than adults. Also, our ratio does not fall within the 95% CI of the slope. The exclusion of two outliers did not change the results. The two matrices, therefore, might not fully share sources and/or times of exposure, and the forms of Hg in hair and nails are likely different. When we reexamined our multiple logistic regression model by testing the influence of dental amalgam fillings on hair-to-nail Hg ratio (HHg/NHg) instead of each separately, after adjusting for confounders as described in Table 4, the OR of HHg/NHg for children with dental amalgam was 2.988 (95% CI: 1.824–4.896) significantly higher than for those without (P=0). UHg-C was also predictive of dental amalgam fillings (OR: 2.034, 95% CI: 1.306–3.165). Nails, like, hair, reflect past exposure but have the potential to integrate multiple routes of exposure over long periods of time because of the slow rate of growth of toenails (Yaemsiri et al., 2010). The reliability of nails to determine associations with dental amalgam fillings may thus be limited, especially if the levels of Hg in nails are low. The hair-to-nail Hg ratio (HHg/NHg), though, was influenced by the presence of dental amalgam fillings, which might suggest that the proportion of inorganic Hg in hair was higher than organic Hg. Organic Hg mainly incorporates in hair but usually transforms into inorganic Hg as it enters the hair follicle and/or as the hair grows (Clarkson, 1997; Dolbec et al., 2001). Without knowledge of the species present, this mechanism cannot be ruled out.

4.3. Risk factors associated with Hg exposure in dental amalgam group

Using multiple regression analyses, the main risk factors associated with urinary Hg body burden in children with dental amalgam fillings were gender, aphthous ulcer, and brushing once a day (Table 6a,b). Girls with dental amalgam had higher urinary Hg levels (both UHg and UHg-C) in comparison to boys, in agreement with Woods et al. (2007). A recent experimental study revealed that differences in whole-body  retention and accumulation of Hg in organs are regulated by genetic factors and gender (Ekstrand et al., 2010) with higher Hg renal retention in males than in females. We also noticed that an increase in HHg was associated with the feeling of a burning mouth-sensation. An immediate  hypersensitivity reaction associated with the Hg in amalgam restorations resulted in erythematous lesions, severe burning, and itching, and difficulty in breathing (Kal et al., 2008).

Children living in the western part of Taif had significantly lower levels of HHg than those living in the northern area. Interestingly, the levels of Hg in the hair of subjects with dental amalgam were not influenced by seafood consumption. This might be due to the small size of children who do not eat seafoods (N=8) versus fish eaters (N=97). In contrast, HHg in children without dental amalgam who ate seafood  (0.275±0.191, N=56) was significantly higher than in those without amalgam (0.141±0.084, N=18), with P=0.001. Our study revealed that children attending B and C dental clinics had higher NHg levels than those attending A dental clinics. This finding indicates the existence of potential Hg source(s) in these particular dental clinics. The model also reinforced our earlier observation (Table 4) that vaccinated children had lower NHg levels than nonvaccinated children, which should be examined further. Our study also revealed some association between white patches and NHg levels. Further research on this issue is needed, especially for children with hypersensitivity to Hg amalgam (Kazantzis, 2002). These findings are of interest and indicate a need for further research with much larger sample sizes for understanding the possible influences of these factors on Hg levels. Lastly and in contrast to many studies, our results found no association between Hg levels in any matrix and the number of dental amalgam fillings (Pesch et al., 2002; Woods et al., 2007; Dunn et al., 2008). Khordi-Mood et al. (2001) also found that the increase in UHg in children 9–12 days after receiving dental amalgam fillings did not correlate with the number of fillings. Mortda et al. (2002) reported similar results for HHg and NHg. According to Maserejian et al. (2008), determining the number of dental amalgam fillings is insufficient if only a rough estimate of current Hg exposure due to amalgam is needed. However, more detailed information on surfaces can provide a better estimate of the association, and simple counts of the number of dental fillings are unlikely to capture the full influence of amalgams in children over time. Finally, some limitations of this study need consideration for the interpretation of our findings:

(1) subjects were limited to the children of military staff, which might affect the ability to generalize to other sectors;

(2) no information on socioeconomic status was collected;

(3) no data on the size of amalgam restoration and surface area were collected; and

(4) some children (17%) had had their amalgam fillings for more than 12 months.