Introduction

The significance of mercury and its compounds with regard to environmental health is well known and has been summarized by WHO (1990, 1991). Within the scope of a study about traffic-related air pollution, it was possible to compare mercury levels in hair, urine, and saliva in a group of 245 school children (8–10 years old) from the area. This study was undertaken in Düsseldorf, the capital of North-Rhine-Westphalia, Germany, which is located in the highly industrialized Rhine-Ruhr region. For biological monitoring of mercury exposure, samples of urine, hair, and saliva were collected. Urine is the most widely used and accepted matrix to assess internal mercury exposure. Concentration levels can be interpreted individually since reference values and toxicologically based values have been established by the German Commission on Human Biological Monitoring (Ewers et al., 1999). Hair samples are also of interest in environmental health surveys, especially in studies conducted on children (Wilhelm and Idel, 1996). Saliva as a readily collectable material has also been discussed for use as a diagnostic fluid (Silbergeld, 1993). However, the analysis of saliva has not yet been accepted as a valuable tool in environmental medicine.

Subjects and methods

Study Population

A total of 245 children (121 boys and 124 girls; 8–10 years of age) from a city located in a highly industrialized region in North-Rhine-Westphalia, Germany, composed the test group. Sampling took place in February and March 1996. Only those children who had lived for at least 2 years at their address were included. All parents consented to the study and completed a questionnaire. In the questionnaire, information on the following variables was collected: age, gender, place of residence, nutrition habits (especially frequency of fish consumption), parents' smoking habits, amount of time spent outdoors, hair color, time of last hair wash, and chemical hair treatment. Furthermore, a basic medical examination took place. The dental status (especially number, condition, and age of amalgam fillings, and of amalgam filling surfaces) was examined by a dentist (N.Schmitz). The determination of the number of surfaces of an amalgam filling served as an approximate indicator of its size.

Sample Collection and Treatment

Urine was collected over a period of 24 h in 2.5-l polypropylene sampling vessels and stored at -20°C prior to analysis.

Hair samples (the first 1.5–2 cm proximal to the scalp, weighing 50–200 mg) were obtained from symmetrical occipital regions and stored in sealed polyethylene bags at room temperature.

Saliva was collected without previous salivary stimulation into 15-ml polypropylene centrifuge tubes and centrifuged at 1000 rpm for 5 min. An amount of 1.5 ml of the liquid phase was transferred into a 1.5-ml polypropylene reaction vessel and stored at -70°C.

Mercury Measurements

Cold vapor atomic absorption spectrometry (AAS), using sodium borohydride as the reducing agent, was employed, combined with an amalgamation technique as analytical method to determine all forms of mercury.

The determination of mercury in urine was carried out using a flow injection system (FIAS 400, Autosampler AS-90, Atomic Absorption Spectrometer AAS 3110; Bodenseewerk Perkin Elmer, Überlingen, Germany) after a sampling treatment according to Vondenhof and Beindorf (1976). For quality assurance purposes, mercury in urine was determined daily using two lyophilized standard reference urines (Lanonorm metal 1, lot no. 625210, Behring Institute, Marburg, Germany; batch no. 009024, Nycomed Pharma, Oslo, Norway). Additionally, the laboratory participated in an external interlaboratory quality assurance program, supervised by the German Society of Occupational Health and Environmental Medicine. The limit of quantification for mercury in urine was 0.1 g/l.

Mercury determination in hair and saliva was performed using a batch design according to Schierling and Schaller (1981). A flow injection system (FIAS 100, Atomic Absorption Spectrometer 1100 B; Bodenseewerk Perkin Elmer) was used as computer-controlled peristaltic pump. The reacting solution consisted of 1 ml of sample solution (saliva or hair digestion solution), 1 ml of nitric acid 65%, 0.5 ml of antifoam solution 2%, and 8 ml of aqua bidest. Hair samples were mineralized by a microwave digestion procedure prior to analysis. In this procedure, a 50-mg portion of the proximal end of the hair was weighed into a PTFE digestion vessel mixed with 1 ml of nitric acid 65% and 0.5 ml of hydrogen peroxide 30% and digested in a closed microwave system at microwave power rates of 200–550 W. The digestion solution was filled up to 4.0 ml with aqua bidest. A human hair standard reference material (GBW 09101 Human Hair) was analyzed for quality assurance purposes. Saliva was used without further treatment. To detect mercury in saliva, no reference material was available. Limits of quantification were 0.06 g/g for mercury in hair and 0.32 g/l for mercury in saliva.

Determination of Creatinine in Urine

Creatinine in urine was measured using a test kit (Merckotest no. 3385; Merck, Darmstadt, Germany), based on the Jaffé reaction.

Statistical Evaluation

For statistical evaluation, the program SAS PC, version 6.09, was applied. The descriptive statistical parameters [arithmetic mean, standard deviation, minimum, maximum, median, percentiles, geometric mean (GM)] and the respective confidence limits were calculated for the mercury content of hair, urine, and saliva.

Using the analytical results and the information recorded via questionnaires, confounders for mercury contents in hair and urine were calculated by stepwise multiple linear regression. The multiple square coefficient of regression R² indicates the fraction of the dependent variables explained by the independent variables. For inclusion and exclusion of variables, a level of significance of 15% was chosen.

As more than 70% of the saliva mercury concentrations were below limit of quantification, regression analysis was not applied to saliva.

Due to the log-normal distribution of mercury concentrations in hair and the close-to-log-normal distribution in case of urine, correlations were calculated using logarithmic values.

Results

Seventy-four of 245 children under study had amalgam fillings (30.2%). Most of them (n=43) had one to three amalgam filling surfaces. One child had 14 amalgam filling surfaces. Twenty-one fillings were defective (related to 14 children). One child received a new amalgam filling during the last 4 weeks prior to the study period, whereas one child had one amalgam filling removed within the last 4 weeks. In no case were dental restorations made out of different metals alloys.

The principal descriptive statistical data for the mercury contents in urine, hair, and saliva are summarized in Table 1.

Table 1 - Mercury levels in urine, hair, and saliva of 8 - to 10- year - old children living in a city located in an industrialized area of Germany.^a.

Full table (48K)

Mercury in Urine

The frequency distribution of levels of mercury in urine was close to a log-normal distribution pattern (Figure 1). There was a good correlation between mercury levels in urine related to volume and 24 h (Figure 2; r=0.866). An even stronger correlation was found between the mercury concentration in urine related to volume and to creatinine excretion (Figure 3; r=0.933). Mercury levels increased with the number of amalgam fillings (Figure 4) and the number of amalgam filling surfaces (Figure 5). There was no significant difference between mercury levels in urine related to the number of amalgam fillings (r=0.50) and the number of amalgam filling surfaces (r=0.49). Data on mercury levels in urine of children with and without amalgam fillings are displayed in Table 2, showing higher mercury levels in individuals with amalgam fillings.

Figure 1.

Log-normal distribution of mercury concentrations in urine of German children aged 8–10 years (n=274).

Full figure and legend (63K)

Figure 2.

Scatter plot of mercury in urine related to volume and to 24-h urine with linear regression line.

Full figure and legend (14K)

Figure 3.

Correlation between mercury in urine by volume and mercury in urine related to urine creatinine in children.

Full figure and legend (14K)

Figure 4.

Correlation between mercury concentration in urine and number of amalgam fillings in children.

Full figure and legend (16K)

Figure 5.

Correlation between mercury concentration in urine and number of amalgam filling surfaces in children.

Full figure and legend (17K)

Table 2 - Mercury concentrations in urine (g/l) in children with and without amalgam fillings.

Full table (22K)

Using multiple linear regression analysis, the number of teeth with amalgam filling surfaces accounted for 23.2% of the variance (P<0.001) while the number of defective amalgam fillings accounted for 2.1% (P<0.016), the overall explanation of variance thus being 25.3%.

Mercury in Hair

Mercury concentrations in hair were distributed log-normally (Figure 6). Using multiple linear regression analysis, three predictors accounted for 20.4% of the variance of mercury levels in hair. The frequency of fish consumption was the most significant predictor, accounting for 14.9% (P<0.001). Mercury levels in hair increased with the number of fish meals consumed per month (Figure 7). The smoking habits of the parents (3.1% of the variance, P<0.004) and the child's age (2.4% of the variance, P<0.011) also revealed to be significant predictors. The mercury levels in hair were higher in children whose parents were smokers when compared with mercury levels in hair of children with nonsmoking parents [median, 95th percentile, geometrical mean (in g/g): 0.21, 0.58, 0.20 vs. 0.15, 0.48, 0.14]. The age of the children correlated negatively with the hair mercury levels.

Figure 6.

Log-normal distribution pattern of mercury levels in hair of German children aged 8–10 years.

Full figure and legend (59K)

Figure 7.

Correlation between mercury concentration in hair and frequency of fish consumption in children.

Full figure and legend (14K)

There was a weak correlation (r=0.297; P<0.05) between mercury concentrations found in hair and urine (Figure 8).

Figure 8.

Scatter plot of the mercury concentration in hair and urine of German children aged 8–10 years.

Full figure and legend (14K)

Mercury in Saliva

Mercury levels in saliva ranged from <0.32 (limit of quantification) to 4.5 g/l. The median was 0.16 g/l. As the mercury concentration in saliva was below limit of quantification in 71% of the samples, it was not meaningful to undertake a multiple linear regression analysis.

Discussion

To the authors' knowledge concerning children, no study was recently done focusing on the combined determination of mercury in urine, hair, and saliva.

Other studies dealt mainly with the influence of amalgam fillings or occupational mercury exposure on mercury levels in adults (e.g., Olstad et al., 1987; Herber et al., 1988; Begerow et al., 1994; Wilhelm et al., 1996; Ganss et al., 2000).

The mercury contents of urine found in the present study point at a further decrease compared to earlier findings. Data from the German Environmental Survey (GerES II) 1990/1992 (Becker et al., 1997; Seifert et al., 2000) are especially useful for a time-related comparison with the results of the present study, as both are Germany-focused surveys that are similar with respect to age of children and time of sampling. With regard to the urine mercury content, GerES II yielded a GM of 0.56 g/l for 8- to 11-year-old children (present study: 0.26 g/l for the ages 8–10). GerES II provided a GM of 0.27 g/l (present study: 0.19 g/l) for 6- to 14-year-old children without amalgam fillings. For children with teeth with amalgam fillings, GerES II differentiated among three groups: one to three teeth (GM 0.55 g/l), four to six teeth (GM 1.37 g/l), more than seven teeth (GM 3.03 g/l). The present study shows an overall GM of 0.45 g/l for children with amalgam fillings.

The most recent GerES III of 1998 provided a GM of 0.43 g/l for adults and, thus, endorses the above-mentioned trend. Unfortunately, data from children are not available from GerES III (Becker et al., 2002a, b).

In a study conducted on 6-year-old children in 1991, Walkowiak et al. (1998) obtained urine mercury values of 0.16 g/l (GM). Arithmetic means found by Schulte et al. (1994) in 3- to 15-year-olds without amalgam fillings were 0.17 g/l urine and 0.19 g/g creatinine, which corresponds to the levels of mercury in urine of the present study. In 1992 and 1993, Trepka et al. (1997) detected urine mercury levels of 0.36 g/g creatinine (GM) in children aged 5–14 years. As the study took place in a heavily polluted (chemical production and coal mining) East German area, the higher levels of mercury in urine are plausible. Nonetheless, also in the Trepka et al. (1997) study, the most significant factor affecting urinary mercury levels was found to be the number of dental amalgam fillings.

The determination of the mercury concentration in urine is considered to be the method of choice to evaluate a long period of exposure to mercury and inorganic mercury species. In this study, except for one value (5.31 g/l), levels of mercury found in urine did not reveal health risks due to mercury exposure when compared with human biological monitoring (HBM) values as given by the German Commission on Human Biological Monitoring (Ewers et al., 1999). HBM I indicates the concentration of an environmental toxin in a human biological material below which there is no risk for adverse health effects in individuals, whereas HBM II values for children and adults denote a concentration above which there is an increased risk for adverse health effects in susceptible individuals. For mercury, HBM I and HBM II are 5 and 20 g/l, respectively, for both children and adults.

The fact, that in three children without amalgam fillings the mercury concentration in urine exceeded the reference value set by the German Commission for Human Biomonitoring of 1.4 g/l (for children from 6 to 12 years), does not predicate any negative health impact. Reference values are derived from population studies intended to measure the concentration of an environmental toxin in blood, urine, or other human biological materials of subjects among the general population, and with respect to mercury in urine refer to individuals without amalgam fillings (Ewers et al., 1999).

The positive correlation between urine mercury concentration and the number of amalgam fillings/amalgam filling surfaces carried out in the present study (Figures 4 and 5) endorses the results of previous studies .

Hair is established as a screening matrix specially targeted for determination of organic mercury, but also of inorganic mercury (Schweinsberg and Kroiher 1994; Wilhelm and Idel, 1996). On the other hand, mercury determination in hair is not yet accepted for assessment on an individual basis (Wilhelm and Idel, 1996). The mercury content in hair reflects mainly the uptake of organic mercury compounds via fish consumption.

Referring to hair mercury levels, the present study yields a GM of 0.18 g/g. All but four hair mercury levels were below the GM of 0.77 g/g (ages 6–16 years) reported from a Spanish study (Batista et al., 1996). This result might be explained by fish consumption being traditionally higher in Spain than in Germany.

In a study performed in the USA between 1995 and 1997, the mean mercury concentration in hair was 0.32 g/g (Pellizzari et al., 1999). Only 21% of the participants were younger than 14 years. Hence, the comparability of data to those of the present study is limited.

As published by the US American National Health and Nutrition Examination Survey "NHANES 1999" (CDC, 2001), the mercury concentration in hair in 1- to 5-year-old children is around 0.4 g/g (90th percentile), which corresponds well with the 90th percentile of the present study (0.47 g/g). In "NHANES 1999," GM values were not calculated for hair mercury contents.

In GerES II, mercury in hair was not determined. The positive association between fish consumption and mercury content in hair (Figure 7) was already found in the above-mentioned study from Batista et al. (1996).

Cigarette smoking correlated positively, whereas the age of the individual correlated negatively with the mercury content in hair. Since cigarettes contain mercury, the relationship between the smoking habits and the mercury content in hair was predictable.

In this study, the correlation between the hair mercury content and urine mercury concentration was rather low (Figure 8). The analysis of hair cannot be considered as an alternative method to the analysis of urine, though the sampling procedure is much easier than urine collection. Nevertheless, both methods contribute to a comprehensive estimation of the total mercury exposure, since mercury analysis in hair is preferred concerning the estimation of the body burden related to organic mercury compounds, whereas the results of the analysis of mercury in urine are a better indicator for exposure to inorganic mercury.

Saliva has already been tested as a substitute matrix for assessing the mercury load (Ott, 1993). A disadvantage of mercury determination in saliva is that the analytical method has not yet been standardized and the interpretation of the obtained data remains unclear.

The saliva levels obtained in this study were very low, with more than 70% being lower than the limit of quantification. Saliva has been used to assess mercury release from amalgam fillings for adults with complaints self-related to amalgam (Lygre et al., 1999). The median of mercury concentration in saliva was 9.5 g/l and therewith far beyond the findings of the present study. The low mercury levels in the present study are probably due to the fact that it dealt with children with a relatively low number of amalgam fillings. Furthermore, saliva was sampled without salivary stimulation and only the supernatant saliva was used for the analysis.

Conventional saliva sampling is a noninvasive procedure and saliva mercury analysis may be of importance in HBM studies, especially in studies dealing with children. However, besides the low mercury levels in saliva of children, it has to be considered that, to the present day, there are not any existing reference values or toxicologically justified values concerning the element mercury in the matrix saliva.

Most probably, behavioral peculiarities such as bruxism and chewing gum increase the saliva mercury concentration.

There is some evidence for adverse behavioral effects associated with low mercury exposures (Echeverria et al., 1998) and cognitive deficits in children due to methylmercury (Grandjean et al., 1997). On the other hand, there is no reasonable suspicion for health impairment caused by correctly posed amalgam fillings (Harhammer, 2001). Although the present study shows a decrease in mercury levels, for preventive purposes in environmental health, a further reduction of mercury exposure is desirable. The US American Center of Disease Control and Prevention proposed that the long-term strategy for reducing exposure to mercury is to lower concentrations of mercury in fish by limiting mercury releases into the atmosphere from burning mercury-containing fuel and waste, and from other industrial processes (CDC, 2001).

Conclusion

The analysis of urine and hair has shown that the mercury burden in children has further decreased compared to preceding surveys referring to children with and without amalgam fillings. For a rough estimate of mercury exposure caused by amalgam, it is sufficient to determine the number of amalgam fillings rather than the less readily available surface of amalgam fillings. At present, saliva is not a suitable material for biological monitoring to assess mercury exposure in children, at least not in cases of low mercury exposure.

References

1. Batista J Schuhmacher M Domingo JL Corbella J Mercury in hair for a child population from Tarragona Province, Spain, Sci Total Environ (1996) 193: 143–148 | PubMed |
2. Becker K Kaus S Krause C Lepom P Schulz C Seiwert M Seifert B German Environmental Survey 1998 (GerEs III): environmental pollutants in blood of the German population, Int J Hyg Environ Health (2002a) 205 4(in press)
3. Becker K Schulz C Kaus S Seiwert M Seifert B German Environmental Survey 1998 (GerEs III): environmental pollutants inurine of the German population, Int J Hyg Environ Health (2002b) (in press)
4. Becker K Seiwert M Bernigau W Hoffmann K Krause C Nöllke P Schulz C Schwabe R Umwelt-Survey 1990/92, Quecksilber-Zusammenhangsanalyse. vol. VII WaBoLu-Hefte 6/96 Umweltbundesamt, Berlin, Germany 1997
5. Begerow J Zander D Freier I Dunemann L Long-term mercury excretion in urine after removal of amalgam fillings, Int Arch Occup Environ Health (1994) 66: 209–212 | PubMed |
6. Björkman L Sandborgh-Englund G Ekstrand J Mercury in saliva and feces after removal of amalgam fillings, Toxicol Appl Pharmacol (1997) 144: 156–162 | PubMed |
7. Center of Disease Control and Prevention (CDC) Blood and hair mercury levels in young children and women of childbearing age—United States, 1999, JAMA, J Am Med Assoc (2001) 285(11): 1436–1437
8. Echeverria D Vasken Aposhian H Woods JS Heyer NJ Aposhian MM Bittner AC Jr Mahurin RK Cianciola M Neurobehavioral effects from exposure to dental amalgam Hg°: new distinctions between recent exposure and Hg body burden, FASEB J (1998) 12: 971–980 | PubMed | ISI |
9. Eneström S Hultman P Does amalgam effect the immune system? A controversial issue, Int Arch Allergy Immunol (1995) 106: 180–203 | PubMed |
10. Ewers U Krause C Schulz C Wilhelm M Reference values and human biological monitoring values for environmental toxins, Int Arch Occup Environ Health (1999) 72: 255–260 | Article | PubMed | ChemPort |
11. Ganss C Gottwald B Traenckner I Kupfer J Eis D Mönch J Gieler U Klimek J Relation between mercury concentrations in saliva, blood, and urine in subjects with amalgam restorations, Clin Oral Invest (2000) 4: 206–211
12. Grandjean P Weihe P White RF Debes F Araki S Yokoyama K Murata K Sorensen N Dahl R Jorgensen PJ Cognitive deficit in 7-year-old children with prenatal exposure to methylmercury, Neurotoxicol Teratol (1997) 19(6): 417–428 | Article | PubMed | ChemPort |
13. Harhammer R Zur Risikobewertung des zahnärztlichen Füllungswerkstoffes Amalgam, Bundesgesundheitsbl-Gesundheitsforsch-Gesundheitsschutz (2001) 44: 149–154 in German | Article |
14. Herber RFM de Gree AJ Wibowo AAE Exposure of dentists and assistants to mercury: mercury levels in urine and hair related to conditions of practice, Commun Dent Oral Epidemiol (1988) 16: 153–158
15. Krämer U Koch T Ranft U Ring J Behrendt H Traffic-related air pollution is associated with atopy in children living in urban areas, Epidemiology (2000) 11: 64–70 | Article | PubMed | ChemPort |
16. Lygre GB Hol PJ Eide R Isrenn R Gjerdet NR Mercury and silver in saliva from subjects with symptoms self-related to amalgam fillings, Clin Oral Invest (1999) 3: 216–218
17. Olstad ML Holland RI Wandel N Petersen A Correlation between amalgam restorations and mercury concentrations in urine, J Dent Res (1987) 66: 1179–1182 | PubMed |
18. Ott KHR Die Messung der Quecksilberbelastung im Speichel, Dtsch Zahna¨rztl Z (1993) 48: 154–157
19. Pellizzari ED Fernando R Cramer GM Meaburn GM Bangerter K Analyses of mercury in hair of EPA Region V population, J Expos Anal Environ Epidemiol (1999) 9(5): 393–401
20. Sällsten G Thorén J Barregard L Schütz A Skarping G Long-term use of nicotine chewing gum and mercury exposure from dental amalgam fillings, J Dent Res (1996) 75: 594–598 | PubMed |
21. Schierling P Schaller KH Einfache und zuverlässige Methoden zur atomabsorptionsspektrometrischen Bestimmung von Quecksilber in Blut und Urin, Zeitschr für Praxis, Klinik, Forschung und Begutachtung (1981) 3: 57–61
22. Schulte A Stoll R Wittich M Pieper K Stachniss V Quecksilberkonzentration im Urin von Kindern mit und ohne Amalgamfüllungen, Schweiz Monatsschr Zahnmed (1994) 104: 1336–1340 | PubMed |
23. Schweinsberg F Kroiher A Quecksilberbelastung durch Fischkonsum bei Rheinfischern, Zbl Hyg (1994) 195: 529–543 in German
24. Seifert B Becker K Helm D Krause C Schulz C Seiwert M The German Environmental Survey 1990/92 (GerES II): reference concentrations of selected environmental pollutants in blood, urine, hair, house dust, drinking water and indoor air, J Expos Anal Environ Epidemiol (2000) 10: 552–565 | Article | ISI | ChemPort |
25. Silbergeld EK New approaches to monitoring environmental neurotoxins In: Malamud D., and Tabak L. (Eds.) Saliva as a Diagnostic Fluid Ann NY Acad Sci (1993) 694: 62–71 | PubMed |
26. Soleo L Elia G Apostoli P Vimercati L Pesola G Gagliardi T Schiavulli N Drago I Lasorsa G Russo A The influence of amalgam fillings on urinary mercury excretion in subjects from Apulia, G Ital Med Lav (1998) 20(2): 75–81
27. Trepka MJ Heinrich J Krause C Schulz C Wist M Popescu M Wichmann H-E Factors affecting internal mercury burdens among eastern German children, Arch Environ Health (1997) 52(2): 134–138 | PubMed |
28. Vondenhof T Beindorf H Beschreibung eines Infrarotstrahlers aus Quarzglas zur Bestimmung von Quecksilber in Wasser und Lebensmitteln, Dtsch Lebensm-Umsch (1976) 5: 158–159
29. Walkowiak J Altmann L Krämer U Sveinsson K Turfeld M Weishoff-Houben M Winneke G Cognitive and sensorimotor functions in 6-year-old children in relation to lead and mercury levels: adjustment for intelligence and contrast sensitivity in computerized testing, Neurotoxicol Teratol (1998) 20: 511–521 | PubMed |
30. WHO Environmental Health Criteria 101, Methylmercury. Geneva, Switzerland 1990 pp 24–37
31. WHO Environmental Health Criteria 118, Inorganic Mercury. Geneva, Switzerland 1991 pp 29–34
32. Wilhelm M Idel H Hair analysis in environmental medicine, Zbl Hyg (1996) 198: 485–501
33. Wilhelm M Lombeck I Kouros B Wuthe J Ohnesorge FK Duplikatstudie zur Aufnahme von einigen Metallen/Metalloiden bei Kindern in Deutschland: Teil I. Arsen und Quecksilber, Zbl Hyg (1995) 197: 345–356
34. Wilhelm M Müller F Idel H Biological monitoring of mercury vapour exposure by scalp hair analysis in comparison to blood and urine, Toxicol Lett (1996) 88: 221–226 | PubMed |
35. Wilhelm M Pesch A Rostek U Begerow J Schmitz N Idel I Ranft U Concentrations of lead in blood, hair and saliva of German children living in three different areas of traffic density, Sci Total Environ (2002) (in press)

Acknowledgements

The authors thank Dr. L. Lajoie-Junge for final linguistic manuscript check.