Introduction
Exposure profiles are one of the fundamental aspects of studies of lead and its impact on neuropsychological and neurobehavioral outcomes in children. As the skeleton is the primary storage compartment for lead in the human body, it is a potential endogenous source of lead that may be released from the bones during pregnancy (1,2), during lactation (2), and after menopause (3). Chronic exposure to lead, such as from mouthing activity in early childhood, may be camouflaged by dilution of lead in bone during periods of rapid skeletal growth and so may not be detected by the normal methods of blood lead (PbB) analysis. Hence, knowledge of past lead exposure, especially in utero, is fundamental to any investigations of lead toxicity.
In many human epidemiological investigations, exposure profiles may be limited to a single PbB level (4), which may provide information only about recent exposure. The use of lead in whole deciduous teeth, enamel, or dentine as an indicator of past exposure of children to lead, and as a proxy for skeletal lead, has been well documented in a number of studies [see Gulson and Wilson for review (5)]. Gulson and Wilson (5) and Edwards-Bert et al. (6) assessed various aspects of tooth lead studies. The advantage of tooth lead is that it is an indicator of lead exposure over several years, from in utero to loss of the tooth, in contrast to blood lead, which has an approximate mean life of 30 days (7). However, because of differences in the type of tooth analyzed, the part of the tooth analyzed, and the analytical techniques used for lead measurement, tooth lead has not been widely accepted as an indicator of lead exposure. In a pilot study, Gulson and Wilson (5) showed that stable lead isotopic analyses and lead concentrations in cross-sectional slices of deciduous teeth provide evidence of in utero and earliest childhood exposure when using enamel, and the results for dentine provided evidence of exposure during early childhood, possibly up until the time the tooth was shed.
We evaluated the efficacy of the isotopic analyses of slices of teeth as an indicator of past exposure in 30 exposed and nonexposed children from a lead mining community. Resulting data would also allow an evaluation of the hypothesis that lead in dentine does not turn over (8,9), a hypothesis that has been misinterpreted to indicate that tooth lead, except from that in the circumpulpal dentine, is a passive reservoir for lead.
Methods
Deciduous teeth, mostly upper and lower incisors, from 30 children from the Broken Hill lead mining community with differing exposure to lead were tested. In one subject, permanent teeth were also available, and these allowed a comparison of exposure through different stages of childhood. Estimates of exposure were based on information obtained from parents and included residence in areas identified as "high risk" by a blood lead survey of 899 children aged 1-4 years (10), early childhood mouthing frequency, learning difficulties, and behavioral problems. With one exception, the mothers of the children were long-term residents of Broken Hill (>10 years).
 |
|
Figure 1. Schematic representation of the sagittal section of a human tooth showing the locations of the cross-sectional slices taken for analysis in this study. |
Except for the permanent teeth and three deciduous canines with roots from the one subject, the samples consisted only of crowns, which had undergone varying amounts of resorption. For upper and lower incisors, the crowns were cut transversely into 1-2 mm thick slices from the incisal and cervical areas using a diamond-impregnated stainless-steel disc. The incisal section consisted of enamel and varying amounts of coronal dentine. In this paper, we use the nomenclature of Shapiro et al. (9): enamel, dentine, secondary dentine (the dentine zone around the pulpal canal, also called circumpulpal dentine) and pulp. The locations from which the incisal, cervical and root sections were cut for the analyses in this paper are shown in Figure 1. As enamel and coronal dentine are formed before most children begin to crawl (11), they provide an indicator for in utero and earliest childhood exposure. For enamel samples, as much attached dentine as possible was removed. The pulpal canal in the cervical section was usually resorbed to varying degrees, but to ensure minimal contribution from secondary (circumpulpal) dentine, approximately 2 mm of the pulpal canal and dentine was reamed out. The dentine in the cervical section provides an integrated exposure to lead from the time of eruption of the tooth until the tooth is shed (12,13). Gulson and Wilson (5) suggested that the sectioning approach was superior to other methods of tooth preparation (or whole teeth), including that of circumpulpal wedges (9,12,13). They suggested that the major drawbacks of the circumpulpal wedge method were 1) the relatively limited quantity of the circumpulpal dentine in deciduous teeth, especially in naturally shed teeth, where there is variable resorption, 2) the high level of technical skill required to remove the wedge, 3) analytical difficulties on a routine basis, and 4) the strongest correlation of tooth lead and PbB was at about 4 years of age (12,13), after the maximum PbB levels in children. For permanent teeth, 1-2 mm sections of the outer crown and the root sample were taken from an area approximately 1-2 mm from the root tip (Fig. 1).
Tooth slices were decontaminated and analyzed as described by Gulson and Wilson (5), except that 1% HNO3 was used, and tooth samples weighing 2-60 mg were completely dissolved in 6 M HCl.
The isotope dilution method was used to analyze the lead isotope abundances and lead concentrations in the same sample. This involves adding to the test sample, before digestion, a known amount of a 46% 202Pb solution of known isotopic abundance (composition). The lead is separated from interfering elements by anion exchange chromatography. Processing "blank" levels were < 150 pg Pb; no corrections for this blank have been made to the data, as it is insignificant compared with the amount of lead in the analyzed sample.
High precision isotope ratios of 208Pb/206Pb, 207Pb/206Pb, and 206Pb/204Pb were measured on a VG Isomass 54E thermal ionization mass spectrometer (VG Isotopes, Winsford, UK) in fully automatic mode. The external precision of the 207Pb/206Pb isotope ratios is 
0.06% (2
), based on over 1800 analyses in the CSIRO laboratories of the internationally recognized Lead Standards SRM 981 and 982 of the National Bureau of Standards (NBS) and natural samples. An analysis of SRM 981 was performed with each batch of samples. Accuracy of the measured isotope ratios in the teeth samples is by way of normalization of the ratios to those given by NBS (now National Institute for Standards and Testing). Validation of the laboratory was a prerequisite for undertaking the project, "Biokinetics of Lead in Human Pregnancy."
Statistical procedures of t-tests for evaluating significances of differences between means of data sets and regression analyses were provided in the Microsoft package Excel 5.0 (Microsoft Corp., Redmond, Washington). The use and interpretation of data with notched box plots are described in McGill et al. (14) and Cleveland (15).
Results and Discussion
The lead isotope technique uses the four isotopes of lead. Three are the stable end products of radioactive decay of uranium and thorium: 238U to 206Pb, 235U to 207Pb, and 232Th to 208Pb. The abundance of the fourth, 204Pb, is essentially constant, and this isotope is commonly used as a reference isotope. Because three isotopes of lead are produced by radioactive decay, the amounts (abundances) have changed over geological time, and this is reflected in the geological source of the lead. The abundances are usually expressed as ratios so that lead from the geologically old (about 1700 million years old) lead-zinc-silver deposits of Broken Hill in New South Wales has an abundance ratio of the 206Pb isotope to the 204Pb isotope (206Pb/204Pb) of 16.0, whereas the ratio is 18.1-18.3 for geologically young deposits (500-400 million years old) on the same continent. The isotopic differences are used to evaluate the source of lead in the environment, humans, and animals. Interpretations of lead isotopic data may not be straightforward because lead in the environment or animals may be a mixture of lead from different sources (mines). Hence, lead that is introduced to the body from soil, dust, air, food, or water is largely dependent on the source of lead in the environment, which in turn is dependent on the age and isotopic composition of the rocks and ores from which the lead in the environment is derived.
The ability to obtain on each sample three sets of lead isotope values, representing variations in the abundances of the isotopes, as well as lead concentrations, allows for more rigorous evaluation of the data than, for example, a single value of lead concentration or isotope ratio. Hence, the statistical results listed in Table 1 incorporate the three sets of measured ratios as well as lead concentrations. For simplicity, usually only one isotope ratio is discussed, most commonly, the abundance of 206Pb to 204Pb, as the 206Pb/204Pb ratio.
Figure 2. Isotopic compositions expressed as the 206Pb/204Pb for diverse sources of lead in the Broken Hill mining community (16).
Broken Hill is a city of approximately 25,000 inhabitants, about 930 km west of Sydney, New South Wales, Australia, and centered about the world's largest currently mined lead-zinc-silver deposit. Mining activities, including underground and open-pit operations and smelters in the latter part of the past century, have been conducted for more than 100 years. This area is desert and subject to severe windstorms, and the dust from the mining activities and potentially from weathering of the lead-zinc-silver mineral deposit over millions of years is considered to be the main point source of lead in children from inhalation and ingestion of contaminated soil and household dust. The isotopic composition of the potential sources of lead in the Broken Hill community are summarized in Figure 2. The 206Pb/204Pb ratio generally ranges from 16.0 to 16.2 for the orebody lead and dust from ceilings, vacuum cleaners, and kitchen wipes (16). The other potential sources of lead are food, water, and air. Food and water contain < 10 ppb and < 3 ppb lead, respectively, and are the main contributors to background PbB concentrations of 6 2 
g/dl, which have been estimated from 38 minimally exposed female adults (16). Apart from orebody lead, another major source of lead in air is from gasoline; at the time of the investigation, approximately 60% of automobiles in Broken Hill used leaded gasoline, with approximately 0.8 g/l lead and with a 206Pb/204Pb ratio of 16.56. Gasoline was considered to be a significant source of lead in adult females from Broken Hill (16). Until the lead isotopic investigations of Gulson et al. (16), paint had been largely ignored as a potential contributor to blood lead in children from Broken Hill. Houses built before the 1970s commonly contain lead paint and far outnumber newer dwellings. Renovation of houses is extremely common in Broken Hill. Isotopic ratios measured in paint (Fig. 2) illustrate the complexity of lead sources but do not negate the use of isotopic analyses for elucidating sources in Broken Hill.
It is possible to estimate relative proportions of lead from some sources using the isotopic ratios, such as the orebody and gasoline. For example, given that the orebody has a 206Pb/204Pb value of approximately 16.0 and gasoline 16.56, a tooth slice with a ratio of 16.28 would have a 50% contribution from each source. Estimating proportions for paint is more complex, but in some cases it is still possible to obtain approximate contributions.
 |
 |
|
Figure 3. Notched box plots for descriptive statistical data for the isotope ratios 208Pb/206Pb and 207Pb/206Pb for within-tooth variation and between groups of subjects from Broken Hill. The upper and lower ends of each box (rectangular areas) are the upper and lower quartiles. The distance between these two values (interquartile ranges) is a measure of the spread of the distribution. The relative distances of the upper and lower quartiles from the median give information about the shape of the distribution of the data. (If one distance is much bigger than the other, the distribution is skewed.) The notches surrounding the median provide a measure of the rough significance of differences between the values. Specifically, if the notches about two medians do not overlap, the medians are roughly significantly different at about a 95% confidence level. The lack of overlap in notches surrounding the medians (the white bar), such as between the low incisal and low cervical data, indicate that the medians are significantly different at the 95% confidence level. The dashed lines extending from the box plots represent adjacent values. If r is the interquartile range, the upper adjacent value is the largest observation that is less than or equal to the upper quartile + 1.5r. The lower adjacent value is the smallest observation that is greater than or equal to the lower quartile - 1.5r. There is one outside value in the 208Pb/206Pb and 207Pb/206Pb data for paint-exposed children and three outside 207Pb/206Pb values for the low incisal group. |
Figure 4. Notched box plots for descriptive statistical data for the isotope ratios 206Pb/204Pb and lead concentrations for within-tooth variation and between groups of subjects from Broken Hill. |
Subjects with low exposure. Descriptive statistics for subjects with low and high lead exposure are illustrated with notched box plots in Figures 3 and 4, and tests of significance are listed in Table 1; detailed isotopic results for subjects with low and high exposure can be obtained from the author. In general, the lead concentrations in the incisal sections for subjects with low exposure range from 0.4 to 3.5 ppm Pb, with a mean value and standard deviation of 1.2 0.8 ppm. These low lead concentrations demonstrate that the children were exposed to low levels of lead in utero and during early childhood. The low lead concentrations are consistent with those found in unexposed populations (17,18) but are up to 100 times greater than natural levels (19). The 206Pb/204Pb ratios range from 16.31 to 16.64 (mean 16.48 0.10), consistent with those found in the blood of female adults (16).
Variable lead concentrations and 206Pb/204Pb ratios are observed in the cervical sections, but there does not appear to be any simple relationship such as increasing lead concentration with decreasing 206Pb/204Pb, which would reflect an increasing amount of lead from an orebody source as a result of hand-to-mouth activity. The correlation coefficients for 206Pb/204Pb versus Pb (ppm) in the incisal and cervical sections are -0.10 (p = 0.75) and -0.12 (p = 0.69), respectively. The lack of simple relationships is also indicated by the observation that the lead concentrations in both cervical and incisal sections of one subject are low, and yet the isotopic composition indicates that the child has a substantial component (approximately 40%) of orebody lead in the teeth.
In general, within-tooth variations show that the cervical section has higher amounts of lead (p = 0.0007; Table 1) and lower 206Pb/204Pb ratio (p = 0.001 in Table 1) than the incisal section, indicating that during early childhood, the individuals' intake of orebody lead exceeded that from other sources.
Subjects with high exposure. Lead concentrations and 206Pb/204Pb ratios are variable in incisal and cervical sections for these subjects (Table 1; Figs. 3 and 4). Low lead concentrations of < 2 ppm and higher 206Pb/204Pb ratios of >16.4 in the incisal tooth slices of some subjects reflect a low-lead exposure in utero and during early childhood. There are higher concentrations of lead in the cervical sections compared with incisal sections (p = 2 X 10-6; 22 df), and there is a stronger correlation of increasing lead concentration with decreasing 206Pb/204Pb in the cervical (correlation of -0.64; p = 0.0004) compared with the incisal section (correlation of 0.15; p = 0.46). The within-tooth variation, represented by the difference between the incisal and cervical sections, is slightly larger (p = 0.003 in 206Pb/204Pb) than in the low-exposure subjects. The differences are also greater in the order: 208Pb/206Pb ~ 207Pb/206Pb > 206Pb/204Pb. In the statistical analyses, the data are included in the high exposure group for deciduous teeth from the one family (Table 3) who suffered high lead exposure from orebody lead during renovations.
For most subjects, the isotopic ratios in the cervical (and often incisal) sections indicate that during early childhood, the individuals' intake of orebody lead exceeded that from other sources.
Differences in tooth components. In addition to within-tooth differences in lead concentration and isotopic ratios, there were significant differences between the high and low lead-exposed children for the incisal and cervical sections (Table 1; Figs. 3 and 4).
Lead from a paint source. The problem of source identification is highlighted by the siblings 517, whose teeth have some of the highest lead concentrations analyzed so far in this cohort. Until paint was identified as a potential lead source in Broken Hill (16), paint was largely ignored because of the obvious point source of the lead orebody and associated mining activities. Results are presented in Table 2. The 206Pb/204Pb ratios in both incisal and cervical sections for the siblings from house 517 are in the range 16.45-16.67. Even though the family lived within 300 m of mining activities and > 90% of the lead in the surface dust and soil from the house was derived from an orebody source, the isotopic data indicated that there was a significant contribution of lead from another source, such as paint. Paint was not, however, considered to be a source of lead by the parents because extensive renovations were supposed to have been carried out several years before the children were born. However, an initial vacuum cleaner dust sample had a 206Pb/204Pb ratio of 16.98, totally different from the orebody and indicative of a source from paint. Inspection of the vacuum dust by optical and scanning electron microscopy identified numerous lead paint flakes with evidence of burning. It is estimated from the isotopic ratios in the paint flakes (18.37, 18.31) that the lead in the children's enamel and dentine consists of approximately 30% paint lead and 70% orebody lead (assuming gasoline to be a minor contributor). After obtaining these results, and upon further discussions with the parents, it was revealed that they had burned off many layers of paint from skirting boards and doorways when the children were very young.
The data for the incisal tooth sections indicate that the children received a lead insult in utero and during early childhood. However, the mother exhibited no evidence of an earlier increased lead burden, as her PbB was 4.7 g/dl and her 206Pb/204Pb ratio was 16.57, consistent with other female adults from Broken Hill. Similarly, there was no evidence of this past exposure in the father, whose urine lead was normal at 4.1 g/l and whose 206Pb/204Pb was 16.42. These normal results for the parents suggest that their exposure was acute and lead from paint was not transferred in significant amounts to long-term bone compartments. In contrast, PbB levels in the children were relatively elevated for their age: approximately 14 g/dl and with similar 206Pb/204Pb ratios of 16.40, perhaps reflecting leakage (mobilization) of lead from earlier accumulated skeletal stores and/or from ongoing exposure. (The deficiencies observed by the mother in reading, bilateral coordination, and balance and visual motor control of the older male sibling may be related to a chronic insult in utero and early childhood.)
There are statistically significant differences in lead concentration for the incisal and cervical sections of the paint-exposed children (p = 0.009) but not in isotopic composition (Table 1; Figs. 3 and 4). There are, however, statistically significant differences in isotopic composition for the incisal and cervical sections between the low-exposed and high-exposed and paint-exposed children, with the largest differences in isotopic composition for the high- and paint-exposed groups (Table 1). Such differences may be expected given the large range in isotopic composition between orebody lead and most paint lead (Fig. 2). The differences in lead concentration for paint and high-exposed children are not statistically significant (p = 0.08, 0.35; Fig. 4).
Variations during pregnancy and between deciduous and permanent teeth. The contribution to enamel lead from different sources and the impact of renovations during pregnancy were mentioned above for siblings from house 517. In another family, it was possible to obtain several teeth, including deciduous and permanent teeth, from two siblings. Extensive contamination from renovations of this house, located within 500 m of the mining activities, occurred during the pregnancy with the older sibling. Additional contamination may have resulted from transport of lead dust to the residence on the clothes of the male adult, who was employed as an underground worker in the mine for over 20 years. The mine work clothes were laundered at home in Broken Hill.
Incisal and cervical tooth sections from the male sibling (subject 534 M) exhibit the highest contribution of orebody lead of any subject so far analyzed in Broken Hill (Table 3). The data are consistent with a larger insult of orebody lead than for the younger female sibling, although the lead concentrations for the same tooth type (primary molar) are similar.
Data for the permanent teeth show an interesting contrast to those for deciduous teeth (Table 3). Crowns (enamel) of the permanent teeth contain higher amounts of lead than the roots, in contrast to that observed in most deciduous teeth from children at Broken Hill and also from this same subject. This is not unexpected as the development of enamel in permanent first molars is completed at 2.5-3 years (11), the time of maximum mouthing activity (20). The roots of the permanent teeth of subject 534 M contain relatively high amounts of lead, and the 206Pb/204Pb ratios are slightly higher than the enamel, indicating a contribution of lead from food, water, and air; the 206Pb/204Pb ratios in these media are generally >16.3 in Broken Hill (16). The PbB in the children, aged 11 and 9 years at the time of blood sampling, were elevated (19.9 and 19.3 g/dl Pb, respectively), and the 206Pb/204Pb ratios (16.19 and 16.24, respectively), were consistent with the data for their teeth. The isotopic data for their blood probably reflect ongoing mobilization of lead from skeletal tissues introduced to the bone compartments during pregnancy and early childhood (21) and/or ongoing lead intake from the high-risk location.
As with subjects 517 M and 517 F, there was no indication from the mother's current PbB of 6.7 g/dl (206Pb/204Pb of 16.48) of this earlier high exposure to lead.
Enamel versus dentine lead. In more than 90% of the analyses of >150 deciduous and permanent teeth, the concentrations of lead have been higher, sometimes up to an order of magnitude, in the coronal dentine compared with the enamel. In contrast, Malik and Fremlin (22), using charged-particle activation analyses, found that in molars of subjects < 25 years old, lead in enamel was higher than in dentine. Incisors and canines are the most common teeth used in studies of children, so investigations of molars may not be relevant to the argument about different concentrations of lead in incisors.
Furthermore, our data have shown that analyses of cervical sections, representative of coronal dentine, substantially underestimate the exposure of a child given the high amounts of lead observed in the roots, when the roots are available for analysis (5). This underestimation is exacerbated if a whole tooth is analyzed because of dilution from the lower lead concentrations in the enamel and also due to the fact that considerably more weight (as against volume) of a tooth is in the highly mineralized apatite in the enamel compared with that in dentine (23). In addition, incorporation of the secondary dentine (circumpulpal dentine) in the tooth analysis may overestimate tooth lead, as the secondary dentine may dominate the tooth lead (9,24).
Steenhout (25) and Bercovitz and Laufer (26) suggested that lead can be added to deciduous and permanent dentine, and Rabinowitz et al. (12) suggested that the permeable nature of dentine allows for loss of lead for as long as 1 year after calcification is complete, only later becoming a sealed repository. Rabinowitz et al. (27) advanced several models to account for relationships between blood leads and tooth lead and suggested that lead was not permanently fixed in dentine. The significant isotopic differences between the lead in enamel and coronal dentine (and roots, when available), demonstrate an ongoing addition of lead to dentine from endogenous and exogenous sources but, at this stage, provide no evidence for the reverse process.
Nevertheless, it should be possible to obtain an estimate of the rate of change, or addition, of lead to dentine in the isotopic composition and lead concentration, assuming that the isotopic composition and lead concentration were the same in enamel and dentine at the time of tooth eruption. The estimation for the rate of change can be calculated from the difference between the incisal and cervical sections from the 207Pb/206Pb, 206Pb/204Pb and lead concentrations in each deciduous tooth. The change from enamel to dentine (incisal to cervical section) in 206Pb/204Pb for high lead-exposed children ranges from 0 to 24 units and 1 to 20 in the low-exposed children. Even though there is a large overall variation, there is fairly good agreement with the change in values for teeth from the same subject. The Student'st-test for 207Pb/206Pb and 206Pb/204Pb in high- and low-exposed children gives p-values of 0.14 and 0.19, respectively (18 df), which are not statistically different. The mean change of about 10 units in 206Pb/204Pb relative to the overall change of 56 units (the difference between petrol at 16.56 and orebody at 16.00) gives an estimate of between 2-3% per year lead addition over a period of 7-8 years. This rate of change of lead in dentine in deciduous teeth is the same as what we measured in permanent teeth for immigrants to Australia who resided in Australia for varying lengths of time (Gulson et al., in preparation). At this stage, the rate of loss of lead from dentine is unknown.
Conclusions
The approach of slicing teeth into incisal and cervical sections, combined with lead isotope measurements, provides an excellent history of lead exposure for children. If there are no changes to lead exposure over the lifetime of the deciduous tooth (e.g., approximately 7 years), then it is feasible to use a whole tooth for analysis.
Teeth from children with low lead exposure generally contained < 2 ppm in the incisal section (predominantly enamel and primary dentine), and the isotopic compositions were consistent with the lead exposure in utero. The cervical sections (predominantly dentine) contained up to 9 ppm Pb, and the isotopic compositions varied depending on the source of the lead. The low lead concentrations in coronal dentine were consistent with those observed in other studies of unexposed children.
In high to moderately lead-exposed children, the amounts of lead and isotopic compositions in incisal and cervical sections varied widely; the cervical sections generally contained more lead, and the isotopic compositions reflect the dominant source(s). Two siblings from Broken Hill with the highest amounts of tooth lead appear to have derived this lead in utero and during early childhood from lead paint released during renovations, even though the dwelling was within 300 m of mining activities. In another subject from Broken Hill, increased exposure in utero and early childhood showed the lead in teeth was derived from an orebody source, probably also during renovations. In both these cases, there was no indication from the mothers' current blood lead of any previous high exposure to lead. Crowns and roots of permanent teeth from one subject yielded data that were consistent with those from his deciduous teeth. Introduction of lead into dentine of deciduous teeth is provisionally estimated to occur at a rate of 2-3% per year, but the rate of loss from dentine is unknown at present.
