Introduction

Lead and zinc are common contaminants at hazardous waste sites. When these metals are present in the dust or soil of nearby residential neighborhoods, there is a potential for exposure, particularly among children. Children are at greatest risk for experiencing lead-induced health effects, including neurobehavioral deficits and impairment of neurological development (Needleman and Gatsonis, 1990; Agency for Toxic Substances and Disease Registry (ATSDR), 1999). Children's activities and behaviors also result in greater opportunities for contact with lead-contaminated dust and soil (Barnes, 1990; Calabrese et al., 1997). In addition to having more opportunities for oral exposure to contaminated dust and soil, children may absorb a greater proportion of ingested lead than adults (ATSDR, 1999). A study of adults residing near a National Priorities List (NPL) site indicated that the bioavailability of ingested soil-borne lead in adults was approximately 26% (Maddaloni et al., 1998). Based on lead absorption studies, the predicted bioavailability of ingested soil-borne lead in children ranged from 25% to 41% (Ziegler et al., 1978; Casteel et al., 1997; ATSDR, 1999).

Several studies have calculated slope factors to describe the association between environmental measures of lead in soil and dust in and around the home and the blood lead levels of children (Barltrop et al., 1974; Roberts et al., 1974; Baker et al., 1977; Yankel et al., 1977; Stark et al., 1982; Angle et al., 1984; Reagan and Silbergeld, 1989). However, there is considerable variability among these slope factor estimates (ATSDR, 1999; Lewin et al., 1999). The US Environmental Protection Agency (USEPA) offered an improvement to the use of slope factors for predicting blood lead levels based on measured lead in environmental media. The Integrated Exposure Uptake Biokinetic Model for Lead (IEUBK) allows for the integration of environmental lead data from different media and variation in ingestion and absorption parameters (USEPA, 1994). However, data on cocontaminants that could influence the bioavailability of lead have not been incorporated into such models.

Zinc is a nutritionally essential metal and functions both as a structural component of many proteins and as a cofactor for several metalloenzymes. At extremely high doses, persistent exposure to zinc can cause adverse health effects such as anemia and pancreatic damage (ATSDR, 1994b). Animal studies suggest that zinc at moderate exposure levels can influence the absorption and toxicity of other metals (Fischer et al., 1981; Hempe and Cousins, 1992). The interaction between environmental exposures of lead and zinc on blood lead levels has not been studied in humans.

Environmental and biological measures of heavy metals had been investigated among residents of an area that was the site of a large zinc smelting operation (target community) and a nearby comparison area (ATSDR, 1994a). Measures of environmental zinc in the target community were significantly higher than measures of environmental zinc in the comparison community. Despite nearly a century of smelting activity, measures of environmental lead were not higher in the target community for most environmental media. Further, blood lead concentrations were lower for children in the target area compared to the comparison area (ATSDR, 1994a). The purpose of this study is to determine if zinc levels present in the environment affect the association between environmental measures of lead and corresponding biological measurements of lead in children residing in two small communities.

Methods

The data for this analysis were collected in 1991 as part of a community health investigation by the ATSDR (ATSDR, 1994a). Study participants included residents of a northeastern community that had been the site of zinc smelting operations from 1898 to 1980. Residents of a community located approximately 10 miles from the site served as the comparison population in the original investigation. The comparison area was selected on the basis of similar demographics, age of housing, and because it was not affected by zinc smelting, coal mining, or other heavy industrial operations. Individuals from both the smelter and comparison communities were combined for this analysis in order to capture the widest range of environmental zinc levels. As part of the original investigation, a complete door-to-door census was taken to determine the total population and to generate a list of eligible residents. Random samples were selected among residents, stratified by town, sex, and age group (6 – 71 months and 6 – 14 years), who had resided in their homes during the previous 6 months. Of the 485 eligible children in the two communities, 311 were successfully contacted and participated in the study. Approximately one-third of households refused to have their homes sampled for environmental contaminants, leaving 214 children with corresponding environmental and biological measures of lead. Trained interviewers administered a standardized questionnaire, and parents or guardians answered questions for children under 12 years of age. Informed consent was obtained from the parent or guardian of minors. The protocol was reviewed and approved by the Institutional Review Board of the Centers for Disease Control and Prevention (CDC).

Specimen Collection and Analyses

A trained phlebotomist drew venous blood from each participant. Lead-free collection materials were provided by the National Center for Environmental Health (NCEH), Division of Environmental Health Laboratory Sciences, CDC, Atlanta, Georgia. Samples were sent by overnight delivery to CDC for analysis. Blood lead determinations were performed using established Zeeman graphite furnace atomic absorption methods (Miller et al., 1987). Quality-control procedures for blood lead determinations included performing 10% replicates and duplicate analysis of four whole blood pools, whose target values were established by thermal ionizing isotope dilution mass spectroscopy.

Soil and dust samples in and around the homes of all participants were tested for lead and zinc. The street dust/soil sample was a composite of dust/soil collected at the intersection of the street and sidewalk or the intersection of the street and driveway. The sample was collected using a stainless-steel scoopula and placed into a glass jar. The perimeter soil sample was a composite soil sample that comprised three subsamples collected from each of the sides of the house. The subsamples were collected approximately 3 feet from the house and placed into a glass jar. The exterior dust sample was collected from the front, back or side entrance to the home with the heaviest loading of dust. The sample was collected from within the area of a template which measured 25 centimeters (cm) square using a modified vacuum cleaner. The interior dust sample was a composite of dust found in the entry way, the most utilized room, and the child's bedroom. These samples were collected from an area inside a 25 cm square template using a personal monitoring pump connected with tubing to a three piece air monitoring cassette. All samples were analyzed by the US EPA for lead and zinc using established procedures (USEPA, 1991). Four exterior dust samples for lead had measures below the detection limit and were assigned a value of one-half the detection limit, or 7.5 mg/kg.

Statistical Analysis

All biological and environmental measurements were log-transformed to approximate normality. The geometric mean, 5th, and 95th percentiles of blood lead were evaluated separately for gender, age, the year the house was built, household income, and the area (smelter or comparison). Differences in mean blood lead for the different categories of these variables were evaluated using analysis of variance. The geometric mean, 5th, and 95th percentiles of the four environmental measures of lead and zinc were also similarly evaluated.

A total of 22% (n=38) of the households had more than one child participant. Children from the same household could have correlated biological measures, implying there is a lack of independence among observations that is required for fixed effects models. Therefore, mixed models, allowing for both random and fixed effects, were used to account for the correlation between members of the same household. The household variable was included as a random effect, while the remaining variables were treated as fixed effects. The dependent variable, blood lead, was evaluated in regression models using the environmental lead measurements as independent variables. The models also were adjusted for age and sex.

To evaluate the influence of high and low environmental zinc, cutpoint values were estimated using the maximum likelihood method (Kleinbaum et al., 1988). Using this method, we considered all possible cutpoint values in the range of the data, and chose the one (along with the regression coefficients) that minimized the mean squared error. A dichotomous variable was created to indicate whether a person was in an area with high or low zinc levels, as determined by the value being above or below the cutpoint for each environmental source. Interaction terms for environmental levels of zinc and lead were added to multiple linear regression models. The interaction term corresponds to the difference in regression coefficients for the environmental lead variable if the high zinc and low zinc groups were to be modeled separately. Thus, a zero for the interaction coefficient suggests that the regression coefficients on the environmental lead terms for the low and high zinc groups are the same. The error degrees of freedom for the regression coefficients were reduced by one to account for the cutpoint being estimated using the data.

Results

Geometric means for samples of environmental lead and environmental zinc from the residences of children age 6 months to 14 years are presented in Table 1. Zinc concentrations were higher in the smelter area versus the comparison area for all four environmental media. Lead concentrations in exterior and interior dust were not statistically different between the smelter and comparison areas. Lead concentrations in street dust/soil were higher in the smelter area, and lead concentrations in perimeter soil were higher in the comparison area.

Table 1 - Geometric mean (5th and 95th percentile) for environmental measures of lead and zinc for smelter and comparison communities.

Full table

Zinc concentrations at the four sampling sites were strongly correlated with one another, but lead concentrations among the four sample media were not strongly correlated (data not shown). Principal components analysis was conducted in an effort to combine the data from the four sample media, but the principal component for lead did not factor well. Therefore, the influence of environmental lead and environmental zinc on blood lead was evaluated separately for each of the sample media.

Blood lead measurements ranged from 1.1 to 18.0 g/dl, and 13% (n=29) of the children had blood lead levels above 10.0 g/dl. Blood lead was inversely correlated with age (Pearson's correlation coefficient=-0.43, P<0.001). Geometric means, 5th and 95th percentiles for blood lead by selected variables are presented in Table 2. Most homes in the study areas were built prior to 1950. Residents living in homes built prior to 1910 had higher geometric mean blood lead levels than more recently built homes. However, environmental lead and environmental zinc concentrations were not associated with age of housing, and age of housing was not found to be a factor in subsequent statistical models. There was no difference in geometric mean blood lead levels between residents of the smelter community versus residents of the comparison community (P=0.21, t-test), and area of residence was not found to be a factor in subsequent statistical models.

Table 2 - Geometric mean (5th—95th percentile) of blood lead by demographic variables.

Full table

The dust and soil lead measurements were positively associated with blood lead. After adjusting for age and sex, the partial correlation coefficients for the environmental lead measurements at participants' residences and corresponding blood lead levels ranged from 0.17 to 0.30 for perimeter soil, interior dust, and exterior dust (Table 3). Street dust/soil was not strongly associated with blood lead levels. The correlations presented in Table 3 do not account for the influence of environmental zinc. Environmental zinc measurements at participants' residences were not directly associated with blood lead levels, resulting in partial correlations for each of the environmental media that did not differ statistically from zero.

Table 3 - Pearson's correlation coefficients (r) for lead concentrations in environmental media collected at participants' residences and corresponding blood lead levels.

Full table

The zinc cutpoint values used to designate individuals with high versus low environmental zinc for each sample medium ranged from 459 to 3677 mg/kg. For exterior dust, the zinc cutpoint was 3677 mg/kg, resulting in 45% of participants being assigned to the low zinc group for this medium. For interior dust, street dust/soil, and perimeter soil, the zinc cutpoints (and percentile of participants assigned to the low zinc group) were 804 mg/kg (10%), 1224 mg/kg (33%), and 459 mg/kg (19%), respectively.

Table 4 presents the regression coefficients for environmental lead versus blood lead, separately for the high and low zinc groups. When comparing regressions for the groups with high versus low environmental zinc, there was a 20% to 46% reduction in the parameter estimate of the associations between residential lead measurements and corresponding measures of blood lead (Table 4). The interaction term for the model described above may be calculated by taking the difference of these coefficients.

Table 4 - Differences in regression coefficients () for the effect of each environmental lead measurement on corresponding blood lead levels by high and low environmental zinc levels.

Full table

The same analysis was completed for children 6 years old and younger (n=102). The magnitude of the effect of high versus low environmental zinc was similar, with a 19% to 40% reduction in parameter estimates, and the difference between estimates remained statistically significant for exterior dust and street dust. Since galvanized pipe can be a source of zinc, the above analyses also were repeated after excluding those participants residing in homes with galvanized pipes (n=7), but the results remained unchanged (data not shown).

Discussion

The results of this study indicated that household exterior and interior dust and soil lead measurements were associated with blood lead levels. These associations remained significant in areas with both high and low environmental levels of zinc, but the association was stronger among those participants whose corresponding zinc measurement was low. This interactive effect is also demonstrated with the data for street dust/soil. Before considering the influence of zinc, there was no association between lead in street dust/soil and blood lead. After including the zinc interaction term, there was a modest but significant association between lead in street dust/soil and blood lead for the low zinc group.

The interactions between environmental lead and zinc levels on blood lead concentrations suggest that zinc may influence the absorption of ingested lead. Animal studies of the influence of zinc on intestinal metallothionein and the absorption of metals provide a biologically plausible explanation for our findings. At low-dose levels, the absorption of zinc in rats was mediated by binding to a cystein-rich intestinal protein, but at higher dose levels, zinc induced metallothionein production in intestinal mucosal cells (Hempe and Cousins, 1992). The induction of intestinal metallothionein due to high zinc levels could account for reduced absorption of other divalent cations that bind more strongly to metallothionein. Intestinal segments taken from rats fed low amounts of zinc transferred more copper from a nutrient medium across the mucosal cells than did intestines from rats fed high levels of zinc (Fischer et al., 1981). Further, the copper was sequestered in the mucosal cells by a protein with a molecular weight similar to that of metallothionein (Fischer et al., 1981). A similar mechanism of increased binding of lead to zinc-induced intestinal metallothionein could explain the influence of zinc on blood lead levels.

The primary limitation of this study was the usage of the measured residential environmental media as the only source for exposure to lead and zinc. First, we did not have environmental measurements from other areas such as schools, day-care facilities, or parks where children could come into contact with contaminated soils. Second, we did not have information on children's behaviors or activities with respect to the sampled media. Third, for dust samples we did not have information on load levels at the children's respective residences. For example, lead concentration in a dust sample could represent a relatively clean home with little dust or a home with higher levels of accessible dust. Despite these limitations in exposure assessment, we found strong associations between three of the four media for environmental lead measurements and blood lead levels.

Another potential limitation of this study is the lack of information on dietary factors or nutritional status that could influence both exposure and absorption. Children with deficiencies in nutritional calcium and iron had higher blood lead levels than children who are calcium and iron replete (Ziegler et al., 1978; Mahaffey and Annest, 1986; Mahaffey et al., 1986; Marcus and Schwartz, 1987). Zinc is a common dietary constituent, and an individual's nutritional level of zinc can influence the rate of zinc absorption (Johnson et al., 1988). Our findings suggest that zinc also may play a role in the absorption of lead. Future studies of the influence of zinc on the absorption of lead should include measures of dietary as well as environmental exposure to zinc.

Lead and zinc are among the top 10 substances with a completed exposure pathway found most frequently at NPL sites, and these metals often can be found as co-contaminants at specific sites (ATSDR, 2001). Our findings suggest that zinc may influence the association between environmental measures of lead in dust and soil and corresponding blood lead levels in children. Models such as the IEUBK that have been widely used in estimating blood lead levels based on lead concentrations in environmental media can allow for variation in exposure, uptake, and biokinetic parameters. These models might be further improved by allowing for the inclusion of the concentrations of other metals such as zinc, when appropriate to site-specific conditions.

References

1. Angle C.R., Marcus A., Cheng I.H., and McIntire M.S. Omaha childhood blood lead and environmental lead: a linear total exposure model. Environ Res 1984: 35: 160–170. | PubMed |
2. ATSDR. Technical Assistance to the Pennsylvania Department of Health. Biologic Indicators of Exposure to Cadmium and Lead, Palmerton, Pennsylvania, Part I. US Department of Health and Human Services, Public Health Service, Atlanta, GA, 1984a.
3. ATSDR. Toxicological Profile for Zinc. US Department of Health and Human Services, Public Health Service, Atlanta, GA, 1994b.
4. ATSDR. Toxicological Profile for Lead. US Department of Health and Human Services, Public Health Service, Atlanta, GA, 1999.
5. ATSDR. FY 2000 Profile and Annual Report. US Department of Health and Human Services, Public Health Service, Atlanta, GA, 2001.
6. Baker Jr. E.L., Hayes C.G., Landrigan P.J., Handke J.L., Leger R.T., Housworth W.J., and Harrington J.M. A nationwide survey of heavy metal absorption in children living near primary copper, lead, and zinc smelters. Am J Epidemiol 1977: 106: 261–273.
7. Barltrop D., Strehlow C.D., Thorton I., and , Webb J.S. Significance of high soil lead concentrations for childhood lead burdens. Environ Health Perspect 1974: 7: 75–82.
8. Barnes R.M. Childhood soil ingestion: how much dirt do kids eat? Anal Chem 1990: 62: 1023A–1026A.
9. Calabrese E.J., Stanek III E.J., Pekow P., and , Barnes R.M. Soil ingestion estimates for children residing on a superfund site. Ecotoxicol Environ Safety 1997: 36: 258–268.
10. Casteel S.W., Cowart R.P., Weis C.P., Henningsen G.M., Hoffman E., Brattin W.J. et al. Bioavailability of lead to juvenile swine dosed with soil from the Smuggler Mountain NPL Site of Aspen, Colorado. Fundamental Appl Toxicol 1997: 36: 177–187.
11. Fischer P.W., Giroux A., and L'Abbe M.R. The effect of dietary zinc on intestinal copper absorption. Am J Clin Nutr 1981: 34: 1670–1675.
12. Hempe J.M., and Cousins R.J. Cysteine-rich intestinal protein and intestinal metallothionein: an inverse relationship as a conceptual model for zinc absorption in rats. J Nutr 1992: 122: 89–95. | PubMed | ISI | ChemPort |
13. Johnson P.E., Hunt J.R., and Ralston N.V. The effect of past and current dietary Zn intake on Zn absorption and endogenous excretion in the rat. J Nutr 1988: 118: 1205–1209.
14. Kleinbaum D., Kupper L., and Muller K. Applied Regression Analysis and Other Multiviable Methods, 2nd edn. Duxbury Press, Belmont, CA, 1988.
15. Lewin M.D., Sarasua S., and Jones P.A., A multivariate linear regression model for predicting children's blood lead levels based on soil lead levels: a study at four superfund sites. Environ Res 1999: 81: 52–61.
16. Maddaloni M., Lolacono N., Manton W., Blum C., Drexler J., and Graziano J. Bioavailability of soilborne lead in adults, by stable isotope dilution. Environ Health Perspect 1998: 106 (Suppl 6): 1589–1594.
17. Mahaffey K.R., and Annest J.L. Association of erythrocyte protoporphyrin with blood lead level and iron status in the second National Health and Nutrition Examination Survey 1976–1980. Environ Res 1986: 41: 327–338.
18. Mahaffey K.R., Gartside P.S., and Glueck C.J. Blood lead levels and dietary calcium intake in 1- to 11-year-old children: the Second National Health and Nutrition Examination Survey, 1976 to 1980. Pediatrics 1986: 78: 257–262.
19. Marcus A.H., and Schwartz J. Dose – response curves for erythrocyte protoporphyrin vs blood lead: effects of iron status. Environ Res 1987: 44: 221–227.
20. Miller D.T., Paschal D.C., Gunter E.W., Stroud P.E., and D'Angelo J. Determination of lead in blood using electrothermal atomisation atomic absorption spectrometry with a L'vov platform and matrix modifier. Analyst 1987: 112: 1701–1704. | PubMed |
21. Needleman H.L., and Gatsonis C.A. Low-level lead exposure and the IQ of children. A meta-analysis of modern studies. JAMA 1990: 263: 673–678. | PubMed |
22. Reagan P.L., and Silbergeld E.K. Establishing a health based standard for lead in residential soils. Environ Geochem Health 1989: 12(Suppl) 198–230.
23. Roberts T.M., Hutchinson T.C., Paciga J., Chattopadhyay A., Jervis R.E., VanLoon J. et al. Lead contamination around secondary smelters: estimation of dispersal and accumulation by humans. Science 1974: 186: 1120–1123.
24. Stark A.D., Quah R.F., Meigs J.W., and DeLouise E.R. The relationship of environmental lead to blood-lead levels in children. Environ Res 1982: 27: 372–383.
25. USEPA. Palmerton Zinc Pile Superfund Site Amended Quality Assurance project plan. Region III, US Environmental Protection Agency, Philadelphia, PA, 1991.
26. USEPA. Guidance Manual for the Integrated Exposure Uptake Biokinetic Model for Lead in Children. US Environmental Protection Agency, Washington, DC, 1994.
27. Yankel A.J., von Linder I.H., and Walter S.D., The Silver Valley lead study: the relationship of childhood lead poisoning and environmental exposure. J Air Pollut and Control Associ 1977: 27: 763–767.
28. Ziegler E.E., Edwards B.B., Jensen R.L., Mahaffey K.R., and Fomon S.J. Absorption and retention of lead by infants. Pediatr Res 1978: 12: 29–34. | PubMed |

Acknowledgements

We thank Drs. James Logue and James Fox, Pennsylvania Department of Health for their contributions to the initial study of these communities. Drs. Robert Amler and Jeffrey A. Lybarger with the ATSDR and Drs. Michael McGeehin and Gina Terracciano, formerly with the ATSDR, provided leadership on the larger study from which the data was derived for the present analysis. We also thank Drs. Dave Campagna and Olivia Harris, ATSDR, for their thoughtful review.