Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Living in a sea of lead — changes in blood- and hand-lead of infants living near a smelter

Abstract

Thirteen infants born into the lead contaminated environment of Port Pirie, South Australia, were followed approximately monthly from birth until they were about 36 months. Blood-lead levels of infants at birth were similar to their mothers but fell rapidly during the first 35 days of life. Thereafter, infants born with blood-lead levels at about 2–4 μg/dl began a slow linear increase until 14–18 months where a plateau occurred of 10.8–17.2 μg/dl. The blood-lead levels were well correlated with hand-lead loadings of infant (r2=0.72, P<0.01, log transformed data) and mother (r2=0.62, P<0.01, log transformed data) unless the birth lead level was exceptionally high. The principle factor determining exposure was thee impact of smelter emissions on the house. Blood-lead increase was caused by the relatively more rapid increase in dose of lead compared with the increasing body mass, which was related directly to the maturation of motor development. Hand-lead of mothers were closely related to both infants' blood- and hand-lead levels until the point of blood-lead plateau then substantially fell as infants began to walk unaided. The estimated slope factor using the ICRP model was 0.75–0.94 μg/dl per μg/day with a maximum daily dose of 3–5 μg/kg/day, assuming 45% absorption. Ingestion appears to be the most likely route for at least 95% of the dose.

Introduction

The toxicity of lead is well established, especially its effect on the intellectual development of children. The continued non-occupational exposure to lead by some communities has prompted further studies of the impact of low-level exposure (Lanphear et al., 2005). The understanding of the severity of the consequence of lead exposure has prompted specific preventative measures such as the removal of lead from food containers and from petroleum products used in motor vehicles. The latter has dramatically reduced lead in air concentrations in major urban populations with a subsequent dramatic reduction to lead exposure to those communities (MMWR, 2005). However, despite this public health progress, some communities are still exposed to excessive levels of lead (e.g., Maynard et al., 2003; Gulson et al., 2004). One of these communities is Port Pirie, where the world's largest lead smelter is located.

The dose of lead required to achieve and maintain blood-lead levels in excess of recognizable action values (10 μg/dl) is extremely small compared with the total mass of substances ingested each day by an infant. It is, therefore, understandable in an environment that is contaminated by continuous emissions of metals, including lead, that infants have elevated blood-lead levels. Port Pirie, with a population 13,450 (2001 census), is located in a semi-arid environment 230 km north of Adelaide, South Australia (Figure 1), being established in 1889 and has been exposed to considerable quantities of lead and other metals (Tiller et al., 1975). Although elevated blood-lead levels were thought to be solely due to lead emitted many years earlier (historic lead contamination hypothesis e.g., Body et al., 1988) it has been demonstrated that the contamination of the city is substantial and ongoing (Maynard et al., 2003, 2006).

Figure 1
figure1

Map indicating location of Port Pirie.

Historically, studies have focused on exposure arising from contamination of surfaces, whether objects or body parts such as the hand, which are placed in the mouth. For instance, Bornschein et al. (1985) showed convincing evidence of a strong relationship between surface lead, hand- and blood-lead of infants aged 9–24 months while Lanphear et al. (1998) pooled data from 12 studies and demonstrated the importance of contaminated surfaces of houses in the exposure pathway of children. An analysis of various lead exposure studies suggests the surface–hand–mouth exposure pathway to be common to all studies for children under 72 months (Succop et al., 1998). However, these studies generally investigated children who were mobile and over 12–18 months of age. There are few data for infants less than 6 months. Infant blood-lead levels in Port Pirie indicate significant exposure occurs before they are mobile, which raises the question of exposure mechanisms or routes.

In the neonate (1–28 days old) and young infant (1–6 months) the surface-hand-mouth route is limited by underdeveloped motor control. Nevertheless, infants (1–23 months old) have been shown to have rising blood-lead from birth. For example, Gulson et al. (1999) found some infants' blood-lead began rising at 60 days, but Manton et al. (2000) found little lead on hands of infants. A more recent study (Kranz et al., 2004) demonstrated that infants as young as 5 weeks have contaminated hands of sufficient magnitude that involuntary mouthing of objects may be the principal exposure route. Significantly, young infants can be exposed to sufficient lead to cause elevated blood-lead levels at a very early age, although identifying at which age and at what level an infant is most vulnerable to lead's toxicity is difficult.

Also of interest is the change in the physiological handling of lead as the infant ages, and in particular, whether the neonates' blood-lead levels are more responsive to the dose of lead than older children, which may be important when determining the susceptibility to lead toxicity. Many have investigated the kinetics of lead with the development of well founded models (e.g., Leggett, 1993; US EPA, 1994; O'Flaherty, 1995). These have defined, although not always clearly, differences between young and older infants. For instance, although it is well known that the fraction of lead absorbed from the gut is much higher in the child than adult, the difference between the neonate and infant is less well understood.

Gulson et al. (1999) argues that the reservoir of lead accumulated in bone while in utero is important for determining later blood lead levels, especially in city non-smelter exposed dwellers. This may be correct for blood levels that are relatively low. However, the size of this store relative to the increasing body mass in a rapidly growing neonate should be considered. Port Pirie blood-lead levels of at least half the child population exceed 10 μg/dl raising the question of the significance of the bone store in the overall blood-lead kinetic picture. As concentration of lead in a given reservoir is related to its size it is clear that in a rapidly growing infant, unless mass of lead entering each compartment is increased at least proportionally, the concentration of lead in that compartment cannot be maintained.

This paper presents a cohort study of 13 infants living in Port Pirie with the primary aim of documenting changes in blood-lead levels from birth along with factors that may be influencing those changes. In particular, the study aimed to answer the questions: at what point did the blood-lead begin to rise after birth and what was the shape of the blood-lead curve particularly during the first 18 months of life? During the study, an attempt was made to explore potential exposure pathways, in particular the role of hand-lead loadings. Few have attempted to measure the lead levels of blood and hands of parents and infants concurrently over a prolonged time period. An attempt was also made to estimate the change in dose of lead with age on a body weight basis in order to explore the magnitude of exposure in terms of mass of lead ingested and the impact of the increases in compartment size with age on blood-lead levels.

Methods

Expectant mothers were recruited after families were invited to participate and given details of the project, highlighting the voluntary nature of ongoing participation. During the study, the usual educational program and case management of infants with elevated blood-lead levels, conducted by the Department of Health's Port Pirie “Environmental Health Centre” (EHC), continued without modification. From 1 week post-birth an employee of the EHC visited each mother approximately 4–6 weekly, conducting 16 sessions in all. The interviewer conducted a blood test (infant and mother), handwipe test (session 3 onward: infant, mother and father, if possible) and executed a questionnaire. Following the study infants continued the usual routine blood testing program conducted by the EHC (see Maynard et al., 2003).

Blood-lead tests were carried out by capillary method as described in the Australian Standard, 1983: AS2636–1994. Blood-lead was analyzed using graphic furnace AAS, by Zinifex Pty Ltd (formally Pasminco) laboratory, in accordance with Australian Standards AS2411–1993 and AS4090–1993. The laboratory participates in several national and international reference programs. To ensure integrity of the capillary method of sampling comparison with venous sampling is performed on a routine basis, by taking blood using both methods from the same subject (Maynard et al., 2003).

The infants' handwipe samples were collected by the interviewer with pre-moistened disposable washcloths (Wet Ones®) after firstly washing her own hands thoroughly with a solution of triclosan (pHisoHex®) then drying with a clean disposable towel. This was followed by thoroughly wiping her own hands with a washcloth (Wet Ones®) that was used as an interviewer “blank”. The caregivers (mother and if possible father), after being shown how, thoroughly wiped their own hands with pre-moistened washcloths (Wet Ones®) with the interviewer present. Details of the activities of the subjects immediately prior to the handwipe, focusing on location of activity and hand-washing, were recorded. Washcloths were placed into pre-labeled, acid washed sample jar, sealed, and stored at 5–7°C until analysis, which was conducted by Chemistry Centre WA (East Perth, Western Australia) using ICP-MS (detection limit 1 μg per wipe).

To ensure tubes, washcloths and bags were not contaminated with lead, samples from each batch (the same batch was used throughout the trial) were tested for lead. Further, with each testing episode (infant and parent) a blank tube, washcloth and bag was also tested along with the environmental sample.

Hand-lead data were collected as mass per two hands, however, for comparative purposes all hand-lead data were converted to mass per unit area of the subject's hands (μg/m2). The adults total hand surface area (two hands) was estimated to be 990 and 817 cm2 for males and females, respectively (US EPA, 1997). The infants' hand surface area was based on calculating their body surface area using the formula of Mostellar (in Kemp et al., 1997) and the percentage area the hands contributed to total body surface area from work of Cowan and Conley (1973) and Nagel and Schunk (1997): 2.8 and 2.6% of total surface area for boys and girls, respectively.

A kinetic model was constructed using Mathematica® (version 5.1, Wolfram) based on the ICRP biokinetic model for lead (Leggett, 1993), with solutions to the ordinary differential equations calculated using a Mathematica module provided by Guillermo Sanchez (SysModel, guillerm@usal.es, version 0.8 2002-09-20, http://web.usal.es/~guillerm/). Partition rate constants were scaled according to age, as was weight, which was calculated using the algorithm of O'Flaherty, 1995. Initialization of compartment loads was obtained by simulating a quasi-fetus model of a growing fetus until 280 days where the fetal input was direct to diffusible plasma. The input dose was derived empirically so that the final blood-lead was equivalent to the geometric mean of mother's blood-lead based on whether they lived in the high-risk or the low-risk areas of Port Pirie. Estimation of infant dose was calculated by iteration using least squares method to determine the best model.

A breast milk sample was taken at the second visit using the following method. The mother washed her hands with triclosan solution (pHisoHex®, GlaxoSmithKline Consumer Healthcare) followed by wiping her breast with a clean disposable washcloth rinsed in warm water. Milk was expressed (usually 10 ml) using a breast pump, which was subsequently placed into an appropriately labeled tube. The breast pump was cleaned between uses, using initially warm water and detergent, followed by soaking for 2–4 h in sodium hypochlorite solution (Milton®). The entire apparatus was subsequently rinsed in distilled water, dried, and stored in a plastic container.

At each session, details were collected by questionnaire on: immediate and past location of child (e.g., in bed, on floor, outside), stage of motor development, health and well-being of the infant and mother, the infant's living environment (sleeping, playing details etc), the family (occupation, smoking, pets, house renovation details), and the weight and length of the infant.

Statistical analyses were carried out using SPSS version 10 (SPSS Inc.). Blood- and hand-lead data were normalized using the natural logarithm of the value. Other variables were tested for normality with the Kolmogorov-Smirnov test and transformed if appropriate. Where possible, summary data were compared. The total area under the curve (AUC) for infant blood-lead was calculated using trapezoids, then normalized to μg/dl, giving a summary estimate of the total lead exposure (Mathews et al., 1990). Repeated measures tests were performed using natural log transformed data. Risk area has been defined by others (Vimpani et al., 1985) and is based on the blood-lead levels. The partial eta squared (η2) value was used to determine the degree of association in general linear models, being a measure of effect size. For the purpose of this report, motor development was categorized by the accomplishment of locomotion milestones only.

Results

Blood- and hand-lead

Thirteen mothers were recruited over a period of 5 months who delivered a total of 14 infants, eight boys and six girls. One infant (male) was withdrawn from the study after the second session (1 month) and is not included in these analyses and another male (infant 7) was withdrawn after session 10 (9 months), whose data are included where possible. The residences of the families were scattered across the city although none were more than 4 km from the center of the smelter. The uneven distribution of gender across risk areas (twice as many males lived in the low-risk area than females) made interpretation of gender-based data difficult.

Within 10 days of birth both mothers and babies had their blood-lead levels measured (range 1–26 μg/dl). Infant levels were approximately 83% (95% CI: 67, 100) of their mothers blood-lead level with the mothers level being a good predictor of the infants (F=67.9, P<0.001). The time course for the blood-lead data for mother and infant is summarized in Figure 2, which indicates the strong upward trend from 2 to 3 months. All infants who stayed in the study exceeded 10 μg/dl at some stage, four infants exceeded 15 μg/dl with the highest recorded level being 44 μg/dl (range at end of study 5–26 μg/dl). (The infant who was withdrawn continued having routine yearly blood-lead monitoring and had not exceeded 8 μg/dl by 36 months).

Figure 2
figure2

Blood-lead levels of infants (closed circles) and their mothers (open circles). The intensive phase of the study finished at approximately 580 days of age. Further data collected as part of routine blood monitoring is also reported. Note the differences in the blood-lead axes scale.

Overall the infant blood-lead profiles had a number of common characteristics. Immediately after birth blood-lead levels of the infants fell for 1–2 months to approximately 47% of birth blood-lead level unless they were born with very low levels of 1 to 2 μg/dl. After this initial fall all infants' blood-lead levels rose with age until approximately 12 months old, after which some began to plateau, with most blood-lead levels beginning to fall after 18 months of age. Peak geometric mean blood-lead level for infants living in the high-risk area occurred at 12.4 months (17.2 μg/dl, range 10–43 μg/dl) and low risk at 18.0 months (10.8 μg/dl, range 7–15 μg/dl). Most monthly changes in blood-lead were of the order 1–2 μg/dl, the largest being between sessions 9 and 10 (7.5–8.5 months). No explanation could be found from the questionnaire data for many of these fluctuations. Residential relocation appeared to be important for infant 11. At 9 months this infant, who had very elevated blood-lead levels, jumped by more than 16 μg/dl between the sessions. Relocation, encouraged by the EHC case workers coincided with an equally precipitous fall. However, the family then relocated to their original address a few weeks later with a subsequent rise in blood-lead to 43 μg/dl. The family finally relocated to a small town just outside of Port Pirie with a corresponding fall in blood-lead.

Blood-lead levels of mothers all remained stable rising or falling by only ±1 to 4 μg/dl over the first year except the mother of the infant 1 whose blood-lead fell, after giving birth, from 25 to 14 μg/dl but subsequently rose to 35 μg/dl before testing discontinued. It was considered that the ingestion of rainwater was the cause of the elevated blood-lead level in this individual although a full investigation was not permitted (rainwater is not considered potable in Port Pirie due to lead and other heavy metal contamination). The overall geometric mean blood-lead level of study mothers was 4.7 μg/dl, which was higher than the general Port Pirie pregnant female population's geometric mean of 3.5–3.7 μg/dl.

Hand-lead loadings of infants were measured from session 3 along with their parents, if available. Unlike blood-lead values, hand-lead levels did not rise uniformly with age (range 10–157 μg/m2/month). The median loadings by session appear to fall into four groups, corresponding to the mobility stage of the infant: (1) 2–5 months: 54 μg/m2 (inter-quartile range: 29–129), (2) 6–9 months: 173 μg/m2 (inter-quartile range: 89–481), (3) 10–15 months: 424 μg/m2 (inter-quartile range: 218–1271), and (4) over 15 months: 336 μg/m2 (inter-quartile range: 259–525). There also appeared to be a gender bias with boys having higher loadings (Table 1), which was mirrored in the mother and perhaps the father as well, with parents of boys having higher hand-lead loadings, but given the small sample size no firm conclusions could be drawn. Fathers generally were unavailable for testing (total number of samples: 54, number of subjects tested: 12). Two fathers were measured on nine occasions and one father 10 times. Adult males had the highest hand-lead loadings with one father's hands loaded with 1250 μg of lead per hand (25253 μg/m2). In all, the loadings were quite variable with approximately an order of magnitude between the median and 95th percentile. In summary: approximately 75% of the samples from male adults had <700 μg/m2 (69 μg/hand), 75% of the samples from mothers had <150 μg/m2 (12 μg/hand) and 75% of the samples from infants had <544 μg/m2 (13 μg/hand).

Table 1 Comparison of infants', mother's and father's hand-lead loadings (μg/m2) with inter-quartile range by sex of infant (n=number of individuals tested)

Two summary variables were used in an attempt to characterize blood-lead profiles and the interactions of factors with them. Given the general upward trend of blood-lead levels from about 2 months of age a linear regression model (age as independent variable) was applied from the time of minimum blood-lead to the maximum. This described the infants' blood-lead profile well: adjusted r2 were >0.62, except for infant number 7, with males and those living in high-risk areas tending to have a steeper slope (Table 2). The median slope for females was 0.022 μg/dl/day and for males 0.046 μg/dl/day. By risk regions, the median values were nearly identical (0.022 μg/dl/day for high-risk and 0.027 μg/dl/day for low-risk). As could be expected the (log transformed) AUC was closely related to the slope (adjusted r2=0.69, which was significant at the 0.01 level: P=0.009) but insufficient data prevented a clearer understanding of the relationship with the point at which the blood-lead level began to increase. Unlike slope, infants from the high-risk area appeared to have larger AUC (median 11.5 vs. 7.0 μg/dl for high and low risk, respectively) but this was not statistically significant.

Table 2 Summary of infant characteristics in ascending order of AUC. Slope is from minimum blood-lead to peak. The risk area is a local designator based on the blood-lead levels. Blood-lead and hand-lead are both geometric mean values

Figure 3 indicates a relationship between blood-lead of the infant and the measured lead parameters of the child (hand-loading) and mother (hand- and blood-lead), with a weaker and non-statistically significant relationship with the father (hand-lead), perhaps due to the incomplete data sets in the case of the fathers. Ten infants had significantly positive correlations (P<0.05) between log transformed blood- and hand-lead loading. Approximately 46% of the variance of the infants AUC could be explained by the geometric mean hand-lead loading of the infant (F=26, P<0.001), with 60% being explained by the mother's geometric mean hand-lead loading (F=19, P<0.001) and geometric mean mother's blood-lead explaining 46% of the infants' AUC (F=11, P<0.006). The slope of the rapid rising portion of the blood-lead curve was not related to hand-lead of the child (F=4.39, P=0.060), but weakly related to mothers hand-lead (F=5.71, P=0.036). In all, eight infants had significantly positive correlations between their hand-lead and that of their mother (P<0.05) on regression. The mothers' own geometric hand-lead loadings were related to their own geometric blood-lead (F=11, P=0.006).

Figure 3
figure3

Summary of lead loading on hands of infants and their parents expressed as μg/m2. The figures show the relationship between the loadings of the infant and age (top left, r2=0.32), hand loadings and blood-lead levels (top right, r2=0.70) and the correlations of the infants' hand-lead loadings with there parents (separate graphs bottom, mother r2=0.66; father r2=0.54). The fitted curves are second-order polynomials along with their maximum ranges.

Factors that influenced lead levels

Data were examined both on an individual and group level to investigate the relationship between increments in blood-lead and achievements in developmental stage. The first observed accomplishment of each particular “mobility” stage is summarized in Table 3 along with blood and hand-lead averages. Infants, as expected, progressed through a series of developmental milestones, however, not all were recorded for each infant. Some jumps in development occurred, perhaps due to parents recalling only the most recent event. By session 9, all infants were sitting and some were creeping or crawling with one standing (supported). Thereafter, infants began to walk supported and by session 15 walked unaided.

Table 3 First record of major developmental accomplishments recorded at time of session, with median age (months) and geometric blood- and hand-lead levels

The two largest increments in blood-lead occurred between sessions 7 and 8, and sessions 10 and 11, with the largest increments in hand-lead between sessions 6 and 7 and 12 and 13. Between sessions 7 and 9 more infants achieved more milestone increments than between any other sessions. However, no single advancement (e.g., progression from creeping or rolling-over to crawling) increased the blood-lead any more than the other, although there appeared to be a consistent fall in hand-lead loadings after the infants began to walk unaided. Between session 13 and 16 cohort geometric hand-lead levels fell from 584 to 356 μg/m2 as progressively more infants walked unaided. Blood-lead levels changed little over this period. Overall, in this small sample, no relationship between blood-lead level and accomplishment of a developmental stage could be found.

For the purpose of management, the city of Port Pirie is divided into risk areas designated high- and low-risk, based on the proportion of children with blood-lead ±15 μg/dl. In a repeated measures model “risk area” showed a weak to no influence on blood-lead (F=5.29, P=0.044), with a small partial η2 value of 0.35. This was even weaker for hand-lead (F=0.488, P=0.058). The summary variables AUC, slope or geometric mean hand-lead were not differentiated by risk area. Risk area can be typified by house envelope type. In this study, house envelope (solid or sheet material, often corrugated iron) did not influence blood-lead (F=4.5, P=0.06, η2=0.33) and had no effect on hand-lead.

Owing to the small numbers of each sex, it is difficult to understand the influence of gender on lead variables. Males tended to have a higher, but not significant AUC compared with females, (P=0.086, males: 11.2 μg/dl (range 6.8–25.9), females: 6.46 μg/dl (range 2.4–13.8), which was also reflected in the slope. In a repeated measures model with blood-lead partitioned by gender there was little interaction, (F=7.4, P=0.030) and there was no interaction between gender and risk area (F=0.01). There was also no significant difference in hand-lead loadings by gender, however, the males (infants and fathers) tended to have higher levels in all cases (Table 1).

Eleven infants were breast fed from birth but by the 14th session only three infants were still breast fed, although solid foods had also been introduced. However, there were some differences in breast feeding among the cohort. For example, only one infant was fed more than 3 months in the high-risk area. Overall, infants from the high-risk area were less likely to be breast fed, and if they were, they were breast fed for a much shorter time than those in low-risk (and more affluent) areas (breast feed >40% diet, months fed vs. risk area, F=9.0, P=0.012). All milk had <0.5 μg/dl except that from the mother of infant 1 whose milk had a lead level of 2.8 μg/dl with a corresponding blood-lead level of 25 μg/dl. It was observed that mothers who breast fed tended to have lower hand-lead loadings while breast feeding but was not statistically significant.

Teething infants tended to have elevated hand and blood-lead, but insufficient data prevents further analysis. The number of illness episodes may also be related to blood-lead (repeated measures, P=0.023 η2=0.123) but classification was difficult, especially when differentiating between teething and illness status.

The location of the infant at time of testing was coded into three categories (on floor, outside or other). The median hand loading (all results) for infants on the floor was 333 μg/m2 (n=58), outside 867 μg/m2 (n=11) compared with all other locations: 110 μg/m2 (n=101). Comparison of these data is confounded by age. Infants <5 months of age group were not found on the floor, while often those found outside were all older than 10 months.

A number of other potential modifiers were examined including season of birth, occupation (too diverse to analyze), smoking, keeping of pet (nine households had no pet), bed type (cot, cradle, bassinette), location within house of sleeping, method of clothes drying (external line vs. mechanical dryer): all of which were not be found to be associated with hand- or blood-lead. Hand-lead tended to be higher if the child had washed 15–30 min prior to the visit, but the data is difficult to interpret, due to poor recall of the mother as to the exact time of washing. The potable water supply meets Australian Drinking Water Guidelines for lead (<10 μg/l) with no sample exceeding 2.5 μg/l (pers. comm. SA Water). Locally collected tank rainwater usually contains much higher lead concentrations.

Two of the female babies were twins (ID 8 and 9), born approximately 1000 g lighter than the others and somewhat shorter in length (6 cm shorter). Their blood-lead time series were very similar, never differing by more than 2 μg/dl. Their maximum hand loadings were also similar (1222 and 1416 μg/m2). There were a number of risk factors for the twins: the father worked at the smelter, the residence was on the border of a high-risk area, up to 10 dogs inhabited the property and the mother smoked for at least 10 months of the study. One mitigating factor was the lounge room carpet had been renewed when the infants were about 3 months old. Nevertheless the twins' ranked (in a ranking of all infants in ascending order) fifth and sixth for slope (age vs. blood-lead) and sixth and seventh for AUC and sixth and seventh for geometric hand-lead loading.

Dose of lead

Dose, for the cohort split by risk area, was best described by a model that gave an increasing dose per body mass per day (μg/kg/day) peaking at approximately 12 months then falling away (parameters given in Table 4). Using the dose model to estimate lead uptake (based on ICRP default fraction absorbed of 45%) a slope factor, which included an intercept, was obtained for high- and low-risk areas, respectively: 0.75 and 0.94 μg/dl per μg/day (i.e., 7.6 and 8.9 μg/dl per μg/kg/day). An estimate of the fraction of dose from air was <2% (assuming 50% absorption) based on South Australian EPA (2004) annual ambient lead in air averages of 0.11–0.45 μg/m3 in high-risk areas (unpublished indoor air seven day average concentrations of lead (PM10) has been found to be 0.2–0.3 μg/m3 in occupied houses located in high-risk areas). The plausibility of hand-mouth transfer was also explored. The median loading of 293 μg/m2 corresponding to 7.3 μg total mass (interquartile values 218–1271 μg/m2) was found at 12 months with an average blood-lead level of 12.1 μg/dl. These parameters corresponded to an estimated oral dose of 4 μg/kg/day. Based on these data and assuming an infant of 12 months spends 9.4 h awake (Brinkman et al., 1999) and weighs 9.5 kg (O'Flaherty, 1995) a complete loading and mouthing cycle would need to occur every 108 min. However, since it is not reasonable to assume complete removal of lead nor complete swallowing, it was estimated that if 5% of the lead was transferred and swallowed (10% of hand mouthed and 50% lead swallowed) (i.e., 0.365 μg/event in this example), then a cycle of 11 times per hour (while awake) would be required to give a complete daily dose of lead. Given that fingers are 7–14% of the total hand surface (Cowan and Conley, 1973), it is quite plausible that the cycling rate is achievable merely by mouthing a couple of fingers each event.

Table 4 Parameters for modeled dose where the dose (μg/day) is given for each risk area as:

Discussion

Although increases in blood-lead from early infancy have been recognized for some time (e.g., McMichael et al., 1985; Rabinowitz et al., 1985), in general the focus has been on the age when infants walk, which also coincides with peak blood-lead levels. For neonates and very young infants the mechanism of exposure to lead has been poorly examined. The blood-lead of every infant in this study increased, usually within a month or so of birth with all except one infant exceeding 10 μg/dl. This indicates the extent of the contamination in Port Pirie and the ubiquitous presence of lead in their living environment. The most plausible exposure route for the older child is via the surface-mouth route (e.g., Roels et al., 1980; Lanphear and Roghmann, 1997; Succop et al., 1998), although contamination of milk, food, or via respiration may offer important routes. Work by Kranz et al. (2004) indicates that even in the very young infant the exposed extremities are easily contaminated by lead. The data reported in this study clearly indicate that sufficient lead is being ingested to arrest the fall in blood-lead within 8 weeks of birth and drive the body burden upward, for at least the first 12 months of life.

The most plausible exposure route of these young infants is contamination of hands that are mouthed from a very young age. Breast milk, except from one mother, was too low to be a major contributor of lead dose. Respiration was estimated to be no more than 2% of the dose in these young infants. Given that even large lead particles only need reach the posterior of the nose in order to travel to an absorptive surface, whether lung or gut (Hsu and Swift, 1999), inhalation may be important, especially among infants close to the floor where in-trained large particles may be present. However, given the level of contamination of hands, evidence is strongly supportive of the hand-mouth route as being the major route. An alternate explanation is lead from bone leached into the blood over time, which has been deposited during pregnancy (Gulson et al., 1999). However, modeling using the ICRP model indicated the bones did not have sufficient lead to solely cause the observed peak blood-lead levels. Indeed, in the modeling scenario lead concentrations fell in all blood compartments for the first 20 days as it repartitioned to soft tissues and bone (Figure 4).

Figure 4
figure4

Predicted blood lead levels using ICRP kinetic model (Leggett, 1993) with close model as described in text vs. geometric mean blood-lead levels split by risk area (high-risk, open circles; low-risk, closed circles).

One question of interest, because it might lead to an understanding of why the blood-lead levels rose, is: why did the blood-lead level fall for 30–90 days from birth? The fall could be due to clearance from the blood compartment exceeding dose. With time dose increases sufficiently to overcome this with a resulting rise in blood-lead. Factors that influence the clearance of lead from the blood and the body as a whole are also not entirely clear. Many of the differences lay in the handling of lead by bone, with the bone uptake from blood being much greater in the growing infant due to the rapid bone turnover and addition of calcium to bone (Leggett, 1993). Other factors such as the changing affinity of lead for binding sites in erythrocytes may also alter lead disposition in the infant. Lead binds to a number of components within the erythrocyte with fetal hemoglobin having a higher affinity for lead than adult hemoglobin (Ong and Lee, 1980). This suggests that the erythrocytes of neonates should be able to store more lead than infants. In addition hematocrit is somewhat higher in the neonate (at birth 51% falling to about 35% at 6 months (Avery et al., 1994)). It is hypothesized that the total binding capacity of blood probably decreases over the first 6 months leading to a redistribution of lead among other tissues including the eliminating organs: the kidney and liver. Other factors such as changes in renal and hepatic function along with fluid redistribution may also have contributed to the fall in blood-lead.

For the blood-lead level to rise there needs to be sufficient dose to cause accumulation. Given the baby is growing rapidly, this means the dose needs to increase at a rate that overcomes increases in volume of organs, altered kidney and liver function and changes to the gut. Hence, the dose per mass of infant needs to rise very steeply in the first 12 months as the infant weight nearly triples. The mechanism for the rise in dose appears related to motor development. As the child becomes more mobile there is an increased chance of body surfaces becoming contaminated along with an increased efficiency in mouthing behavior. The latter point is speculative: observation of neonates and young infants suggest that as fine motor movement is developed the intensity of the mouthing increases, however, it has been found that frequency of mouthing vary little in the first 2 years of life (Tulve et al., 2002; Black et al., 2005). The plateau in blood-lead is brought about by a reduction in dose per unit body mass. Data from this study indicates a lower hand-lead loading in the older infant, and clearly other dose-delivery activities such as mouthing tend to be attenuated in the older infant (Brinkman et al., 1999; Black et al., 2005). However, data from most studies indicate a wide variability in mouthing with no great attenuation between 12 and 18 months corresponding to the plateau for blood-lead levels. This suggests a combination of events that reduces dose delivery. Additional modeling of data from this study found that if weight did not increase from 12 months then blood-lead levels would be 28–39% higher than found. This highlights the importance of weight gain in preventing blood-lead levels rising more rapidly than they did, an aspect that is perhaps overlooked.

One objective of the study was to investigate if one development stage was more important than another in determining the extent of the exposure. The data presented here shows there is clearly a continuum of exposure that increases with age, reflected by a steady change in blood-lead levels. A significant finding was the plateau that occurs at walking. It is hypothesized that hands of young infants are in contact with surfaces such as carpet the majority of the time while crawling or creeping. Once walking unaided there are fewer contacts with heavily contaminated surfaces, and therefore a reduction of the ingested dose occurs. Of course, mouthing still occurs and with the extended range of the infant, once walking, the potential for heavy exposure by other scenarios such as outdoor dust and soil ingestion is increased. This indicates that prevention of contamination of hands is important, and as suggested by others, the removal of contaminated surfaces provides the best amelioration strategy (e.g., Succop et al., 1998). Further to this, Succop et al. (1998) stated that a reduction in hand-lead from 10 to 1 μg per two hands would produce a 14 μg/dl fall in blood-lead level. Using infant data presented in this paper a fall in hand-lead from 10 to 1 μg would produce approximately a 6.2 μg/dl fall in blood-lead (using geometric mean blood- and hand-lead). These data clearly indicate the difference between Port Pirie children and those in houses not being contaminated from active smelter sources. It is interesting to note that Succop excluded Cincinnati (USA) soil lead study and Trail (Canada) data as outliers.

In our study, infant hand-lead loadings were similar to literature values where gross environmental contamination had occurred (e.g., Bornschein et al., 1986; Hilts et al., 1995; Manton et al., 2000). The limitation of the study was that hand-lead loadings were not taken in the first three sessions as it was believed at the beginning of the study hand loadings would be insignificant. Our study, like others, showed hand-lead loadings are indicative of the environment's lead loading (e.g., Davies et al., 1990). Estimating exposure by taking multiple media samples can be complex but hand-loading offers a method that gives a reasonable estimate of exposure, being an integrator. Four factors support this: (1) the relationship between blood- and hand-lead, (2) the relationship between the hand-loading and location of infant at time of test, (3) the close relationship between the mother and the child and, (4) the extreme narrow range found in the twins.

Infants had higher hand loadings than their mothers, but not their fathers, noting that the latter were not routinely measured. This suggests either a concentrating effect by the infant or the existence of a concentration gradient in the house, with the infant residing in the higher end of the gradient and the mother in the lower extremity. We have no evidence to prove or refute the former hypothesis. In the Cincinnati study, a gradient existed from outside to the midst of the dwelling (Bornschein et al., 1986), which was proven in Port Pirie homes by Kutlaca (1998) who found gradients within rooms, in particular between surfaces proximal to windows and doors and those distal from these openings. This study and that by Kranz et al. (2004) suggests a close paralleling between the mother's and the infant's environment. Intuitively this is probably correct given the infant's close association with the mother, especially in the first few months of life. The fact the two appear to reside in the same “lead” environment, albeit at different levels, is further demonstrated by the good correlations between AUC of the infant and mothers hand-lead loadings. The correlation between the infant and mothers hand-lead was good at each age, but the study was limited by its small size to explore this aspect further. Although there were fewer adult male data, those collected indicated a distinctly different environment. This is not surprising considering the adult males in this study were not principal child carers and usually spent considerable time outside the home.

The twins seemed to load their hands to the same extent, irrespective of variations in the conditions: for example, the exact time the last mouthing incident had taken place, the degree of hand moisture, the exact location of play etc. Hence, hand-lead loadings, although mere points in time, appear to be relatively robust in determining the exposure potential. These findings suggest that hand-loadings alone may be useful in determining risk, and may be a useful measure prior to the birth of the infant, so that appropriate strategies could be put into place to remediate or at least attenuate the exposure risk. A better understanding of the individual's variations in loadings is required before this could be put into practice.

The simple model used to test the plausibility of hand-mouth activity being sufficient to dose the infant appears useful and an important step in any mechanistic explanation for the exposure. In the scenario presented in this study, respiration was found not to be a major route, although further in-house measurements of lead in air are needed because gradients may exist, especially from resuspension (Ferro et al., 2004). The data lacking in the model is the actual mass of lead transferred from a hand to the gut, given a certain loading, along with a method to quantify ingestion exposures of children (Cohen Hubal et al., 2000). Although there are data on mouthing rates, the surface area actually mouthed in each event along with the transfer efficiency is difficult to measure. Our model indicates the required mouthing rate of approximately 11 times per hour, with a maximum of two fingers per event, is well within the range determined by Tulve et al. (2002). There are also other surfaces mouthed by infants that have somewhat different lead loadings from the hands, each contributing to the dose.

It is often assumed that the dose of lead rises by small increments when in fact the dose may be highly variable. The normal regimen may consist of a few large doses interspersed amongst a number of lower doses. For instance, if the hands were loaded near the 90 percentile level once daily (e.g., 1825 μg/m2 for a 12 month child in this study) a quarter to half of the daily dose could be delivered in one event. If this was the case, identifying reasons for the extreme loadings may be more important than reducing the ambient lead levels in the infant's environment.

This study was too limited to identify all of the factors that influence exposure. Within the power of the study three major predictors of blood-lead were identified: age of infant, hand-lead loading of mother or infant, and location of residence. Also mothers with high pre-natal blood-lead levels usually had infants that developed high blood-lead levels. The data also shows that, in essence, the increments in dose of lead are related to maturation of the motor system (perhaps by increased efficiency of contaminating dermal surfaces and the efficiency of mouthing). Further, the study indicates the importance of measuring lead levels at an early age, if not prior to birth, to enable implementation of strategies to ameliorate exposure. To this end the Port Pirie Lead Implementation Program now tests mothers ante- and postnatally and infants at age 6 and 12 months then near birthdays and if blood-lead has exceeded 10 μg/dl at any time, at additional intervals, with case-management of clients if blood-lead exceeds 6 μg/dl at 6 months of age.

Proximity to the smelter, as explained by others (Roels et al., 1980; Calder et al., 1994; Baghurst et al., 1998), is the major determinant of the level of exposure to lead, at least in Port Pirie. There is also a possible interaction between house type and location, with older, less well-constructed houses being in the highest risk areas (being older suburbs). The consequence of the existence of high-risk areas can be dramatically demonstrated, as was shown when infant 11 was relocated. These data and others from the Port Pirie Lead Implementation Program indicate that relocation of infants is an important strategy in recalcitrant cases, although prohibitively expensive for all children. Location within the residence is also a contributing factor with infants found on the floor or outside the house having the highest hand loadings.

What needs to be determined are the factors that contribute to the ongoing contamination of the environment of the infant, especially if houses are remediated, so that recontamination can prevented. The Port Pirie Lead Implementation Program data indicate large reservoirs of lead in some houses, which are difficult to extract and keep clean. As ongoing emissions from the smelter continually adds to the environmental load they have become the main priority in the overall effort to reduce exposure to lead.

Conclusions

Our data demonstrates that for infants living in Port Pirie, exposure to lead increments from birth, demonstrated by a steady increase in their blood-lead. The data suggests the notion of a sea of lead in which there are gradients from which infants, even perinatally, cannot escape, especially when considering the minuteness of the dose required. Findings from this study indicate that the principal intervention strategy must be the reduction in ongoing contamination of the household environment followed by a comprehensive house remediation program. Further investigation of reservoirs of dust within homes is needed with a special regard in finding methods of both removing them and sustainably preventing recontamination.

References

  1. Australian Standard 2411-1993. Venous Blood — Determination of Lead Content — Flame Atomic Absorption Spectrometric Method. Standards Association of Australia, Sydney, 1993.

  2. Australian Standard 2636-1983. Sampling of Venous and Capillary Blood for the Determination of Lead Content. Standards Association of Australia, Sydney, 1983.

  3. Australian Standard 4090-1993. Whole Blood — Determination of Lead Content — Graphite Furnace Atomic Absorption Spectrometric Method. Standards Association of Australia, Sydney, 1993.

  4. Avery G.B., Fletcher M.A., and MacDonald M.G. Neonatology, Pathophysiology and Management of the Newborn, 4th edn. Lippincott, Philadelphia, 1994.

    Google Scholar 

  5. Baghurst P.A., McMichael A.J., Vimpani G., Robertson E.F., Clark P.D., and Wigg N.R. Determinants of blood lead concentrations of pregnant women living in Port Pirie and surrounding areas. Med J Aust 1998: 146: 69–73.

    Google Scholar 

  6. Black K., Shalat S.L., Freeman N.C.G., Jimenez M., Donnelly K.C., and Calvin J.A. Children's mouthing and food-handling behavior in an agricultural community on the US/Mexico border. J Exp Anal Environ Epidemiol 2005: 15: 244–251.

    CAS  Article  Google Scholar 

  7. Body P.E., Inglis G.R., and Mulcahy D.E. Lead Contamination in Port Pirie, South Australia: A Review of Environmental Lead Which Could Give Rise to an Increase in Children's Blood Lead Levels. Department of Environment and Planning, South Australia, Adelaide, 1988.

    Google Scholar 

  8. Bornschein R.L., Succop P., Dietrich K.N., Clark C.S., Que Hee S., and Hammond P.B. The influence of social and environmental factors on dust lead, hand lead and blood lead levels in young children. Environ Res 1985: 38: 108–118.

    CAS  Article  Google Scholar 

  9. Bornschein R.L., Succop P., Kraft K.M., Clark C.S., Peace B., and Hammond P.B. Exterior surface dust lead, interior house dust lead and childhood lead exposure in an urban environment. Hemphill, D. D. Trace Subst Environl Health 1986: 2: 322–332.

    Google Scholar 

  10. Brinkman S., Gialamas A., Jones L., Edwards P., and Maynard E.J. Child Activity Patterns for Environmental Exposure Assessment in the Home. National Environmental Health Forum, Adelaide, 1999.

    Google Scholar 

  11. Calder I.C., Maynard E.J., and Heyworth J.S. Port Pirie lead abatement program, 1992. Environ Geochem Health 1994: 16: 137–145.

    CAS  Article  Google Scholar 

  12. Cohen Hubal E.A., Sheldon L.S., Burke J.M., McCurdy T.R., Berry M.R., Rigas M.L., Zartarian V.G., and Freeman N.C.G. Children's exposure assessment: a review of factors influencing children's exposure, and the data available to characterize and assess that exposure. Environ Health Perspect 2000: 108: 475–486.

    CAS  Article  Google Scholar 

  13. Cowan R.J., and Conley M.M. Surface area of the hand and digits. Can J Surg 1973: 16: 187–194.

    CAS  PubMed  Google Scholar 

  14. Davies D.J.A., Thornton I., Watt J.M., Culbard E.B., Harvey P.G., Delves H.T., Scherlock J.C., Smart G.A., Thomas J.F.A., and Quinn M.J. Lead intake and blood lead in two-year-old UK urban children. Sci Total Environ 1990: 90: 13–29.

    CAS  Article  Google Scholar 

  15. Ferro A.R., Kopperud R.J., and Hildemann L.M. Elevated personal exposure to particulate matter from human activities in a residence. J Exp Anal Environ Epidemiol 2004: 14: S34–S40.

    CAS  Article  Google Scholar 

  16. Gulson B.L., Gray B., Mahaffey K.R., Jameson C.W., Mizon K.J., Patison N., and Korsch M. Comparison of the rates of exchange of lead in the blood of newly born infants and their mothers with lead from their current environment. J Lab Clin Med 1999: 133: 171–178.

    CAS  Article  Google Scholar 

  17. Gulson B.L., Mizon K.J., Davis J.D., Palmer J.M., and Vimpani G. Identification of sources of lead in children in a primary zinc-lead smelter environment. Environ Health Perspect 2004: 112: 52–60.

    CAS  Article  Google Scholar 

  18. Hilts S.R., Hertzman C., and Marion S.A. A controlled trial of the effect of HEPA vacuuming on childhood lead exposure. Can J Public Health 1995: 86: 345–350.

    CAS  PubMed  Google Scholar 

  19. Hsu D.J., and Swift D.J. The measurements of human inhalability of ultralarge aerosols in calm air using manikins. J Aerosol Sci 1999: 30: 1331–1343.

    CAS  Article  Google Scholar 

  20. Kemp C.A., McDowell J.M., Lilley B.J., Munari G.A., O'Sullivan C.E., South S., and Tibballs J. Paediatric Pharmacopoeia, 12th ed. Pharmacy Dept., Royal Children's Hospital, Melbourne, 1997.

    Google Scholar 

  21. Kranz B., Simon D.L., and Leonardi B. The behavior and routes of lead exposure in pregrasping infants. J Exp Anal Environ Epidemiol 2004: 14: 300–311.

    CAS  Article  Google Scholar 

  22. Kutlaca A. Mechanisms of Entry of Lead-Bearing Dusts into Houses in Port Pirie. Ph.D. Thesis The Mawson Graduate Centre for Environmental Studies, University of Adelaide, 1998.

    Google Scholar 

  23. Lanphear B.P., and Roghmann K. Pathways of lead exposure in urban children. Environ Res 1997: 74: 67–73.

    CAS  Article  Google Scholar 

  24. Lanphear B.P., Hornung R., Khoury J., Yolton K., Baghurst P., Bellinger D.C., Canfield R.L., Dietrich K.N., Bornschein R.L., Greene T., Rothenberg S.J., Needleman H.L., Schnaas L., Wasserman G., Graziano J., and Roberts R. Low-level environmental lead exposure and children's intellectual function: an international pooled analysis. Environ Health Perspect 2005: 113: 894–899.

    CAS  Article  Google Scholar 

  25. Lanphear B.P., Matte T.D., Rogers J., Clickner R.P., Dietz B., Bornschein R.L., Succop P., Mahaffey K.R., Dixon S., Galke W., Rabinowitz M.B., Farfel M., Rohde C.A., Schwartz J., Ashley P., and Jacobs D.E. The contribution of lead-contamination house dust and residential soil to children's blood lead levels. Environ Res 1998: 79: 51–68.

    CAS  Article  Google Scholar 

  26. Leggett R.W. An age-specific kinetic model of lead metabolism in humans. Environ Health Perspect 1993: 101: 598–616.

    CAS  Article  Google Scholar 

  27. Manton W.I., Angle C.R., Stanek K.L., Reese Y.R., and Keuhnemann T.J. Acquisition and retention of lead by young children. Environ Res 2000: 82: 60–80.

    CAS  Article  Google Scholar 

  28. Mathews J.N.S., Altman D.G., Campbell M.J., and Royston P. Analysis of serial measurements in medical research. BMJ 1990: 300: 230–235.

    Article  Google Scholar 

  29. Maynard E.J., Franks L.J., and Malcolm M.S. The Port Pirie Lead Implementation Program. Future Focus and Directions. Department of Health, Adelaide, South Australia, 2006.

    Google Scholar 

  30. Maynard E.J., Thomas R., Simon D., Phipps C., Ward C., and Calder I. An evaluation of recent blood lead levels in Port Pirie, South Australia. Sci Total Environ 2003: 303: 25–33.

    CAS  Article  Google Scholar 

  31. McMichael A.J., Baghurst P.A., Robertson E.F., Vimpani G., and Wigg N.R. The Port Pirie cohort study. Blood lead concentrations in early childhood. Med J Aus 1985: 143: 499–503.

    CAS  Google Scholar 

  32. MMWR. Blood Lead Levels — United States, 1999–2002. MMWR 2005: 54: 513–516: 1997.

  33. Nagel T.R., and Schunk J.E. Using the hand to estimate the surface area of a burn in children. Pediatr Emerg Care 1997: 13: 254–255.

    CAS  Article  Google Scholar 

  34. O'Flaherty E.J. Physiologically based models for bone-seeking elements. V. Lead absorption and disposition in childhood. Toxicol Appl Pharmacol 1995: 131: 297–308.

    CAS  Article  Google Scholar 

  35. Ong C.N., and Lee W.R. High affinity of lead for fetal haemoglobin. Br J Ind Med 1980: 37: 292–298.

    CAS  PubMed  PubMed Central  Google Scholar 

  36. Rabinowitz M.B., Leviton A., Needleman H., Bellinger D., and Waternaux C. Environmental correlates of infant blood lead levels. Environ Res 1985: 38: 96–107.

    CAS  Article  Google Scholar 

  37. Roels H.A., Buchett J., Lauwerys R.R., Bruaux P., Claeys-Thoreau F., LaFontaine A., and Verduyn G. Exposure to lead by the oral and the pulmonary routes of children living in the vicinity of a primary lead smelter. Environ Res 1980: 22: 81–94.

    CAS  Article  Google Scholar 

  38. Succop P., Bornschein R.L., Brown K., and Tsung C.Y. An empirical comparison of lead exposure pathway models. Environ Health Perspect 1998: 106(Suppl 6): 1577–1583.

    CAS  Article  Google Scholar 

  39. Tiller K.G., Merry R.H., Cartwright B., and Bartlett N.R. Dispersal of lead emissions from an isolated lead smelter within an agricultural region of South Australia. Search 1975: 6: 437–439.

    Google Scholar 

  40. Tulve N.S., Suggs J.C., McCurdy T., Cohen Hubal E.A., and Moya J. Frequency of mouthing behavior in young children. J Exp Anal Environ Epidemiol 2002: 12: 259–264.

    Article  Google Scholar 

  41. US EPA 1994. Technical Support Document for the Integrated Exposure Uptake Biokinetic Model for Lead in Children (v.099d) Technical Review Group for Lead (US EPA) NTIS #PB94-963505, EPA 9285.7–22, 1994.

  42. US EPA Exposure Factors Handbook: Volume 1- General Factors. US EPA, Washington, 1997.

    Google Scholar 

  43. Vimpani G., McMichael A.J., Robertson E.F., and Wigg N.R. The Port Pirie Study: A prospective study of pregnancy outcome and early childhood growth and development in a lead-exposed community — protocol and status report. Environ Res 1985: 38: 19–23.

    CAS  Article  Google Scholar 

Download references

Acknowledgements

Early work on this study was undertaken by Sally Brinkman (formally with Port Pirie Environmental Health Centre, Public Health, Department of Health: South Australia) and Raylene Thomas (Public Health, Department of Health: South Australia). The blood-lead sampling and interviews were conducted by Raylene Thomas and Charmayne Jebb (Public Health, Department of Health: South Australia). Data entry was conducted by Charmayne Jebb and Kathy Thomas (Public Health, Department of Health: South Australia). Analytical expertise for the testing of hand-lead loadings was provided by Mike van Alphen (formally with Port Pirie Environmental Health Centre, Public Health, Department of Health: South Australia) with Angela Gialamas (formally with Port Pirie Environmental Health Centre, Public Health, Department of Health: South Australia) and Anna Swanepoel (formally with Port Pirie Environmental Health Centre, Public Health, Department of Health: South Australia) preparing them for analysis.

Author information

Affiliations

Authors

Corresponding author

Correspondence to David L Simon.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Simon, D., Maynard, E. & Thomas, K. Living in a sea of lead — changes in blood- and hand-lead of infants living near a smelter. J Expo Sci Environ Epidemiol 17, 248–259 (2007). https://doi.org/10.1038/sj.jes.7500512

Download citation

Keywords

  • lead exposure
  • blood-lead
  • hand-lead
  • infants

Further reading

Search

Quick links