SEARCH

SEARCH BY CITATION

Keywords:

  • environmental issues;
  • industrial chemicals;
  • maternal-fetal interactions;
  • metals

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EPIDEMIOLOGY OF LEAD EXPOSURE IN PREGNANT WOMEN
  5. LEAD EXPOSURE AND ADVERSE PREGNANCY EVENTS AND OUTCOMES
  6. KINETICS OF LEAD DURING PREGNANCY AND LACTATION
  7. CALCIUM INTAKE, CALCIUM SUPPLEMENTATION, AND MOBILIZATION OF LEAD FROM BONE
  8. POPULATION SUBGROUPS AT INCREASED RISK
  9. CONCLUSION
  10. REFERENCES

This review focuses on the impacts of lead exposure on reproductive health and outcomes. High levels of paternal lead exposure (>40 μg/dl or >25 μg/dl for a period of years) appear to reduce fertility and to increase the risks of spontaneous abortion and reduced fetal growth (preterm delivery, low birth weight). Maternal blood lead levels of approximately 10 μg/dl have been linked to increased risks of pregnancy hypertension, spontaneous abortion, and reduced offspring neurobehavioral development. Somewhat higher maternal lead levels have been linked to reduced fetal growth. Some studies suggest a link between increased parental lead exposure and congenital malformations, although considerable uncertainty remains regarding the specific malformations and the dose-response relationships. Common methodological weaknesses of studies include potential exposure misclassifications due to the frequent unavailability of exposure biomarker measurements at biologically appropriate times and uncertainty regarding the best exposure biomarker(s) for the various outcomes. A special concern with regard to the pregnant woman is the possibility that a fetus might be exposed to lead mobilized from bone stores as a result of pregnancy-related metabolic changes, making fetal lead exposure the result of exposure to exogenous lead during pregnancy and exposure to endogenous lead accumulated by the woman prior to pregnancy. By reducing bone resorption, increased calcium intake during the second half of pregnancy might reduce the mobilization of lead from bone compartments, even at low blood lead levels. Subgroups of women who incurred substantial exposures to lead prior to pregnancy should be considered to be at increased risk. Birth Defects Research (Part A), 2005. © 2005 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EPIDEMIOLOGY OF LEAD EXPOSURE IN PREGNANT WOMEN
  5. LEAD EXPOSURE AND ADVERSE PREGNANCY EVENTS AND OUTCOMES
  6. KINETICS OF LEAD DURING PREGNANCY AND LACTATION
  7. CALCIUM INTAKE, CALCIUM SUPPLEMENTATION, AND MOBILIZATION OF LEAD FROM BONE
  8. POPULATION SUBGROUPS AT INCREASED RISK
  9. CONCLUSION
  10. REFERENCES

Despite the long history of investigation into lead toxicity, progress continues to be made in understanding its mechanisms and manifestations, including its influences on the course and outcomes of pregnancy. Updating a previous review (Bellinger,1994), this article addresses the following topics: 1) the current epidemiology of blood lead levels among U.S. women, 2) adverse pregnancy events and outcomes associated with higher lead burdens, 3) the kinetics of lead during pregnancy and lactation, specifically the mobilization of lead from bone stores, 4) the role of calcium supplementation in reducing the mobilization of bone lead, and 5) subgroups of pregnant women who should be considered to be at increased risk.

EPIDEMIOLOGY OF LEAD EXPOSURE IN PREGNANT WOMEN

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EPIDEMIOLOGY OF LEAD EXPOSURE IN PREGNANT WOMEN
  5. LEAD EXPOSURE AND ADVERSE PREGNANCY EVENTS AND OUTCOMES
  6. KINETICS OF LEAD DURING PREGNANCY AND LACTATION
  7. CALCIUM INTAKE, CALCIUM SUPPLEMENTATION, AND MOBILIZATION OF LEAD FROM BONE
  8. POPULATION SUBGROUPS AT INCREASED RISK
  9. CONCLUSION
  10. REFERENCES

In the National Health and Nutrition Examination Survey (NHANES), 1999–2000 survey, the geometric mean blood lead level among females older than 1 year of age (n = 4057) was 1.37 μg/dl, with a 95th percentile of 4 μg/dl (U.S. Centers for Disease Control and Prevention [CDC,2003]). This represents a remarkable reduction from the levels measured two decades ago in NHANES II, when the mean levels of women between 20 and 50 years old ranged from 10 to 13 μg/dl (Mahaffey et al.,1983). Crocetti et al. (1990) estimated, based on 1984 prevalence data, that 9% of white women and 20% of African-American women 20–40 years old had a blood lead level of 10 μg/dl or greater. The overall percentage had dropped to 0.5% by 1988–1991, when NHANES II, phase 1 was conducted (Brody et al.,1994).

Elevated exposures still occasionally occur among pregnant women, however. In a recent 3-year period, 7 severely lead-poisoned pregnant women (mean initial blood lead of 72 μg/dl) were seen in a pediatric environmental health specialty unit in Boston (Shannon,2003). Fletcher et al. (1999) surveyed 135 of 197 18- to 45-year-old women in New York State who had a blood lead level between 10 and 25 μg/dl. The primary source of exposure was judged to be occupational in 46% of cases and home renovation in 24% of cases. In the remaining 30% of cases, no current source of exposure could be identified.

LEAD EXPOSURE AND ADVERSE PREGNANCY EVENTS AND OUTCOMES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EPIDEMIOLOGY OF LEAD EXPOSURE IN PREGNANT WOMEN
  5. LEAD EXPOSURE AND ADVERSE PREGNANCY EVENTS AND OUTCOMES
  6. KINETICS OF LEAD DURING PREGNANCY AND LACTATION
  7. CALCIUM INTAKE, CALCIUM SUPPLEMENTATION, AND MOBILIZATION OF LEAD FROM BONE
  8. POPULATION SUBGROUPS AT INCREASED RISK
  9. CONCLUSION
  10. REFERENCES

Male Fertility

Although most concern regarding the effects of lead on pregnancy focuses on women's exposures, several recent studies have evaluated the contribution of male exposures to fertility. Some were conducted by linking blood lead registries assembled by occupational health authorities and population data registries of medically diagnosed pregnancies, whereas others included collection of data on pregnancy outcome by means of interview or questionnaire.

Likelihood of conception/number of live births.

Linking the Heavy Metal Registry of New York State with birth certificates reported to the New York State Vital Statistics Department for the period 1981–1992, Lin et al. (1996) evaluated the association between blood lead level and number of children among 4256 men who had had at least one blood lead level >25 μg/dl and a referent group of 5148 men frequency matched for age and residence. Overall, the group exposed to lead had fewer births than expected (standardized fertility ratio of 0.88; 95% confidence interval [CI], 0.81–0.95). Workers' exposure was classified by blood lead level (high ≥50, medium 35–49, low 25–34) and duration (>5 years, 1–5 years, ≤1 year). Additional analyses revealed that the standardized fertility ratio was close to 1.0 among men with cumulative exposure (mean blood lead level × duration of exposure) below the 85th percentile. A dose-dependent decline in ratio was observed at levels above the 85th percentile. Workers with exposure of at least 5 years in duration were largely responsible for the decreased fertility rate. In this group, the ratio was 0.3 (95% CI, 0.2–0.5).

Bonde and Kolstad (1997) identified 1349 men who worked in any of 3 battery plants in Denmark during the period 1964–1992. Blood lead levels were available only for 400 workers in 1 of the plants during a portion of this period. The mean of all samples was 36 μg/dl (SD = 15). The national population register was used to identify number of live-born children sired by each man during the period 1968–1992 (when the man was 20–49 years old) compared to the number sired by 9656 men who worked in industries that did not involve lead exposure. The findings were essentially null, as the birth rate in the lead-exposed workers during the period of lead exposure did not differ from the birth rate of the same men when they were not being exposed. Similarly, the birth rate during the period of lead exposure did not differ from the birth rate in the referent group. Finally, the birth rate did not change as a function of duration of employment.

Using a nationwide database on medically diagnosed pregnancies treated in Finnish hospitals during the period 1973 to 1983, Sallmen et al. (2000a) evaluated the likelihood of conception in couples in which the male had been biologically monitored for lead exposure at the Finnish Institute of Occupational Health. Men's exposure was classified probable (n = 2111) or potential (n = 664) based on measured blood lead level and its timing in relation to marriage. Five categories were defined for the group with probable exposure: 10–20, 21–30, 31–40, 41–50, and >50 μg/dl. For men with possible exposure, 2 groups were defined: <20 and ≥20 μg/dl. The success rate for pregnancy was compared to that in a referent group of 681 men whose mean blood lead was <10 μg/dl. The risk of failing to achieve a pregnancy was significantly higher in all strata of the probable exposure group: 10–20 μg/dl: 1.27 (95% CI, 1.08–1.51); 21–30: 1.35 (95% CI, 1.12–1.63); 31–40: 1.37 (95% CI, 1.08–1.72); 41–50: 1.50 (95% CI, 1.08–2.02); >50: 1.90 (95% CI, 1.30–2.59). No clear association was found between lead exposure and reduced fertility when the analyses were restricted to couples who had achieved a pregnancy, suggesting that exposure increases the risk of childlessness rather than delays pregnancy among fertile couples.

In a study of 365 men occupationally exposed to metals, Gennart et al. (1992) identified 74 who were exposed to lead continuously for >1 year and who had at least 1 blood lead level >20 μg/dl (mean 46.3, SD 11.2, range 24–75). In the control group, the mean blood lead level was 10.4 (SD 3.3, range 4.4–19). The probability of a live birth was significantly lower among the exposed workers than among the controls. Within the exposed group, the probability was significantly lower after the onset of exposure than in the period prior to the onset of exposure (0.43; 95% CI, 0.25–0.73). Fertility tended to decrease with increasing duration of exposure but was not significantly associated with blood lead level so that the blood lead level at which fertility begins to decrease could not be identified.

Time to pregnancy (TTP).

Several studies evaluated the time needed to conceive in relation to male lead exposure. Apostoli et al. (2000) assembled a group of male workers in lead-related factories in Italy who had sired at least 1 child. Using records of the worker monitoring system, they identified, for each man, the blood lead level closest to the time at which he and his partner began attempting to conceive. They identified 251 men exposed to lead in the course of their jobs and compared them to a referent group of 45 men who were not. Survival analysis (Kaplan-Meier statistic) was used to model TTP, the interval between starting unprotected intercourse and the last menstrual period before the index pregnancy. Contrary to expectation, the exposed group had a significantly shorter TTP than the unexposed group, as well as a greater mean number of children. By limiting the analyses to selected subgroups of the study sample, the authors did find some associations consistent with the hypothesis that lead reduces male fertility. For instance, among the lead-exposed men, a higher blood lead level was associated with a longer TTP, although this association was not statistically significant (p = 0.23). Among men who had sired only 1 child, TTP was significantly longer for those with blood lead levels >40 μg/dl.

The data reported by Apostoli et al. (2000) were included in a larger cross-sectional study of TTP, conducted in four European countries, of 1104 male workers, 638 of whom were exposed to lead (Joffe et al.,2003). The mean blood lead of exposed workers was approximately 30 μg/dl in each country. An internal control group was assembled, consisting of lead workers for whom the timing of exposure could not have affected fertility. Only pregnancies that concluded in live births were included in analyses. No consistent increase in TTP was found in association with blood lead level. The hazard ratios for exposure strata were: <20 μg/dl: 1.12 (95% CI, 0.84–1.49); 20–29 μg/dl: 0.96 (95% CI, 0.77–1.19); 30–39 μg/dl: 0.88 (95% CI, 0.70–1.10); ≥40 μg/dl: 0.93 (95% CI, 0.76–1.15). Similar findings resulted when workers were classified by duration of exposure or cumulative exposure. The authors concluded that fertility is not reduced at the current levels of occupational lead exposure in these 4 countries (Belgium, Finland, Italy, and England).

Using the same data sources in which they investigated fertility as the outcome of male lead exposure, Sallmen et al. (2000b) assembled a sample of 502 couples and estimated the blood lead levels of the males during the 80-day period that preceded the attempt to conceive (i.e., the period of spermatogenesis). For 62% of the men, a blood lead level that was measured during that period was available. For the remaining 38%, this blood lead level was estimated based on blood lead levels that had been measured at other times or on the basis of a self-description of work. The critical endpoint was the “relative fecundability density ratio” (RFDR), which was essentially equivalent to TTP, ascertained by a questionnaire sent to couples 8 to 18 years after the index pregnancies. In a proportional hazards regression, using men with a blood lead level <10 μg/dl as the referent group (n = 176), all strata in which men had periconceptional blood lead levels >10 had an RFDR < 1, indicating reduced fecundability. The change in RFDR across strata was not clearly dose-related, however (10–20 μg/dl: 0.92; 95% CI, 0.73–1.16; n = 203; 21–30 μg/dl: 0.89; 95% CI, 0.66–1.20; n = 79; 31–40 μg/dl: 0.58; 95% CI, 0.33–0.96; n = 21; >40 μg/dl: 0.83; 95% CI, 0.50–1.32; n = 23). Despite the large sample size, however, the numbers of observations in the highest blood lead strata were relatively small, limiting the power of group comparisons.

Shiau et al. (2004) determined the TTP for 280 pregnancies of 133 married couples, 153 pregnancies that occurred when the men, battery workers, were not exposed to lead during the biologically relevant period, and 127 when they were. Cox proportional hazards regression was used to calculate a fecundability ratio, defined as the odds of a conception during exposure divided by the odds of conception during unexposed periods. The ratio decreased with increasing male blood lead levels, with the decreases being significant for the group with blood lead levels of 30–39 μg/dl (0.50; 95% CI, 0.34–0.74) and >39 μg/dl (0.38; 95% CI, 0.26–0.56). For the 41 couples with at least 1 pregnancy before exposure and another after the male began work in a lead occupation, TTP was longer for the pregnancy that occurred during exposure. Specifically, TTP was prolonged by 0.15 menstrual cycles when blood lead level increased by 1 μg/dl above a blood lead level of 10 μg/dl, but <40 μg/dl. Fecundability appeared to be more strongly associated with recent than cumulative exposure. This is the only study to report decreased fecundability or prolonged TTP for men with blood lead levels <40 μg/dl.

Finally, higher lead concentrations in seminal plasma are associated with lower pregnancy rates following artificial insemination (Benoff et al.,2003a) and in vitro fertilization (Benoff et al.,2003b).

Summary.

Considering these studies in aggregate, it appears likely fertility is reduced among couples during periods in which the male has a blood lead level exceeding 40 μg/dl or a level exceeding 25 μg/dl for a period of years. It might be expressed as a reduction in the number of live births, reduced odds of conception, or increased time to pregnancy. Whether scenarios involving less exposure also compromise a couple's fertility is less clear, although some evidence does suggest this. Because some studies involved capturing information about exposure and outcome from databases assembled for other reasons, the potential for some misclassification is high. Whether this introduced biases or merely decreased the precision of the estimates of lead's association with fertility cannot be determined.

Spontaneous Abortion

Paternal exposure.

Using national databases to identify the outcomes of 99,186 pregnancies in Finland between 1973 and 1982, Lindbohm et al. (1991a) determined that paternal employment in a job considered likely to involve lead exposure was not associated with an increased risk of spontaneous abortion (odds ratio [OR], 0.9; 95% CI, 0.8–1.0). In a second study, using a case-referent design (213 cases, 500 matched referents), Lindbohm et al. (1991b) classified men according to their lead exposure during the period of spermatogenesis of the index pregnancy. For 6% of men, classification was based on a blood lead level measured during this period. For other men, a regression equation was used to estimate this level from blood lead levels measured prior to or after this period. For the remaining men, classification was based solely on occupation title or work description. The risk of spontaneous abortion was not significantly associated with paternal lead exposure. When analyses were restricted to men for whom blood lead level was measured during the specific target period, the risk of spontaneous abortion for couples in which men had blood lead levels >30 μg/dl was significantly elevated (3.8; 95% CI, 1.2–2.0). However, this OR was based on a small number of men (12 cases, 6 referents).

In a study of the 203 spontaneous abortions among 2021 pregnancies, Alexander et al. (1996) found ORs of 0.8 (95% CI, 0.5–1.5) and 1.4 (95% CI, 0.7–2.5) for men in “lead smelting-moderate exposure” and “lead smelting-high exposure” jobs, respectively. When evaluated in relation to blood lead level 1 year prior to the index pregnancy, however, the risk was not elevated even among men with levels exceeding 40 μg/dl (OR, 0.7; 95% CI, 0.4–1.5). The low participation rate among eligible workers (37.6%) raises the possibility of a substantial selection bias in this study.

Maternal exposure.

Hertz-Picciotto (2000) reviewed 8 studies that evaluated the risk of spontaneous abortion among women with blood lead levels <30 μg/dl. Overall, the studies were considered to provide little evidence that higher exposures within this range were associated with increased risk, but the many methodological deficiencies identified in the studies limited the confidence that could be placed in this conclusion. Among these were small sample sizes, inadequate definition or ascertainment of the outcome, and inadequate control for potential confounding bias. In addition, exposure assessment tended to be weak, consisting in some cases of ecological indices and in others of biomarkers that were not collected at a time that was biologically appropriate for the outcome of interest.

One study that seemed to be less beset by these problems was conducted in Mexico City by Borja-Aburto et al. (1999). Through prenatal clinics, 668 women were identified with medically confirmed pregnancies still in the first trimester. By contacting women biweekly to determine pregnancy status, the investigators identified 35 cases of spontaneous abortion. For each case, a control was selected, matched on age, calendar time of entry into the study, public versus private clinic, and gestational age at study entry. Matching on the latter factor (called “incidence density matching”) ensured that a case and her control were similar in terms of the length of time available for the outcome to occur. Furthermore, it reduced the potential bias that might have attended a lack of comparability in the timing of exposure assessment. This is important because of evidence that a woman's blood lead level changes over the course of pregnancy (see section “Kinetics of Lead during Pregnancy and Lactation”). The mean blood lead levels for both the cases and controls (12.0 μg/dl, range 3.1–29; 10.1 μg/dl, range 1.3–26, respectively) were considerably higher than the current mean among U.S. women. Among the women who were followed to 20 weeks of gestation, 6.4% experienced a spontaneous abortion. Adjusting for spermicide use, active and passive smoking, use of alcohol and coffee, age, education, income, physical activity, hair dye use, use of a video display terminal, and medical conditions, the risk of spontaneous abortion increased by a factor of 1.3 for each μg/dl increase in blood lead level. Thus, a 5 μg/dl increase in blood lead level was associated with an OR of 1.8 (95% CI, 1.1–3.1). ORs were also calculated for groups defined by 5 μg/dl increments in blood lead level. Using women with levels <5 μg/dl as the referent group, the ORs for the 5–9, 10–14, and ≥15 μg/dl groups were 2.3, 5.4, and 12.2, respectively, a statistically significant trend, suggesting a coherent dose-response relationship. The CIs for these ORs were not provided, but they are likely to be broad. The interquartile range of the blood lead levels of cases was 6.5–16 μg/dl (Hertz-Picciotto,2000), which implies that approximately 9 cases had a blood lead level >15 (25% of 35 cases).

Summary.

The evidence that paternal lead exposures <30 μg/dl are associated with an increased risk of spontaneous abortion is limited, although this conclusion should be tempered by recognition of the limitations in the quality of the data used to classify men's lead exposure. Maternal lead exposures as low as 5–9 μg/dl might be associated with a doubling of the risk. Although this conclusion is based on the results of a well-designed study, an independent replication of the findings would be desirable.

Pregnancy Hypertension

Although many studies have explored the association between higher lead burdens and hypertension in adult males (Nawrot et al.,2002) and, to a lesser extent, adult females (Nash et al.,2003), few have focused specifically on hypertension during pregnancy. An early report suggested such a link (Rabinowitz et al.,1987), although in that study umbilical cord blood lead level was the sole exposure biomarker measured, and pregnancy hypertension was identified using obstetrical records rather than by direct measurement of blood pressure or assessment of clinical status. Three recent reports provide additional perspectives. In a case-control study nested within a cross-sectional study of 133 third-trimester primigravid women, Magri et al. (2003) found that the 33 women with gestational hypertension, defined as a systolic blood pressure ≥140 mm of Hg or a diastolic blood pressure ≥90 mm of Hg, had a significantly higher mean blood lead level than women who did not (9.6 vs. 5.8 μg/dl). Furthermore, both systolic and diastolic blood pressures bore a significant positive relationship with blood lead level, without apparent threshold.

In another case-control study, Sowers et al. (2002) compared, at 4 time points (12, 20, 28 weeks, and delivery), the blood lead levels of 71 women classified as “hypertension in pregnancy (HIP)/toxemia” (systolic blood pressure >140 mm of Hg; diastolic blood pressure >90 mm of Hg) to the blood lead levels of a referent group of 631 healthy women. Lead exposures were low in this sample, as the average blood lead level ranged between 1.08 (20 weeks) and 1.32 μg/dl (delivery). At all time points except 12 weeks of pregnancy, the HIP/toxemia group had a higher mean lead level, adjusting for age, ethnicity, and calcium intake. In both groups, however, the mean level never exceeded 1.5 μg/dl over the course of pregnancy.

In a prospective study, Rothenberg et al. (2002) found that the concentration of lead in the calcaneus (heel bone), but not tibia lead or concurrent blood lead level (mean: 1.9 μg/dl) was significantly associated with third trimester hypertension (OR, 1.86; 95% CI, 1.04–3.32), adjusting for age, parity, immigrant status, body mass index, smoking, education, and postpartum hypertension. Among normotensive women, a 10 μg/gm increase in calcaneus lead, representing approximately 1 standard deviation in the study sample, was associated with an increase of 0.70 mmHg (95% CI, 0.04–1.36) in third trimester systolic blood pressure and an increase of 0.54 (95% CI, 0.01–1.08) in third trimester diastolic blood pressure. Women with calcaneus lead values at the 95th percentile had an estimated systolic blood pressure 2.6 mmHg higher than that of women with calcaneus lead levels at the fifth percentile. In this study, a woman's postnatal blood lead level was, paradoxically, significantly inversely related to both systolic and diastolic blood pressures in the postpartum period. For each natural log unit increase in blood lead level, systolic blood pressure decreased 1.52 mmHg (95% CI, −2.83 to −0.20) and diastolic blood pressure decreased 1.67 mmHg (95% CI, −2.85 to −0.50).

Insight into a possible mechanism of lead-induced pregnancy hypertension was provided by Factor-Litvak et al. (1993). Studying women living near a large lead smelter in Kosovo, they found that among women with blood lead levels in the highest 10th percentile (mean of approximately 40 μg/dl, ranging up to 56.7), the adjusted OR for ≥1+ proteinuria, interpreted as a marker of renal damage, was 4.5 (95% CI, 1.5–13.6), using as a referent group women with lead levels in the lowest 10th percentile (mean of approximately 3 μg/dl). The OR began to rise above 1.0, however, when blood lead levels were >5.8 μg/dl.

Given the dramatic decline in women's blood lead levels in recent decades, one might expect that, if lead exposure is causally related to pregnancy hypertension, the incidence of this complication would also have declined. Drawing inferences about the likelihood of causality from data muddied by possible cohort effects is always risky, however. Coincident with the decrease in lead exposure were increases in other risk factors for pregnancy hypertension (Cunningham et al.,2001), such as older age at pregnancy, greater body mass index, multiple gestations, and reduced smoking.

Summary.

The 2 recent case-control studies suggest that the risk of pregnancy hypertension might be increased even at blood lead levels below 10 μg/dl. The recent cohort study of Rothenberg et al. (2002) suggests, however, that bone lead level, an index of long-term exposure, might be a better predictor of a woman's risk than concurrent blood lead level, which is an index of short-term exposure.

Other Pregnancy Complications

In a series of nested case-control analyses carried out in 100 pregnant women from the general population in India, Kaul et al. (2002) found that whereas the mean blood lead level of women with uncomplicated pregnancies was 6.2 μg/dl, the means among women with anemia, proteinuria and essential hypertension, toxemia, and hyperemesis were all elevated, 12.9, 18.4, 26.0, and 35.5 μg/dl, respectively. Falcon et al. (2003) reported higher lead concentrations in trophoblastic placental tissues of women with premature rupture of the membranes or preterm delivery (n = 18) than in the tissues of women with uncomplicated pregnancies (n = 71).

Offspring Outcomes

Congenital malformations.

Using the same Finish national data sources they used in studies of other reproductive outcomes, Sallmen et al. (1992) conducted a case-control study to evaluate the hypothesis that paternal lead exposure during the period of spermatogenesis increases the risk of congenital malformations in offspring. The unadjusted OR among lead-exposed men (blood lead level >20 μg/dl) was 2.4 (95% CI, 0.9–6.5). When adjusted for one potential confounder at a time (because of small cell counts), it ranged from 1.9 to 3.2. The offspring of the 5 men with the highest exposures all had a different type of malformation, however (congenital heart disease, oral cleft, club foot, polydactyly, adrenal gland). Alexander et al. (1996) found ORs of 1.9 (95% CI, 0.6–6.3) and 2.7 (95% CI, 0.7–9.6) for men in “lead smelting-moderate exposure” and “lead smelting-high exposure” jobs, respectively, when they combined the 30 birth defects and 12 stillbirths that occurred among the 2021 pregnancies in their study cohort. Separate analyses for birth defects alone were not reported. In a Norwegian study, neither maternal nor paternal lead exposure was associated with an increased risk of “serious birth defects” (maternal OR, 1.25; 95% CI, 0.80–1.90; paternal OR, 0.94; 95% CI, 0.82–1.08) (Irgens et al.,1998). Kristensen et al. (1993) also failed to find a significant association between paternal lead exposure and total birth defects, although they did report a 4-fold increased risk of cleft lip in the male offspring of lead-exposed men. Irgens et al. (1998) found isolated cleft palate in male (but not female) offspring to be associated with paternal lead exposure (OR, 1.73; 95% CI, 0.97–2.89). Unlike Kristensen et al. (1993), however, the risk of cleft palate was not increased among male offspring of lead-exposed men.

In a case-control study in England involving 88,449 births, Bound et al. (1997) found that the risk of delivering an infant with a neural tube defect (NTD) (n = 363) was greater for women who lived in areas in which elevated water lead levels (>10 μg/liter) were more common. This association remained when cases of anencephaly (n = 169) and cases of spina bifida and cranium bifidum (n = 195) were examined separately. The associations were reduced but remained close to significance when adjustment was made for an index of deprivation. The authors speculated that the association could reflect a direct effect of lead on the risk of an NTD or an indirect effect resulting from a lead-induced reduction in zinc bioavailability (resulting in a zinc-induced reduction in folate uptake). In the Irgens et al. (1998) study, the odds of an NTD among the 1886 infants born to women classified as lead-exposed was 2.87 (95% CI, 1.05–6.38). The odds of an NTD were not elevated among the 35,930 infants born to men classified as lead-exposed, however.

A significant association between parental lead exposure during pregnancy (self-reported, industrial hygiene assessment, or job exposure matrix) and a specific cardiac defect in offspring, total anomalous pulmonary venous return, was found in a case-control study of the Baltimore-Washington Infant Study (Jackson et al.,2004). For women, the OR associated with exposure during the periconceptional period (3 months prior to conception through first trimester) was 1.57 (95% CI, 0.64–3.47). For men, the OR was 1.83 (95% CI, 1.00–3.42).

Summary.

Overall, the evidence that prenatal parental lead exposure, except at very high levels, causes congenital malformations remains modest, although studies have linked it to specific malformations of the brain and heart. Until independently replicated, however, these associations should be viewed cautiously. In some studies demonstrating an association, classification of exposure was not based on measurement of a biomarker but on job title. The resulting uncertainty regarding the actual magnitude of exposure (and the possibility of teratogenic coexposures) make it difficult to characterize the dose-response relationship quantitatively.

Birth size, birth weight, and head circumference.

Torres-Sanchez et al. (1999) conducted a case-control study involving 459 full-term births and 161 preterm births in Mexico City. They found that cord blood lead levels were significantly higher in infants born preterm (9.8 ± 2 μg/dl) than in infants born at term (8.4 ± 2.2), but this was true only among primiparous women.

Using a Norwegian birth registry and census records for the period 1970–1993, Irgens et al. (1998) found that lead-exposed women (none, low/moderate, high) were slightly more likely than nonexposed women to deliver a low birth weight (OR, 1.34; 95% CI, 1.12–1.60) or preterm infant (OR, 1.13; 95% CI, 0.98–1.29). Paternal exposure, classified ordinally in the same manner, did not appear to increase the risk of either outcome.

Min et al. (1996) identified 220 low birth weight infants in the Baltimore-Washington Infant Study database and compared their parental lead exposure histories to those of 522 controls. Maternal lead exposure was infrequent (12 women overall) and not associated with risk of low birth weight (OR, 0.89; 95% CI, 0.24–3.33). Overall, paternal lead exposure did not increase the risk (OR, 1.00; 95% CI, 0.67–1.50), although additional subgroup analyses suggested that the risks of term small-for-gestational-age, preterm small-for-gestational-age, and preterm appropriate-for-gestational-age were all elevated for infants of men with lead exposures >10% of the threshold limit value (TLV) (2.93; 95% CI, 0.94–9.19; 2.44; 95% CI, 1.92–3.11; 2.08; 95% CI, 0.66–6.52; respectively). Lin et al. (1998) reported similar findings in a study comparing the offspring of lead-exposed men to those of nonexposed bus drivers. Overall, no significant group differences were found in the risk of low birth weight, preterm delivery, small-for-gestational age, or Apgar score <7. Among men who had had an elevated blood lead level (>25 μg/dl) for >5 years, however, the risks of having a preterm or low birth weight infant were both significantly elevated (relative risk, 3.0; 95% CI, 1.6–6.8; relative risk, 3.4; 95% CI, 1.4–8.4; respectively). The risk of an infant being small-for-gestational-age was not elevated, however, in contrast to the findings of Min et al. (1996)

Gonzalez-Cossio et al. (1997) found, in a cohort of 272 infants in Mexico City, that higher tibia lead levels, but not patella lead or umbilical cord blood lead level, were associated with reduced birth weight. The bone lead levels were measured 1 month after delivery. In the highest quartile, infants were, on average, 156 gm lighter than infants in the lowest quartile. In addition, infants with tibia lead levels in the upper quintile had a 1.79 greater risk (95% CI, 1.10–3.22), using a cumulative odds model, of having a shorter birth length, compared to infants in the other quintiles (Hernandez-Avila et al.,2002).

In a study conducted among 233 women in Mexico City, Hernandez-Avila et al. (2002) found that a higher maternal patella lead level (measured at 1 month postpartum; mean 14.14 μg of lead/gm of bone mineral; SD = 13) placed an infant at increased risk of having a smaller head circumference (1.02 cm/μg of lead/gm of bone mineral (95% CI, 1.01–1.04). In another study conducted in Mexico City, Rothenberg et al. (1999a) found a similar inverse association. Specifically, an increase from 1 to 35 μg/dl in maternal blood lead at 36 weeks of pregnancy was associated with a reduction of 1.9 cm (95% CI, 0.9–3.0) in infant head circumference at 6 months of age. This association was temporary, however. Moreover, infants' head circumference was measured 9 times between birth and 48 months and examined in relation to 6 indices of prenatal lead exposure (12, 20, 28, and 36 weeks, maternal blood at delivery, and cord blood). Therefore, this was the only 1 of the 54 possible associations that was significant.

Summary.

The findings of recent studies are generally consistent with those of earlier studies (Andrews et al.,1994) in suggesting that maternal lead exposure during pregnancy is inversely related to fetal growth, as reflected in duration of pregnancy and infant size. Identifying a blood lead level at which the risks begin to increase is difficult either because of the manner in which exposure was classified within a study (i.e., categorically rather than quantitatively) or because these adverse outcomes were associated with bone lead levels but not with blood lead levels. High paternal exposures (defined as a blood lead level >10% of the TLV or >25 μg/dl for at least 5 years) appear to increase an infant's risk of preterm birth and low birth weight.

Neurobehavioral development.

Although the CDC (1991) proposed a blood lead level of 10 μg/dl as a risk management guideline rather than as a “threshold,” it has often been imbued with special biological meaning. Recent evidence suggests that the latter interpretation is incorrect and that postnatal blood lead levels <10 μg/dl are associated with enduring adverse neurodevelopment (Lanphear et al.,2000; Canfield et al.,2003; Bellinger and Needleman,2003; Chiodo et al.,2004). Whether this is true, as well, for prenatal lead levels <10 μg/dl is less certain. Emory et al. (1999) evaluated the association between blood lead levels, measured at 6 to 7 months of gestation and at delivery, and scores on the Neonatal Behavior Assessment Scale (NBAS), administered at 1 to 2 days of age, in 103 African-American neonates. The findings were weak and equivocal. No associations were found when conventional methods of clustering NBAS items were applied. In secondary analyses involving only 40 of the children, the investigators compared the scores of neonates in the lowest (≤1 μg/dl; mean 0.86) and highest (≥2.5 μg/dl; mean 4.0) blood lead quartiles on individual NBAS items that were considered, on the basis of previous studies, most likely to be related to lead exposure: attention, motor function, and social interaction. Applying 1-tailed tests of significance without adjustment for potential confounding, the investigators reported group differences, favoring the lowest quartile group, on 4 of the 9 items evaluated. Trends in the same direction were found for the other items.

Emory et al. (2003) administered the Fagan Test of Infant Intelligence, a recognition memory paradigm, to 79 of the infants in this cohort at age 7 months. As with the neonatal assessments, the investigators chose to include in the analyses only the subset of infants with extreme values on both the Fagan test (“high” >85th percentile; “low” 15th percentile) and blood lead level (“high” fourth quartile; “low” first, second, or third quartile). The chi-square statistic for the resulting 2 × 2 contingency table was statistically significant, although the difference between the mean blood lead levels of the high- and low-performing infants on the Fagan test was very small (0.44 vs. 0.94 μg/dl).

Tang et al. (1999) studied a group of 244 infants who had been exposed to low levels of lead in utero (mean umbilical cord blood lead level: 3.9 μg/dl, 5th–95th percentile: 2.5–7.5). At 9 months of age, infants were administered the Brunet-Lezine scales, which yield a global score as well as scores on posture, coordination, language, and sociability. To explore potential neurochemical mechanisms of lead toxicity, two monoamine metabolites were measured in cord plasma, homovanillic acid (HVA) and 5-hydroxyindoleacetic acid (5-HIAA). Only the sociability score was significantly associated, in an inverse direction, with cord blood lead level. The association remained statistically significant if 5-HIAA concentration was controlled by means of partial correlation, but it did not if HVA concentration was controlled. The investigators speculated that HVA, and thus the dopaminergic system rather than the serotonergic system, mediated the association between lead and sociability. This interpretation seems unlikely, however, insofar as lead in cord blood was not correlated with HVA (r = 0.018; p = 0.82). In addition, controlling for HVA, a significant inverse association was found between cord blood lead and the global score. Controlling for 5-HIAA, significant inverse associations were found between cord blood lead and global score, posture, and coordination. This pattern is consistent with the inference that monoamine metabolite levels confounded rather than mediated the associations between cord blood lead level and neurodevelopment.

Among infants in a cohort born in Mexico City, Gomaa et al. (2002) found that umbilical cord blood lead (mean 6.7 μg/dl, range 1.2–21.6) and bone lead levels were each independently associated with covariate-adjusted scores at 2 years of age on the Mental Development Index (MDI) on the Bayley scales. Infants with patella lead levels in the lowest quartile had a mean MDI score that was 5.4, 7.2, and 6.5 points higher than the mean scores of children in the second, third, and fourth quartiles, respectively. The findings of this study suggest that cord blood and bone lead levels provide complementary rather than redundant information about fetal lead exposure.

Shen et al. (1998) classified 133 infants in Shanghai into 2 groups based on umbilical cord blood lead levels, which ranged from 1.6 to 17.5 μg/dl: <30th percentile or >70th percentile. The geometric mean level in the entire sample of 348 was 9.2 μg/dl. At 3, 6, and 12 months of age, the infants in the high exposure group had significantly lower covariate-adjusted scores on the MDI of the Bayley scales, with the differences ranging from 3.4 to 6.3 points. Thus, the studies of Gomaa et al. (2002) and Shen et al. (1998) largely replicate the results of several older studies indicating that modest elevations in cord blood lead levels are associated with deficits in early cognitive development (e.g., Bellinger et al.,1987).

Several of the long-running prospective studies of lead exposure and child development have recently reported on the associations between indices of prenatal lead exposure and late neurodevelopment. The findings have been mixed. In some of the studies, prenatal lead exposures were not associated with intelligence quotient (IQ) scores measured in middle and later childhood (e.g., Bellinger et al.,1992; Tong et al.,1996). In 2 of the studies, however, such associations were found. Among children living near a lead smelter in Kosovo, IQ scores at the age of 8 years were significantly inversely associated with a composite index of prenatal lead exposure (average of mothers' blood lead levels at midpregnancy and at delivery) (Wasserman et al.,2000). This association was independent of changes in postnatal blood lead levels. Among 15- to 17-year-old inner-city children in Cincinnati, maternal blood lead levels in the first trimester, which ranged from 1 to approximately 30 μg/dl, were inversely related to children's scores on tests of attention and visuoconstruction (Ris et al.,2004) and to the frequency of self-reported delinquent behaviors (Dietrich et al.,2001). The explanation(s) for the differences in findings across studies are not known.

Summary.

The findings of recent cohort studies suggest that even within the blood lead range with an upper bound of 10 μg/dl, prenatal lead level, like postnatal blood lead level, is inversely related to neurobehavioral status in childhood. Replication of these provocative findings is needed to clarify their import, however. To the extent that this can be determined, the results are roughly consistent with those of earlier studies in terms of the estimated effect sizes. While the lead-associated differences in test score are small when viewed as a potential change in an individual child's score, they acquire substantially greater importance when viewed as a shift in the mean score within a population (Bellinger,2004). The mechanism(s) by which low-level lead exposure, whether incurred prenatally or postnatally, might adversely affect neurobehavioral development remains uncertain, although many hypotheses have been proposed (Lidsky and Schneider,2003).

KINETICS OF LEAD DURING PREGNANCY AND LACTATION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EPIDEMIOLOGY OF LEAD EXPOSURE IN PREGNANT WOMEN
  5. LEAD EXPOSURE AND ADVERSE PREGNANCY EVENTS AND OUTCOMES
  6. KINETICS OF LEAD DURING PREGNANCY AND LACTATION
  7. CALCIUM INTAKE, CALCIUM SUPPLEMENTATION, AND MOBILIZATION OF LEAD FROM BONE
  8. POPULATION SUBGROUPS AT INCREASED RISK
  9. CONCLUSION
  10. REFERENCES

The concentration of lead in blood during pregnancy is a function of exogenous exposures to lead, plasma volume, red cell mass, and the redistribution of lead among different body pools stimulated by physiologic changes that occur during pregnancy. Numerous studies document nonlinearity in blood lead level over the course of pregnancy. In one study of 105 healthy women in Mexico City, in whom blood lead levels were measured at 12, 20, 28, 36 months and at delivery and averaged 7.0 μg/dl throughout pregnancy, blood lead declined about 1 μg/dl between 12 and 20 weeks (Rothenberg et al.,1994). It is not clear whether this was due to hemodilution, organ growth, or increased glomerular filtration rate and/or lead excretion. Between 20 weeks of gestation and delivery, blood lead increased by 1.6 μg/dl. This U-shaped pattern of change in blood lead level over pregnancy is consistently observed (Hertz-Picciotto et al.,2000; Schell et al.,2000,2003; Sowers et al.,2002; Gulson et al.,2004). Plasma lead level appears to follow a similar pattern (Tellez-Rojo et al.,2004). The magnitude of the increase varies across studies (14–40%), presumably because the lead exposure histories of the study samples differed. This temporal pattern suggests an increased influx of lead into blood in later pregnancy. In principle, this could be the result of changes in one or more aspects of lead metabolism, including increased gut absorption, increased retention, or the mobilization into blood of lead stored in deeper body pools. It is the latter possibility that has attracted the greatest interest among lead researchers.

Changes in blood lead level have been observed in association with various physiologic states associated with increased metabolic activity in bone. One example is menopause. In the Hispanic NHANES survey (1982–1984), blood lead level was higher in postmenopausal women than in premenopausal women (Symanski and Hertz-Picciotto,1995). The highest levels were in women whose menopause had occurred within the previous 4 years, and the mean blood lead level declined slowly with each year since menopause. The difference between the blood lead levels of premenopausal and postmenopausal women was greatest among women who had never been pregnant. Using data from NHANES-III (1988–1994), Nash et al. (2004) found that bone mineral density was inversely related to blood lead level in perimenopausal women. These observations suggest that bone lead is mobilized at rates that are consistent with the pattern of bone loss that occurs in association with menopause.

A number of observations provide indirect evidence that the increase in blood lead in the latter half of pregnancy might be due, at least in part, to the mobilization of some portion of the large fraction of the adult body lead burden that is stored in bone (90–95%). The timing of the blood lead increase coincides with a period of increased fetal need for and increased maternal provision of calcium. Women with lower milk intakes, and thus lower calcium intakes, have higher late pregnancy blood lead levels than do women with higher milk intakes. Women reporting more abortions or more prior pregnancies have lower late pregnancy blood lead levels than do primiparous women, suggesting that prior pregnancies deplete bone lead stores. In predicting plasma lead level during pregnancy in a cohort of 193 Mexican women, a significant interaction was found between bone lead level (measured at 4 weeks postpartum) and urinary cross-linked N-telopeptides of type 1 collagen (NTx). Specifically, plasma lead level was highest among women who had both a high bone lead level and greater bone resorption activity (Tellez-Rojo et al.,2004). Furthermore, it appears that lead in bone stores might be mobilized into plasma without producing an increase in whole blood lead level (Gonzalez-Cossio et al.,1997; Chuang et al.,2001).

Calcium needs during lactation exceed those even during late pregnancy, and blood lead changes associated with lactation are also consistent with an increase in the mobilization of lead from bone during this period. In 1 study of 24 women of child-bearing age in Mexico, calcaneus lead level was inversely related to the number of months of lactation (Moline et al.,2000). In the first postpartum month, blood lead level decreases in women who do not breastfeed their infants but increases among women who do. Furthermore, this increase is greater among women who are exclusive breast feeders than it is among women who are partial breast feeders (Tellez-Rojo et al.,2002). Similarly, among women who breastfed, the rise observed in blood lead level that begins in late pregnancy continued after delivery, declining only when lactation ceased (Manton et al.,2003). Furthermore, the higher the blood lead level in pregnancy, the greater the rise in blood lead level during lactation.

Natural experiments involving stable lead isotopes provide more direct evidence of bone lead mobilization during pregnancy and lactation and the factors that influence its magnitude. The distribution of isotopes in a particular sample of lead ore varies depending on its geologic origins. In effect, the distribution corresponds to the ore's isotopic “signature” or “fingerprint.” Gulson et al. (2003) measured blood lead longitudinally in a group of 15 women who became pregnant shortly after immigrating to Australia from Eastern Europe, where environmental lead has a different isotopic composition than environmental lead in Australia. By measuring the changes in the ratios of lead isotopes in the blood of such immigrant women as they became pregnant, delivered, and lactated, Gulson et al. (2003) were able to estimate the contribution of skeletal (endogenous) lead to blood lead during pregnancy. Biologic and environmental samples were collected soon after a woman's arrival in Australia, prepregnancy (if possible), throughout pregnancy, and for at least 6 months postpartum. The mean blood lead level of the pregnant immigrants (2.8 μg/dl; range 1.5–20) did not differ significantly from the mean level of the native Australian women who served as controls (2.9 μg/dl; range 1.9–4.3). In women who arrived in Australia before becoming pregnant, the lead isotope ratio (206/204) in blood steeply declined over their first 100 days of residence, moving closer to the ratio characteristic of environmental lead in Australia. This change presumably reflected the clearance of European lead from the women's blood. The ratio then stabilized, reflecting an equilibrium between skeletal and current environmental lead.

When an immigrant woman became pregnant, however, the 206/204 ratio increased, in the direction of the European ratio, suggesting that lead to which a woman had been exposed prior to arriving in Australia was now entering her blood. Although the contribution of skeletal lead to blood lead was greatest in women who conceived within 100 days of arriving in Australia, an increase in the ratio was observed even in a woman who had resided in Australia for more than 2 years prior to conception (800 days). The 206/204 ratio in blood did not change when native Australian women became pregnant, although their blood lead levels showed the usual pregnancy-related pattern of temporal change. Comparison of the isotopic ratio of the lead in blood upon a woman's arrival in Australia to the isotopic ratio of lead in her infant's umbilical cord blood suggested that maternal skeletal lead mobilized during pregnancy accounted, on average, for 79% of the lead transferred gestationally.

Manton et al. (2003) also examined lead isotope ratio (206/207) changes during pregnancy (n = 12), reporting findings generally similar to those of Gulson et al. (2003). They observed that, over successive pregnancies, the direction of the change in isotopic ratio remained the same, but its magnitude was diminished, suggesting a gradual depletion of a woman's “old” bone stores. By measuring isotopic ratios in women's diet and in house dust samples, they confirmed that it was pregnancy-related changes in bone resorption rather than in dietary absorption that influenced the blood lead changes during pregnancy.

Manton et al. (2003) concluded that in early pregnancy, it is only trabecular bone that is resorbed. This type of bone, which they postulated has a lower lead concentration than does cortical bone, releases amounts of lead that are not sufficient to compensate for the tendency of hemodilution to reduce blood lead concentration in early pregnancy (the rough U-shaped function). They argued that, in late pregnancy, resorption shifts to cortical bone, which is responsible for the increase in blood lead that occurs at that time. Even in women who only bottle-feed their infants, and for whom cortical bone contribution to blood lead ceases at delivery, blood lead level does not necessarily decline immediately. The hemoconcentration caused by postpartum loss of plasma volume might be sufficient to compensate for the interruption in the mobilization of bone lead stores. Gulson et al. (2004) disputed this interpretation, however, citing evidence that lead concentration tends to be greater in trabecular than cortical bone and that because of its higher vascularity and turnover rate, trabecular bone has the greater influence on blood lead level. A slew of questions clearly remain to be answered regarding these processes.

Summary.

The major lesson of the elegant isotopic studies by Gulson et al. (2003,2004) and Manton et al. (2003) is that it is not only a woman's exposures to exogenous lead sources during pregnancy that are germane in determining her risk and that of her fetus. Because of changes in bone metabolism during pregnancy, her past exposures, represented by the lead stored in deep pools with long biological half-life periods, represent potential endogenous sources of lead both for her and her fetus.

CALCIUM INTAKE, CALCIUM SUPPLEMENTATION, AND MOBILIZATION OF LEAD FROM BONE

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EPIDEMIOLOGY OF LEAD EXPOSURE IN PREGNANT WOMEN
  5. LEAD EXPOSURE AND ADVERSE PREGNANCY EVENTS AND OUTCOMES
  6. KINETICS OF LEAD DURING PREGNANCY AND LACTATION
  7. CALCIUM INTAKE, CALCIUM SUPPLEMENTATION, AND MOBILIZATION OF LEAD FROM BONE
  8. POPULATION SUBGROUPS AT INCREASED RISK
  9. CONCLUSION
  10. REFERENCES

As noted, among pregnant women, a lower calcium intake is associated with higher blood lead levels in late pregnancy. In a sample of 195 women who entered prenatal care by 13 weeks of pregnancy and in whom blood lead level was measured 5 times prior to delivery (Hertz-Picciotto et al.,2000), calcium intake, determined by means of a food frequency questionnaire, was inversely related to blood lead level, although only in the latter half of pregnancy. Furthermore, the apparent protective effect of higher calcium intake increased as pregnancy progressed. These associations were observed at blood lead levels that were low, as only 2 of the 878 measurements made over the course of the study were >5 μg/dl. The magnitude of effect was modest, however, insofar as women with a calcium intake <600 mg/day were estimated to have a blood lead level that was approximately 1 μg/dl higher than the level of women with a calcium intake >2000 mg/day.

Given the metabolic interactions between lead and calcium, the possibility that calcium supplementation might reduce the mobilization of lead from bone during pregnancy and lactation and thus reduce fetal and neonatal lead exposure has generated considerable interest given that it is a relatively low cost and low risk intervention. Using a rodent model, Han et al. (2000) showed that the administration of dietary calcium during pregnancy to dams for which lead exposure ended well before pregnancy reduced fetal lead transfer, although it did not reduce adverse effects of lead on birth weight or length.

Gulson et al. (2004) reported a case series of 10 women who were given 1 gm/day of calcium during pregnancy and for 6 months postpartum. Whereas the pregnancy-associated increase in blood lead level usually occurs between 3 and 6 months of pregnancy, in the women receiving calcium, it was delayed until 6 to 8 months. The magnitude of the increase in blood lead level was similar to that usually observed, but, because of the delay in its onset, the net effect was, as in calcium-supplemented rodents, reduced efflux of lead from bone.

Insight into the physiological basis for the potential benefits of calcium supplementation during pregnancy was provided by a small study of bone resorption by Janakiraman et al. (2003). Each of 31 pregnant women (25–35 weeks of gestation) received 1200 mg of calcium for 10 consecutive days, and then a multivitamin that did not contain any calcium for the next 10 days. Levels of urinary NTx tended to be lower when women were receiving calcium (27/31 women). The authors concluded that calcium supplementation reduced bone resorption during pregnancy by approximately 14%.

In the Gulson et al. (2004) study, calcium supplementation did not appear to be similarly effective in reducing lead efflux from bone during lactation. The reasons are not clear. The mechanisms of bone resorption during pregnancy and lactation are poorly understood, with some arguing that bone mineral changes during lactation are independent of the amount of dietary calcium (Prentice,2000). If so, calcium supplementation during lactation would not be expected to affect the efflux of lead from bone during this period. The only randomized trial addressing this issue did, however, suggest a modest benefit of supplementation during lactation. Hernandez-Avila et al. (2003) enrolled 617 lactating women and provided them with either 1200 mg of calcium per day or a placebo. Blood lead level, which averaged 8.5 μg/dl at study entry, was measured at 3 and 6 months later. Intent-to-treat analyses indicated that supplementation produced a reduction of 0.29 μg/dl in blood lead level (95% CI, −0.85 to 0.26). Additional analyses indicated that the blood lead reduction was greater, 16% or 1.16 μg/dl (95% CI, −2.08 to −0.23) among women who were compliant and who had higher bone lead levels at study entry.

Summary

Experimental animal and observational human studies are consistent in suggesting that increased calcium intake during the second half of pregnancy, by reducing bone resorption, is inversely related to mobilization of lead from bone compartments. This process might occur at blood lead levels well below 10 μg/dl. Whether calcium supplementation during lactation is similarly efficacious is uncertain.

POPULATION SUBGROUPS AT INCREASED RISK

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EPIDEMIOLOGY OF LEAD EXPOSURE IN PREGNANT WOMEN
  5. LEAD EXPOSURE AND ADVERSE PREGNANCY EVENTS AND OUTCOMES
  6. KINETICS OF LEAD DURING PREGNANCY AND LACTATION
  7. CALCIUM INTAKE, CALCIUM SUPPLEMENTATION, AND MOBILIZATION OF LEAD FROM BONE
  8. POPULATION SUBGROUPS AT INCREASED RISK
  9. CONCLUSION
  10. REFERENCES

The data reviewed regarding the mobilization of lead from bone during pregnancy and lactation suggest that women who enter pregnancy with large amounts of lead stored in this compartment should be considered to be at increased risk of lead-associated reproductive adversities. This could result from several exposure scenarios. One such scenario is growing up near a point source such as a smelter or in inner-city neighborhoods that, particularly in past decades, provided numerous sources and pathways of exposure (air, dusts, soils, paint, and water).

Another exposure scenario associated with increased risk is early residence in a developing country in which general environmental exposures to lead remain high, perhaps due to industrial practices or to the continued use of lead as a gasoline additive. The prevalence of elevated blood lead levels is also higher among internationally adopted children (CDC,2000). In a study of 1428 pregnant women in south central Los Angeles, of whom 74% were immigrants, mostly Latino, the number of years living in the United States was among the strongest predictors of a blood lead level exceeding 10 μg/dl (Rothenberg et al.,1999b). Although a strong inverse relationship was found between blood lead level and number of years in the United States, the rate of decline was shallow, suggesting that the women arrived in the United States with substantial bone lead burdens. Similarly, in a study of pregnant women in New York City, the proportion of “foreign born” women who had blood lead levels >20 μg/dl was twice as high as the proportion of foreign-born women among all women giving birth in New York City during the same period (Klitzman et al.,2002). As in the studies of Rothenberg et al. (1999b,2000), blood lead level was inversely related to the length of time a woman had resided in the United States.

Women with long-standing habits associated with increased lead exposure are another group at risk. These habits include smoking and alcohol consumption (Rhainds and Levallois,1997), pica for soil or clay (Rothenberg et al.,1999b; Klitzman et al.,2002; Shannon,2003), the use of certain ethnic remedies, cosmetics, and complementary medications (e.g., Ayervedic medicines) (Anderson et al.,2001; Tait et al.,2002; CDC,2004), and food preparation methods that involve lead-glazed vessels that were not fired at temperatures that were sufficiently high (Azcona-Cruz et al.,2000; Brown et al.,2000; Rothenberg et al.,2000). Women with low dietary calcium intakes are also at increased risk. This conclusion is supported by observational studies (Rothenberg et al.,1999b; Hertz-Picciotto et al.,2000), as well as by randomized (Hernandez-Avila et al.,2003) and nonrandomized intervention trials (Janakiraman et al.,2003; Gulson et al.,2004). It should be noted, however, that calcium supplements have sometimes been found to be lead-contaminated (Bourgoin et al.,1993). Finally, poorly conducted home renovations involving the disturbance of leaded surfaces is also a risk factor (Fletcher et al.,1999).

A validated questionnaire for triaging pregnant women in terms of lead risk is not available. Stefanak et al. (1996) evaluated the utility of an instrument that consisted of 4 of the 5 questions recommended by the CDC for evaluating a child's risk (time spent in an older house with paint in poor repair, renovation or remodeling of an older house, a sibling or friend with lead poisoning, or an adult whose job or hobby involved exposure to lead; the question about exposure to point sources was omitted). Among the 299 pregnant women who participated, 13% had a blood lead level >10 μg/dl. Answering “yes” to one or more of the questions was associated with more than a doubling of the risk of having an elevated blood lead level (95% CI, 1.2–4.9). Interpreting one “yes” response as a positive screen, the sensitivity of the instrument was 76% and the negative predictive value 93%. Adding questions about smoking and consumption of canned foods improved these numbers to 89% and 96%, respectively. This suggests good concordance between the risk factors for elevated blood lead levels in children and adults.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EPIDEMIOLOGY OF LEAD EXPOSURE IN PREGNANT WOMEN
  5. LEAD EXPOSURE AND ADVERSE PREGNANCY EVENTS AND OUTCOMES
  6. KINETICS OF LEAD DURING PREGNANCY AND LACTATION
  7. CALCIUM INTAKE, CALCIUM SUPPLEMENTATION, AND MOBILIZATION OF LEAD FROM BONE
  8. POPULATION SUBGROUPS AT INCREASED RISK
  9. CONCLUSION
  10. REFERENCES

As in other areas of lead toxicology, adverse health impacts of lead exposure during pregnancy are being observed at lower and lower biomarker levels. Focusing on lower and lower exposures, an investigator is faced with methodological challenges of increasing complexity. These involve the choice of the exposure biomarker that is biologically most relevant, quality assurance (QA)/quality control (QC) issues in biomarker collection and measurement, and the identification, measurement, and statistical control for potential confounders. An innovative feature of recent research is its focus on explicating the importance of endogenous sources of exposure resulting from pregnancy-related mobilization, from long-term bone stores, of lead to which a woman might have been exposed in the distant past. The studies using changes in the ratios of stable lead isotopes to distinguish the contributions of exogenous and endogenous sources to fetal exposure have been especially informative. It is notable that many of the recent studies were conducted in regions where exposures are higher than those of most contemporary U.S. women, suggesting that the morbidities associated with gestational lead exposure in U.S. women are almost certainly milder than the morbidities observed in the studies. Nevertheless, because of uncertainties regarding the “no observed adverse effect levels” for adverse pregnancy events and outcomes, primary prevention of lead exposure, including during the years prior to pregnancy, remains the best approach, at present, to reducing the burden of lead-associated morbidities.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EPIDEMIOLOGY OF LEAD EXPOSURE IN PREGNANT WOMEN
  5. LEAD EXPOSURE AND ADVERSE PREGNANCY EVENTS AND OUTCOMES
  6. KINETICS OF LEAD DURING PREGNANCY AND LACTATION
  7. CALCIUM INTAKE, CALCIUM SUPPLEMENTATION, AND MOBILIZATION OF LEAD FROM BONE
  8. POPULATION SUBGROUPS AT INCREASED RISK
  9. CONCLUSION
  10. REFERENCES