Relationship of Lead, Mercury, Mirex, Dichlorodiphenyldichloroethylene, Hexachlorobenzene, and Polychlorinated Biphenyls to Timing of Menarc
http://www.100md.com
《小儿科》
the University at Albany, State University of New York, Albany, New York
ABSTRACT
Background. Children are commonly exposed at background levels to several ubiquitous environmental pollutants, such as lead and persistent organic pollutants, that have been linked to neurologic and endocrine effects. These effects have prompted concern about alterations in human reproductive development. Few studies have examined the effects of these toxicants on human sexual maturation at levels commonly found in the general population, and none has been able to examine multiple toxicant exposures. The aim of the current investigation was to examine the relationship between attainment of menarche and levels of 6 environmental pollutants to which children are commonly exposed at low levels, ie, dichlorodiphenyldichloroethylene (p,p'-DDE), hexachlorobenzene (HCB), polychlorinated biphenyls (PCBs), mirex, lead, and mercury.
Methods. This study was conducted with residents of the Akwesasne Mohawk Nation, a sovereign territory that spans the St Lawrence River and the boundaries of New York State and Ontario and Quebec, Canada. Since the 1950s, the St Lawrence River has been a site of substantial industrial development, and the Nation is currently adjacent to a US National Priority Superfund site. PCB, p,p'-DDE, HCB, and mirex levels exceeding the US Food and Drug Administration recommended tolerance limits for human consumption have been found in local animal species. The present analysis included 138 Akwesasne Mohawk Nation girls 10 to 16.9 years of age. Blood samples and sociodemographic data were collected by Akwesasne community members, without prior knowledge of participants' exposure status. Attainment of menses (menarche) was assessed as present or absent at the time of the interview. Congener-specific PCB analysis was available, and all 16 PCB congeners detected in >50% of the sample were included in analyses (International Union of Pure and Applied Chemistry numbers 52, 70, 74, 84, 87, 95, 99, 101 [+90], 105, 110, 118, 138 [+163 and 164], 149 [+123], 153, 180, and 187). Probit analysis was used to determine the median age at menarche for the sample. Binary logistic regression analysis was used to determine predictors of menarcheal status. Six toxicants (p,p'-DDE, HCB, PCBs, mirex, lead, and mercury) were entered into the logistic regression model. Age, socioeconomic status (SES), and BMI were tested as potential cofounders and were included in the model at P < .05. Interactions among toxicants were also evaluated.
Results. Toxicant levels were measured in blood for this sample and were consistent with long-term exposure to a variety of toxicants in multiple media. Mercury levels were at or below background levels, all lead levels were well below the Centers for Disease Control and Prevention action limit of 10 μg/dL, and PCB levels were consistent with a cumulative, continuing exposure pattern. The median age at menarche for the total sample was 12.2 years. The predicted age at menarche for girls with lead levels above the median (1.2 μg/dL) was 10.5 months later than that for girls with lead levels below the median. In the logistic regression analysis, age was the strongest predictor of menarcheal status and SES was also a significant predictor but BMI was not. The logistic regression analysis that corrected for age, SES, and other pollutants (p,p'-DDE, HCB, mirex, and mercury) indicated that, at their respective geometric means, lead (geometric mean: 0.49 μg/dL) was associated with a significantly lower probability of having reached menarche ( = –1.29) and a group of 4 potentially estrogenic PCB congeners (E-PCB) (geometric mean: 0.12 ppb; International Union of Pure and Applied Chemistry numbers 52, 70, 101 [+90], and 187) was associated with a significantly greater probability of having reached menarche ( = 2.13). Predicted probabilities at different levels of lead and PCBs were calculated on the basis of the logistic regression model. At the respective means of all toxicants and SES, 69% of 12-year-old girls were predicted to have reached menarche. However, at the 75th percentile of lead levels, only 10% of 12-year-old Mohawk girls were predicted to have reached menarche; at the 75th percentile of E-PCB levels, 86% of 12-year-old Mohawk girls were predicted to have reached menarche. No association was observed between mirex, p,p'-DDE, or HCB and menarcheal status. Although BMI was not a significant predictor, we tested BMI in the logistic regression model; it had little effect on the relationships between menarcheal status and either lead or E-PCB. In models testing toxicant interactions, age, SES, lead levels, and PCB levels continued to be significant predictors of menarcheal status. When each toxicant was tested in a logistic regression model correcting only for age and SES, we observed little change in the effects of lead or E-PCB on menarcheal status.
Conclusions. The analysis of multichemical exposure among Akwesasne Mohawk Nation adolescent girls suggests that the attainment of menarche may be sensitive to relatively low levels of lead and certain PCB congeners. This study is distinguished by the ability to test many toxicants simultaneously and thus to exclude effects from unmeasured but coexisting exposures. By testing several PCB congener groupings, we were able to determine that specifically a group of potentially estrogenic PCB congeners affected the odds of reaching menarche. The lead and PCB findings are consistent with the literature and are biologically plausible. The sample size, cross-sectional study design, and possible occurrence of confounders beyond those tested suggest that results should be interpreted cautiously. Additional investigation to determine whether such low toxicant levels may affect reproduction and disorders of the reproductive system is warranted.
Key Words: puberty sexual maturation endocrine disruptors Native Americans Haudenosaunee Iroquois
Abbreviations: HCB, hexachlorobenzene NHANES III, Third National Health and Nutrition Examination Survey PCB, polychlorinated biphenyl POP, persistent organic pollutant p,p'-DDE, dichlorodiphenyldichloroethylene SES, socioeconomic status E-PCB, sum of 4 potentially estrogenic polychlorinated biphenyl congeners
Today, children in many countries are chronically exposed at background levels to a range of common pollutants, including lead and persistent organic pollutants (POPs). POPs, such as mirex, dichlorodiphenyldichloroethylene (p,p'-DDE), hexachlorobenzene (HCB), and polychlorinated biphenyls (PCBs), are environmentally persistent chemicals that tend to be lipophilic and bioaccumulate.1 Lead and POPs have been shown to affect the endocrine system in laboratory animals and in wildlife.2–4
The endocrine-disrupting effects of these environmental pollutants in animal studies have prompted interest in possible reproductive effects.5 Recent observations6 of pubic hair and breast development among many girls <7 years of age, especially black girls, have been viewed by some as evidence of an ongoing secular trend.7 It has been suggested that a secular trend toward earlier maturation in the past few decades has been fueled by exposure to POPs.8 Animal studies indicate that POPs and lead have a variety of adverse effects on reproductive development and functioning.9–13
Comparatively little research has investigated the relationship of pollutants to human sexual maturation. Perinatal exposure to POPs has been associated with earlier age at menarche and pubic hair staging among girls14 and higher rates of precocious puberty,15 as well as delayed sexual maturation among both boys and girls16 and smaller penis size.17 Recent analyses of lead have found delays of menarche and sexual maturation (Tanner staging) among girls with relatively low lead levels.18,19
Although each exposure may produce different effects and most populations are exposed to mixtures of chemical pollutants, to date studies of pollutants and human sexual maturation have not examined the concurrent effects of the most common pollutants to which children may be exposed. Here we assess the relationship between attainment of menarche and PCBs, p,p'-DDE, HCB, mirex, lead, and mercury among Mohawk girls, to determine the effects, if any, of low levels of these pollutants on female sexual maturation.
METHODS
Setting
The study was conducted with residents of the Akwesasne Mohawk Nation. The Nation is a sovereign territory lying on both sides of the St Lawrence River and spanning the boundaries of New York State and Ontario and Quebec, Canada. Since the 1950s, the St Lawrence River has been a site of substantial industrial development. Today, several industrial complexes are in close proximity to Akwesasne, including a US National Priority Superfund site (General Motors Central Foundry Division) and 2 New York State Superfund sites (Reynolds Metal Company and Aluminum Company of America). Exposure assessment studies in the 1990s found some local animal species to have PCB, p,p'-DDE, HCB, and mirex levels above the US Food and Drug Administration human consumption tolerance limits.20,21 Current levels of POPs in this study sample of 10- to 16.9-year-old participants largely reflect prenatal and lactational exposure22 that is consistent with known levels of contaminants among adult women, attributable to past maternal consumption of locally caught, contaminated fish.23 The history of local environmental pollution, the traditional reliance on locally caught fish and game, and the concerns of community members about the health effects of environmental pollutants prompted human health studies at Akwesasne, including an investigation of adolescent development.
Participants
The study participants, recruitment and data collection protocols, laboratory analysis methods, and substitution methods for toxicant levels below the laboratory minimum detection limits were previously described in detail.22 Briefly, all data were collected by Akwesasne community members, without prior knowledge of participants' exposure status. All study protocols were approved by the institutional review board at the University at Albany, State University of New York, and informed consent was obtained from all participants. The target population was defined as residents of Mohawk households located in the Akwesasne Mohawk Nation and in neighboring communities within 10 miles of Akwesasne. Of the households meeting the eligibility requirements for the study, 294 mother/adolescent pairs enrolled in the study. Of these, 19 were lost to follow-up, 1 was found to be ineligible, and 3 had blood samples that were broken in transit, for a final sample size of 271 (140 females and 131 males). Of the 140 eligible female participants who finished the study, 138 had complete data for this analysis.
Data Collection and Variables
Menarche
Menarche was measured as present or absent, based on self-report at the time of the interview and blood sample collection.
Socioeconomic Status
A weighted index of socioeconomic status (SES) was created in consultation with Native American informants from the community. The index used data from interviews with mothers and included: employment, marital status, education, house size and condition, number of motor vehicles, and age of newest motor vehicle.
Pollutants
Fasting blood specimens were collected at first rising by trained Mohawk staff members. Toxicants analyzed were lead, mercury, PCBs, p,p'-DDE, HCB, and mirex. Analyses of PCBs and organochlorine pesticides were conducted at the University at Albany School of Public Health Analytical Laboratory. High-resolution, ultratrace, congener-specific analysis was performed with parallel dual-column (splitless injection) gas chromatography with electron capture detection.24 This method quantitates up to 83 individual PCB congeners and 18 PCB congeners as pairs and triplets, as well as p,p'-DDE, HCB, and mirex. Lead and mercury analyses were conducted by Le Centre de Toxicologie due Quebec (Sainte-Foy, Quebec, Canada). Mercury analysis was based on cold-vapor atomic absorption spectrometry. Lead was analyzed with Zeeman-corrected graphite furnace atomic absorption spectrometry.25
The US Environmental Protection Agency recommended method26 was used to impute values below the minimum detection limits for lead, mercury, PCBs, and HCB, all of which had rates of detection of 50%. Lead, mercury, PCBs, p,p'-DDE, and HCB were natural-logarithmically transformed, because of the skewness of the toxicants. Mirex levels were categorized into 3 groups because of the high rate of undetectable levels, ie, nondetects (below the minimum detection limit of 0.02 ppb; 51.4% of the sample), low detects (0.02–0.03 ppb; 20.3% of the sample), and high detects (0.04–1.17 ppb; 28.3% of the sample).
We considered the 16 PCB congeners with a >50% rate of detection in our sample for hypothesis testing. Three representations of PCB burden were tested, on the basis of the PCB congener groups described by Wolff and co-workers,27,28 ie, estrogenic/neurotoxic, antiestrogenic/dioxin-like, and enzyme-inducing (Table 1).
Statistical Analyses
We used probit and logistic regression to model the probability that study participants reported the presence (scored 1) or absence (scored 0) of menarche and how that probability was altered by changes in relevant independent variables. The motivation for both models can be found in the generalized linear model approach to describing models of binary outcomes. The cumulative distribution functions of the probit and logistic regression models are quite similar and generate predicted probabilities that are almost identical. The choice is often dictated by disciplinary preference rather than inferential properties.29,30 We used the probit model to estimate the median age at menarche but we used logistic regression to interpret the effects of toxicants on the attainment of menarche, because of its greater flexibility in describing model results.
The median age at menarche was estimated to establish the comparability of the sample to other populations, as well as to compare the effect of high versus low lead levels on age at menarche without adjustment for additional toxicants or covariates. The median age at menarche for the sample was estimated with probit analysis31 using the status quo method.32 With the status quo method, age and the presence or absence of menarche are assessed at the time of the interview. The combination of probit analysis and the status quo method is the best method for estimating the median age at menarche for a population, because it eliminates recall bias and solves the problem posed by individuals who have not yet reached menarche.33 Instead of recall of the age at menarche, the probit transformation is applied to the distribution of the percentage of postmenarcheal girls in each age group. The age corresponding to 50% of the probit distribution is the predicted median age at menarche for the population.
Binary logistic regression analysis was used to test the influence of 6 toxicants (mirex, p,p'-DDE, HCB, PCBs, lead, and mercury) on the likelihood (log odds) of having attained menarche (yes/no), controlling for other influences on menarche. The control variables (age, SES, and BMI) tested for inclusion in the model were retained in the model if they reached statistical significance (P < .05). Any variable not included on the basis of statistical significance was reconsidered, to estimate the effects on results and to determine whether its exclusion would alter any observed effects of toxicants. All continuous independent variables were mean-centered. Squared terms were included to test for nonlinear effects. All statistical testing was performed with SPSS software, version 11.5.0 (SPSS, Chicago, IL). An level of .1 was chosen for inclusion of nonlinear terms. An level of .05 was used for testing of all other effects. Squared terms that were maintained in the model and their corresponding variable main effects were tested at several toxicant levels.
In addition, a logistic regression model with age, SES, and only 1 toxicant was run for each toxicant. We compared the results of the single-toxicant models with the model containing all 6 toxicants simultaneously. The purpose of this comparison was to test the influence of additional toxicants in the multiple-toxicant model on the observed effect of a given toxicant.
RESULTS
Toxicant Levels
Toxicant levels for this sample of Mohawk girls are shown in Table 2. Levels were consistent with long-term exposure to a variety of toxicants in multiple media. Mercury levels in the current study were at or below background levels observed for the general population (0.1–0.8 μg/dL).34 The highest lead level was less than one half the Centers for Disease Control and Prevention action limit of 10 μg/dL. PCB levels were consistent with a pattern of both cumulative and recent exposure.22
Probit Analysis
Girls included in this analysis ranged in age from 10 through 16.9 years; 83 (59.7%) had reached menarche. With probit analysis, the median predicted age at menarche for the total sample was 12.2 years (95% confidence interval: 11.9–12.5 years). This group of Mohawk girls was comparable to a larger sample of 10- to 16-year-old American girls from the Third National Health and Nutrition Examination Survey (NHANES III),18 in terms of the distribution of menarcheal status according to age (Table 3).
Girls at or above the median blood lead level of 1.2 μg/dL had a predicted age at menarche of 12.7 years (95% confidence interval: 12.2–13.1 years), without adjustment for any other factors that can influence age at menarche. Girls below the median lead level had a predicted age at menarche of 11.8 years (95% confidence interval: 9.9–12.8 years), 10.6 months earlier than that for girls with higher lead levels.
Logistic Regression
Binary logistic regression analysis predicting menarcheal status (premenarcheal or postmenarcheal) was used to consider simultaneously the effects of multiple toxicants while controlling for other influences on menarche (Table 4). The effect associated with each predictor variable takes into account all other variables in the model. As expected, age was the strongest predictor of attainment of menarche in the binary logistic regression analysis and was positively associated with having reached menarche. Lower SES was related to higher odds of having reached menarche. In other words, same-aged girls with lower SES were more likely to have reached menarche than were girls with higher SES. Mirex, p,p'-DDE, and HCB were unrelated to menarcheal status. The effect of BMI was tested, but BMI was found not to be a significant predictor in the model, perhaps because of its association with age (r = 0.322, P < .001), which is included in the model.
A nonlinear effect was observed for lead, as indicated by the squared term in Table 4. At the geometric mean (0.49 μg/dL) and 75th percentile (1.66 μg/dL), lead was associated with a lower likelihood of having attained menarche, after adjustment for age, SES, and other toxicants. This main effect of lead was not significant below the geometric mean. A 100% increase in lead levels above the mean, from 0.49 to 0.98 μg/dL, decreased the odds of reaching menarche by 72% when all other variables were held constant at their respective means. The odds of reaching menarche decreased by 98% with a doubling of lead levels above the 75th percentile, from 1.66 to 3.32 μg/dL.
The estrogenic PCB (E-PCB) group described by Wolff and co-workers27,28 was associated with a higher likelihood of having reached menarche, after adjustment for age, SES, and other toxicants. A 100% increase in E-PCB levels above the mean, from 0.12 to 0.24 ppb, was associated with 8.4 times greater odds of having reached menarche. Using the same logistic regression analytic technique, we tested the effects of the 2 other classes of PCBs (antiestrogenic and enzyme-inducing) described by Wolff and co-workers.27,28 No relationship was observed between menarche and the PCB groups that Wolff and co-workers27,28 classed as antiestrogenic or enzyme-inducing.
A nonlinear effect of mercury was observed, as indicated by the squared term in Table 4. The main effect of mercury was not significant at the mean and was marginally significant ( = 2.57, P = .08, 2-tailed test; results not shown) at the geometric 95% percentile of mercury levels (0.28 μg/dL).
To substantiate the interpretation of coefficients in the logistic regression model containing all 6 toxicants (Table 4), single-toxicant logistic regression models that corrected only for age and SES were considered. We compared these single-toxicant models with the model shown in Table 4 and found little change in the measure of effect ( coefficients), with the exception of mercury (results not shown). Neither main nor nonlinear effects of mercury were found in the model containing only age, SES, and mercury. Also, adding BMI to the logistic regression model (Table 4) had little effect on the relationships between menarche and either lead or E-PCB, as indicated by the small changes in coefficients (lead: without BMI, = –1.29; with BMI, = –1.10; E-PCB: without BMI, = 2.13; with BMI, = 1.98).
We assessed nonadditivity in the effects of toxicants in the 6-toxicant model. Interactions among all pairs of the 6 toxicants were tested. Three statistically significant interactions were detected, namely, lead and E-PCB, lead and HCB, and lead and mirex (results not shown). In the 3 models with significant interaction terms, age, SES, E-PCB, and lead remained significant at P < .05. To estimate how the effects of toxicants were conditioned by levels of the other toxicants, predicted probability plots were compared, those based on the presented model without interactions (Table 4) and those based on models including interactions. The patterns of effects for each toxicant were indistinguishable across models.
DISCUSSION
Analysis of multichemical exposure among Akwesasne Mohawk adolescent girls provides evidence that lead and certain PCB congeners may affect attainment of menarche. These associations were observed after controlling for potential confounders. No association was observed between mirex, p,p'-DDE, or HCB and menarcheal status.
In this study, lead was associated with a later median predicted age at menarche, without controlling for other variables, as well as a reduced likelihood of having reached menarche, controlling for other toxicants, age, and SES. This finding is consistent with analyses of the NHANES III data that showed 1- to 6-month delays in attainment of menarche and Tanner breast and pubic hair stages among girls 8 to 18 years of age with only modest increases in blood lead levels from 1 to 3 μg/dL.18,19 The NHANES III effects were seen after adjustment for socioeconomic factors and body size, which suggests that lead might have delayed maturation through changes in the endocrine system, rather than as a consequence of delayed growth. Although the effect we observed was larger, it was stable across analyses, both with and without control for SES, BMI and other toxicants. One possible explanation for the larger effect size is our greater control for other toxicant levels, which could have maximized the observed effect of lead.
The plausibility of the observed relationships between higher lead levels and a lower likelihood of having reached menarche and delayed menarche is supported by animal studies showing that lead exposure delayed pubertal development and reduced reproductive organ growth.10–12 Two biological mechanisms might be at work here. Lead might alter sexual maturation through direct effects on all levels of the hypothalamic-pituitary-gonadal axis.3,13 Lead might also indirectly affect sexual maturation by restricting or delaying growth. Several cross-sectional studies found higher lead levels to be associated with decreased weight and height.35–38 Body size, in turn, has been directly related to the timing of sexual maturation.39,40
In addition to the effect of lead, a group of 4 potentially estrogenic PCB congeners was found to be associated with a greater likelihood of having attained menarche, controlling for age, SES, lead, mercury, p,p'-DDE, HCB, and mirex. Few other studies are available for comparison, and among these there is some variation in results. Among breastfed girls, perinatal polybrominated biphenyl exposure was associated with earlier pubic hair staging, and in utero polybrominated biphenyl exposure was associated with earlier menarche.14 Similarly, significantly higher concentrations of p,p'-DDE were found among girls with precocious puberty who immigrated to Belgium from developing countries.15 Among girls in a North Carolina cohort, there was some evidence that PCB exposure was associated with younger ages at attainment of pubertal stages, but results were not conclusive because of the small sample size.41 These associations of POP exposure with earlier sexual maturation are consistent with the present study. However, a study of Flemish adolescents found an association between dioxin-like compounds and delayed attainment of Tanner adult breast stage, and neither PCB levels nor polychlorinated aromatic hydrocarbon exposure was significantly correlated with the timing of menarche or pubic hair staging among girls.14,16,42 These differences in results (delay, acceleration, or no effect) might be attributable to exposure to different types and/or combinations of POPs, which can have estrogenic, androgenic, and antiestrogenic effects. In addition, some of these studies focused on 1 or 2 toxicants or classes of toxicants in their analyses, leaving open the possibility that other covarying but unmeasured toxicants modify the observed effects.
PCBs and related compounds have been associated with several other effects related to the reproductive system, including decreased implantation rate, litter size, and sexual receptivity among animals, as well as reduced fecundity, menstrual irregularity, decreased sperm motility, and hormonal changes among humans.9 Although these effects are consistent with our results, the exact mechanisms underlying the effects of POPs are unclear. Furthermore, >1 pathway may be involved, depending on the combination, concentration, and timing of POP exposure, as well as the species affected. Although the specific mechanisms of effects are not known, modification of the hypothalamic-pituitary-gonadal axis has been proposed as a likely means of action.43–46
The results for mercury reported here cannot be assessed against other research because no investigations of the relationship between mercury levels and sexual maturation among humans have been published. Animal studies indicate the possibility of endocrine and reproductive effects of mercury.47,48
The relationship observed here between lower SES and earlier menarche requires interpretation. Higher SES has been shown to be associated with earlier menarche.49,50 However, this negative relationship was observed in settings in which nutritional status is highly variable and strongly related to socioeconomic factors. Other studies that found no relationship involved well-off populations without threats to maturation from nutritional stress.51,52 The Akwesasne Mohawk sample demonstrates no evidence of nutritional deficiency that would affect menarche,53,54 and the observed relationship between menarche and SES fits the pattern observed for well-off samples.
Here our focus was not on the relationship of SES to menarche but on the role of SES in influencing the relationship of lead and PCBs to menarche. In the present study, the removal of SES from the logistic regression model had little impact on the effects of age and toxicants on menarche (on the basis of changes in coefficients) or on the pseudo-variance explained (2.8% reduction in both Cox-Snell and Nagelkerke pseudo-R2).
This study is distinguished by its ability to consider multiple common toxicants simultaneously and to include congener-specific PCB analysis. Despite relatively low toxicant levels in our sample, both lead and E-PCB were found to be associated with attainment of menarche after controlling for age, SES, and other toxicants. By testing several PCB congener groupings, we were able to determine that specifically a group of potentially estrogenic PCB congeners affected the odds of reaching menarche. Although a nonlinear effect of mercury on menarche was suggested, with the effect strengthening at higher levels, mercury did not reach statistical significance in our model. Additional tests of single-toxicant models suggested that the results observed for lead and E-PCB were robust and not influenced by the other model variables, as indicated by the stable coefficients. Body size had little impact on the associations of lead or E-PCB levels with menarcheal status in this sample. Inclusion of terms representing interactions among the toxicants did not substantially change the effects of each toxicant. The small sample size and very small number of participants at the extremes of the interaction space limit our ability to generalize from analyses of interactions.
CONCLUSIONS
We observed an association between low levels of 4 potentially estrogenic PCB congeners and lead and the attainment of menarche, controlling for SES, age, and other toxicants. These findings are consistent with the literature and are biologically plausible. However, cautious interpretation is warranted because of the sample size, the cross-sectional study design, and the possible occurrence of confounders beyond those tested, including genetic factors. In this, as in other observational studies, the results are derived from a specific range of values, and it is inappropriate to generalize the effects to levels above or below the values in this sample. At much higher or lower levels of lead and/or PCBs, different effects may occur. Findings reported here suggest that the timing of sexual maturation may be sensitive to relatively low levels of lead and certain PCB congeners. Additional investigation to determine whether such low levels may affect reproduction and disorders of the reproductive system is warranted.
ACKNOWLEDGMENTS
This work was supported by grants from the National Institute of Environmental Health Sciences (grants ES04913-09, ES07231-07, and ES10904-05).
We thank the Akwesasne Mohawk community for their cooperation and participation in this research and Dr David O. Carpenter for his leadership in the Superfund Basic Research Program at the University at Albany, State University of New York.
FOOTNOTES
Accepted Sep 29, 2004.
No conflict of interest declared.
REFERENCES
Jones KC, de Voogt P. Persistent organic pollutants (POPs): state of the science. Environ Pollut. 1999;100 :209 –221
Kavlock RJ, Daston GP, DeRosa C, et al. Research needs for the risk assessment of health and environmental effects of endocrine disruptors: a report of the U.S. EPA-sponsored workshop. Environ Health Perspect. 1996;104 :715 –740
Ronis MJ, Gandy J, Badger TM. Endocrine mechanisms underlying reproductive toxicity in the developing rat chronically exposed to dietary lead. J Toxicol Environ Health. 1998;54 :77 –99
Kelce WR, Stone CR, Laws SC, Gray LE Jr, Kemppainen JA, Wilson EM. Persistent DDT metabolite p,p'-DDE is a potent androgen receptor antagonist. Nature. 1995;375 :581 –585
Cooper RL, Kavlock RJ. Endocrine disruptors and reproductive development: a weight-of-evidence overview. J Endocrinol. 1997;152 :159 –166
Herman-Giddens ME, Slora EJ, Wasserman RC, et al. Secondary sexual characteristics and menses in young girls seen in office practice: a study from the Pediatric Research in Office Settings Network. Pediatrics. 1997;99 :505 –512
Herman-Giddens ME, Kaplowitz PB, Wasserman R. Navigating the recent articles on girls' puberty in Pediatrics: what do we know and where do we go from here Pediatrics. 2004;113 :911 –917
Parent A-S, Teilmann G, Juul A, Skakkebaek NE, Toppari J, Bourguignon J-P. The timing of normal puberty and the age limits of sexual precocity: variations around the world, secular trends, and changes after migration. Endocr Rev. 2003;24 :668 –693
Faroon OM, Keith S, Jones D, de Rosa C. Effects of polychlorinated biphenyls on development and reproduction. Toxicol Ind Health. 2001;17 :63 –93
Dearth RK, Hiney JK, Srivastava V, Burdick SB, Bratton GR, Dees WL. Effects of lead (Pb) exposure during gestation and lactation on female pubertal development in the rat. Reprod Toxicol. 2002;16 :343 –352
Corpas I, Castillo M, Marquina D, Benito MJ. Lead intoxication in gestational and lactation periods alters the development of male reproductive organs. Ecotoxicol Environ Saf. 2002;53 :259 –266
McGivern RF, Sokol RZ, Berman NG. Prenatal lead exposure in the rat during the third week of gestation: long-term behavioral, physiological, and anatomical effects associated with reproduction. Toxicol Appl Pharmacol. 1991;110 :206 –215
Ronis MJ, Badger TM, Shema SJ, Roberson PK, Shaikh F. Reproductive toxicity and growth effects in rats exposed to lead at different periods during development. Toxicol Appl Pharmacol. 1996;136 :361 –371
Blanck HM, Marcus M, Tolbert PE, et al. Age at menarche and Tanner stage in girls exposed in utero and postnatally to polybrominated biphenyl. Epidemiology. 2000;11 :641 –647
Krstevska-Konstantinova M, Charlier C, Craen M, et al. Sexual precocity after immigration from developing countries to Belgium: evidence of previous exposure to organochlorine pesticides. Hum Reprod. 2001;16 :1020 –1026
Den Hond E, Roels HA, Hoppenbrouwers K, et al. Sexual maturation in relation to polychlorinated aromatic hydrocarbons: Sharpe and Skakkebaek's hypothesis revisited. Environ Health Perspect. 2002;110 :771 –776
Guo Y-LL, Lai TJ, Ju SH, Chen YC, Hsu CC. Sexual developments and biological findings in Yu-Cheng children. Organohalogen Compounds. 1993;14 :235 –238
Wu T, Buck GM, Mendola P. Blood lead levels and sexual maturation in U.S. girls: the Third National Health and Nutrition Examination Survey, 1988–1994. Environ Health Perspect. 2003;111 :737 –741
Selevan SG, Rice DC, Hogan KA, Euling SY, Pfahles-Hutchens A, Bethel J. Blood lead concentration and delayed puberty in girls. N Engl J Med. 2003;348 :1527 –1536
Forti A, Bogdan KG, Horn E. Health Risk Assessment for the Akwesasne Mohawk Population From Exposure to Chemical Contaminants in Fish and Wildlife. Albany, NY: New York State Department of Health; 1995
Sloan RJ, Jock K. Chemical Contaminants in Fish From the St Lawrence River Drainage on Lands of the Mohawk Nation at Akwesasne and Near the General Motors Corporation/Central Foundry Division, Massena, NY Plant. Albany, NY: New York State Department of Environmental Conservation; 1990
Schell LM, Hubicki LA, DeCaprio AP, et al. Organochlorines, lead, and mercury in Akwesasne Mohawk youth. Environ Health Perspect. 2003;111 :954 –961
Fitzgerald EF, Hwang S-A, Bush B, Cook K, Worswick P. Fish consumption and breast milk PCB concentrations among Mohawk women at Akwesasne. Am J Epidemiol. 1998;148 :164 –172
DeCaprio A, Tarbell AM, Bott A, Wagemaker DL, Williams RL, O'Heir CM. Routine analysis of 101 polychlorinated biphenyl congeners in human serum by parallel dual-column gas chromatography with electron capture detection. J Anal Toxicol. 2000;24 :403 –420
Parsons P, Slavin W. A rapid Zeeman graphite-furnace atomic absorption spectrometric method for the determination of lead in blood. Spectrochim Acta B. 1993;48B :925 –939
US Environmental Protection Agency. Guidance for Data Quality Assessment: Practical Methods for Data Analysis. Washington, DC: Environmental Protection Agency; 1998. Report QA/G-9
Wolff MS, Toniolo PG. Environmental organochlorine exposure as a potential etiologic factor in breast cancer. Environ Health Perspect. 1995;103 :141 –145
Wolff MS, Camann D, Gammon M, Stellman SD. Proposed PCB congener groupings for epidemiological studies. Environ Health Perspect. 1997;105 :13 –14
Liao TF. Interpreting Probability Models: Logit, Probit, and Other Generalized Linear Models. Thousand Oaks, CA: Sage; 1994. Sage University Series on Quantitative Applications in the Social Sciences publication 07–101
Aldrich JH, Nelson FD. Linear Probability, Logit, and Probit Models. Beverly Hills, CA: Sage; 1991. Sage University Series on Quantitative Applications in the Social Sciences publication 07–045
Finney DJ. Probit Analysis: A Statistical Treatment of the Sigmoid Response Curve. 2nd ed. Cambridge, United Kingdom: Cambridge University Press; 1952
Tanner JM. Growth at Adolescence. 2nd ed. Oxford, United Kingdom: Blackwell Scientific Publications; 1962
Marshall WA, Tanner JM. Puberty. In: Falkner F, Tanner JM, eds. Human Growth: A Comprehensive Treatise. 2nd ed. New York, NY: Plenum Press; 1986; 2:171–209
Agency for Toxic Substances and Disease Registry. Toxicological Profile for Mercury. Atlanta, GA: US Department of Health and Human Services; 1999
Bellinger DC, Leviton A, Rabinowitz M, Allred E, Needleman HL, Schoenbaum S. Weight gain and maturity in fetuses exposed to low levels of lead. Environ Res. 1991;54 :151 –158
Shukla R, Dietrich KN, Bornschein RL, Berger O, Hammond PB. Lead exposure and growth in the early preschool child: a follow-up report from the Cincinnati Lead Study. Pediatrics. 1991;88 :886 –892
Frisancho AR, Ryan AS. Decreased stature associated with moderate blood lead concentrations in Mexican-American children. Am J Clin Nutr. 1991;54 :516 –519
Ballew C, Khan LK, Kaufmann R, Mokdad A, Miller DT, Gunter EW. Blood lead concentration and children's anthropometric dimensions in the Third National Health and Nutrition Examination Survey (NHANES III) 1988–1994. J Pediatr. 1999;134 :623 –630
Rubin K, Schirduan V, Gendreau P, Sarfarazi M, Mendola R, Dalsky G. Predictors of axial and peripheral bone mineral density in healthy children and adolescents, with special attention to the role of puberty. J Pediatr. 1993;123 :863 –870
Kaplowitz PB, Slora EJ, Wasserman RC, Pedlow SE, Herman-Giddens ME. Earlier onset of puberty in girls: relation to increased body mass index and race. Pediatrics. 2001;108 :347 –353
Rogan WJ, Ragan NB. Evidence of effects of environmental chemicals on the endocrine system in children. Pediatrics. 2003;112 :247 –252
Staessen JA, Nawrot T, Hond ED, et al. Renal function, cytogenetic measurements, and sexual development in adolescents in relation to environmental pollutants: a feasibility study of biomarkers. Lancet. 2001;357 :1660 –1669
Gore AC. Organochlorine pesticides directly regulate gonadotropin-releasing hormone gene expression and biosynthesis in the GT1–7 hypothalamic cell line. Mol Cell Endocrinol. 2002;192 :157 –170
Jansen HT, Cooke PS, Porcelli J, Liu T-C, Hansen LG. Estrogenic and antiestrogenic actions of PCBs in the female rat: in vitro and in vivo studies. Reprod Toxicol. 1993;7 :237 –248
Hany J, Lilienthal H, Sarasin A, et al. Developmental exposure of rats to a reconstituted PCB mixture or Aroclor 1254: effects on organ weights, aromatase activity, sex hormone levels, and sweet preference behavior. Toxicol Appl Pharmacol. 1999;158 :231 –243
Khan IA, Thomas P. Aroclor 1254-induced alterations in hypothalamic monoamine metabolism in the Atlantic croaker (Micropogonias undulatas): correlation with pituitary gonadotropin release. Neurotoxicology. 1997;18 :553 –560
Hammerschmidt CR, Sandheinrich MB, Wiener JG, Rada RG. Effects of dietary methylmercury on reproduction of fathead minnows. Environ Sci Technol. 2002;36 :877 –883
Matta MB, Linse J, Cairncross C, Francendese L, Kocan RM. Reproductive and transgenerational effects of methylmercury or Aroclor 1268 on Fundulus heteroclitus. Environ Toxicol Chem. 2001;20 :327 –335
Prabhakar AK, Sundaram KR, Ramanujacharyulu TK, Taskar AD. Influence of socio-economic factors on the age at the appearance of different puberty signs. Indian J Med Res. 1972;60 :789 –792
Lee MMC, Chang KSF, Chan MMC. Sexual maturation of Chinese girls in Hong Kong. Pediatrics. 1963;32 :389 –398
Lindgren G. Pubertal stages 1980 of Stockholm schoolchildren. Acta Paediatr. 1996;85 :1365 –1367
Wu T, Mendola P, Buck GM. Ethnic differences in the presence of secondary sex characteristics and menarche among US girls: the Third National Health and Nutrition Examination Survey, 1988–1994. Pediatrics. 2002;110 :752 –757
Gallo MV, Ravenscroft J, Denham M, Schell LM, DeCaprio A, Akwesasne Task Force on the Environment. Environmental contaminants and growth of Mohawk adolescents at Akwesasne. In: Gilli G, Schell LM, Benso L, eds. Human Growth From Conception to Maturity. London, United Kingdom: Smith-Gordon; 2002: 279–287
Ravenscroft J, Gallo MV, Denham M, Akwesasne Task Force on the Environment, Schell LM. Dietary patterns of Mohawk youth. Am J Hum Biol. 2003;16 :220 –221(Melinda Denham, MA, Lawre)
ABSTRACT
Background. Children are commonly exposed at background levels to several ubiquitous environmental pollutants, such as lead and persistent organic pollutants, that have been linked to neurologic and endocrine effects. These effects have prompted concern about alterations in human reproductive development. Few studies have examined the effects of these toxicants on human sexual maturation at levels commonly found in the general population, and none has been able to examine multiple toxicant exposures. The aim of the current investigation was to examine the relationship between attainment of menarche and levels of 6 environmental pollutants to which children are commonly exposed at low levels, ie, dichlorodiphenyldichloroethylene (p,p'-DDE), hexachlorobenzene (HCB), polychlorinated biphenyls (PCBs), mirex, lead, and mercury.
Methods. This study was conducted with residents of the Akwesasne Mohawk Nation, a sovereign territory that spans the St Lawrence River and the boundaries of New York State and Ontario and Quebec, Canada. Since the 1950s, the St Lawrence River has been a site of substantial industrial development, and the Nation is currently adjacent to a US National Priority Superfund site. PCB, p,p'-DDE, HCB, and mirex levels exceeding the US Food and Drug Administration recommended tolerance limits for human consumption have been found in local animal species. The present analysis included 138 Akwesasne Mohawk Nation girls 10 to 16.9 years of age. Blood samples and sociodemographic data were collected by Akwesasne community members, without prior knowledge of participants' exposure status. Attainment of menses (menarche) was assessed as present or absent at the time of the interview. Congener-specific PCB analysis was available, and all 16 PCB congeners detected in >50% of the sample were included in analyses (International Union of Pure and Applied Chemistry numbers 52, 70, 74, 84, 87, 95, 99, 101 [+90], 105, 110, 118, 138 [+163 and 164], 149 [+123], 153, 180, and 187). Probit analysis was used to determine the median age at menarche for the sample. Binary logistic regression analysis was used to determine predictors of menarcheal status. Six toxicants (p,p'-DDE, HCB, PCBs, mirex, lead, and mercury) were entered into the logistic regression model. Age, socioeconomic status (SES), and BMI were tested as potential cofounders and were included in the model at P < .05. Interactions among toxicants were also evaluated.
Results. Toxicant levels were measured in blood for this sample and were consistent with long-term exposure to a variety of toxicants in multiple media. Mercury levels were at or below background levels, all lead levels were well below the Centers for Disease Control and Prevention action limit of 10 μg/dL, and PCB levels were consistent with a cumulative, continuing exposure pattern. The median age at menarche for the total sample was 12.2 years. The predicted age at menarche for girls with lead levels above the median (1.2 μg/dL) was 10.5 months later than that for girls with lead levels below the median. In the logistic regression analysis, age was the strongest predictor of menarcheal status and SES was also a significant predictor but BMI was not. The logistic regression analysis that corrected for age, SES, and other pollutants (p,p'-DDE, HCB, mirex, and mercury) indicated that, at their respective geometric means, lead (geometric mean: 0.49 μg/dL) was associated with a significantly lower probability of having reached menarche ( = –1.29) and a group of 4 potentially estrogenic PCB congeners (E-PCB) (geometric mean: 0.12 ppb; International Union of Pure and Applied Chemistry numbers 52, 70, 101 [+90], and 187) was associated with a significantly greater probability of having reached menarche ( = 2.13). Predicted probabilities at different levels of lead and PCBs were calculated on the basis of the logistic regression model. At the respective means of all toxicants and SES, 69% of 12-year-old girls were predicted to have reached menarche. However, at the 75th percentile of lead levels, only 10% of 12-year-old Mohawk girls were predicted to have reached menarche; at the 75th percentile of E-PCB levels, 86% of 12-year-old Mohawk girls were predicted to have reached menarche. No association was observed between mirex, p,p'-DDE, or HCB and menarcheal status. Although BMI was not a significant predictor, we tested BMI in the logistic regression model; it had little effect on the relationships between menarcheal status and either lead or E-PCB. In models testing toxicant interactions, age, SES, lead levels, and PCB levels continued to be significant predictors of menarcheal status. When each toxicant was tested in a logistic regression model correcting only for age and SES, we observed little change in the effects of lead or E-PCB on menarcheal status.
Conclusions. The analysis of multichemical exposure among Akwesasne Mohawk Nation adolescent girls suggests that the attainment of menarche may be sensitive to relatively low levels of lead and certain PCB congeners. This study is distinguished by the ability to test many toxicants simultaneously and thus to exclude effects from unmeasured but coexisting exposures. By testing several PCB congener groupings, we were able to determine that specifically a group of potentially estrogenic PCB congeners affected the odds of reaching menarche. The lead and PCB findings are consistent with the literature and are biologically plausible. The sample size, cross-sectional study design, and possible occurrence of confounders beyond those tested suggest that results should be interpreted cautiously. Additional investigation to determine whether such low toxicant levels may affect reproduction and disorders of the reproductive system is warranted.
Key Words: puberty sexual maturation endocrine disruptors Native Americans Haudenosaunee Iroquois
Abbreviations: HCB, hexachlorobenzene NHANES III, Third National Health and Nutrition Examination Survey PCB, polychlorinated biphenyl POP, persistent organic pollutant p,p'-DDE, dichlorodiphenyldichloroethylene SES, socioeconomic status E-PCB, sum of 4 potentially estrogenic polychlorinated biphenyl congeners
Today, children in many countries are chronically exposed at background levels to a range of common pollutants, including lead and persistent organic pollutants (POPs). POPs, such as mirex, dichlorodiphenyldichloroethylene (p,p'-DDE), hexachlorobenzene (HCB), and polychlorinated biphenyls (PCBs), are environmentally persistent chemicals that tend to be lipophilic and bioaccumulate.1 Lead and POPs have been shown to affect the endocrine system in laboratory animals and in wildlife.2–4
The endocrine-disrupting effects of these environmental pollutants in animal studies have prompted interest in possible reproductive effects.5 Recent observations6 of pubic hair and breast development among many girls <7 years of age, especially black girls, have been viewed by some as evidence of an ongoing secular trend.7 It has been suggested that a secular trend toward earlier maturation in the past few decades has been fueled by exposure to POPs.8 Animal studies indicate that POPs and lead have a variety of adverse effects on reproductive development and functioning.9–13
Comparatively little research has investigated the relationship of pollutants to human sexual maturation. Perinatal exposure to POPs has been associated with earlier age at menarche and pubic hair staging among girls14 and higher rates of precocious puberty,15 as well as delayed sexual maturation among both boys and girls16 and smaller penis size.17 Recent analyses of lead have found delays of menarche and sexual maturation (Tanner staging) among girls with relatively low lead levels.18,19
Although each exposure may produce different effects and most populations are exposed to mixtures of chemical pollutants, to date studies of pollutants and human sexual maturation have not examined the concurrent effects of the most common pollutants to which children may be exposed. Here we assess the relationship between attainment of menarche and PCBs, p,p'-DDE, HCB, mirex, lead, and mercury among Mohawk girls, to determine the effects, if any, of low levels of these pollutants on female sexual maturation.
METHODS
Setting
The study was conducted with residents of the Akwesasne Mohawk Nation. The Nation is a sovereign territory lying on both sides of the St Lawrence River and spanning the boundaries of New York State and Ontario and Quebec, Canada. Since the 1950s, the St Lawrence River has been a site of substantial industrial development. Today, several industrial complexes are in close proximity to Akwesasne, including a US National Priority Superfund site (General Motors Central Foundry Division) and 2 New York State Superfund sites (Reynolds Metal Company and Aluminum Company of America). Exposure assessment studies in the 1990s found some local animal species to have PCB, p,p'-DDE, HCB, and mirex levels above the US Food and Drug Administration human consumption tolerance limits.20,21 Current levels of POPs in this study sample of 10- to 16.9-year-old participants largely reflect prenatal and lactational exposure22 that is consistent with known levels of contaminants among adult women, attributable to past maternal consumption of locally caught, contaminated fish.23 The history of local environmental pollution, the traditional reliance on locally caught fish and game, and the concerns of community members about the health effects of environmental pollutants prompted human health studies at Akwesasne, including an investigation of adolescent development.
Participants
The study participants, recruitment and data collection protocols, laboratory analysis methods, and substitution methods for toxicant levels below the laboratory minimum detection limits were previously described in detail.22 Briefly, all data were collected by Akwesasne community members, without prior knowledge of participants' exposure status. All study protocols were approved by the institutional review board at the University at Albany, State University of New York, and informed consent was obtained from all participants. The target population was defined as residents of Mohawk households located in the Akwesasne Mohawk Nation and in neighboring communities within 10 miles of Akwesasne. Of the households meeting the eligibility requirements for the study, 294 mother/adolescent pairs enrolled in the study. Of these, 19 were lost to follow-up, 1 was found to be ineligible, and 3 had blood samples that were broken in transit, for a final sample size of 271 (140 females and 131 males). Of the 140 eligible female participants who finished the study, 138 had complete data for this analysis.
Data Collection and Variables
Menarche
Menarche was measured as present or absent, based on self-report at the time of the interview and blood sample collection.
Socioeconomic Status
A weighted index of socioeconomic status (SES) was created in consultation with Native American informants from the community. The index used data from interviews with mothers and included: employment, marital status, education, house size and condition, number of motor vehicles, and age of newest motor vehicle.
Pollutants
Fasting blood specimens were collected at first rising by trained Mohawk staff members. Toxicants analyzed were lead, mercury, PCBs, p,p'-DDE, HCB, and mirex. Analyses of PCBs and organochlorine pesticides were conducted at the University at Albany School of Public Health Analytical Laboratory. High-resolution, ultratrace, congener-specific analysis was performed with parallel dual-column (splitless injection) gas chromatography with electron capture detection.24 This method quantitates up to 83 individual PCB congeners and 18 PCB congeners as pairs and triplets, as well as p,p'-DDE, HCB, and mirex. Lead and mercury analyses were conducted by Le Centre de Toxicologie due Quebec (Sainte-Foy, Quebec, Canada). Mercury analysis was based on cold-vapor atomic absorption spectrometry. Lead was analyzed with Zeeman-corrected graphite furnace atomic absorption spectrometry.25
The US Environmental Protection Agency recommended method26 was used to impute values below the minimum detection limits for lead, mercury, PCBs, and HCB, all of which had rates of detection of 50%. Lead, mercury, PCBs, p,p'-DDE, and HCB were natural-logarithmically transformed, because of the skewness of the toxicants. Mirex levels were categorized into 3 groups because of the high rate of undetectable levels, ie, nondetects (below the minimum detection limit of 0.02 ppb; 51.4% of the sample), low detects (0.02–0.03 ppb; 20.3% of the sample), and high detects (0.04–1.17 ppb; 28.3% of the sample).
We considered the 16 PCB congeners with a >50% rate of detection in our sample for hypothesis testing. Three representations of PCB burden were tested, on the basis of the PCB congener groups described by Wolff and co-workers,27,28 ie, estrogenic/neurotoxic, antiestrogenic/dioxin-like, and enzyme-inducing (Table 1).
Statistical Analyses
We used probit and logistic regression to model the probability that study participants reported the presence (scored 1) or absence (scored 0) of menarche and how that probability was altered by changes in relevant independent variables. The motivation for both models can be found in the generalized linear model approach to describing models of binary outcomes. The cumulative distribution functions of the probit and logistic regression models are quite similar and generate predicted probabilities that are almost identical. The choice is often dictated by disciplinary preference rather than inferential properties.29,30 We used the probit model to estimate the median age at menarche but we used logistic regression to interpret the effects of toxicants on the attainment of menarche, because of its greater flexibility in describing model results.
The median age at menarche was estimated to establish the comparability of the sample to other populations, as well as to compare the effect of high versus low lead levels on age at menarche without adjustment for additional toxicants or covariates. The median age at menarche for the sample was estimated with probit analysis31 using the status quo method.32 With the status quo method, age and the presence or absence of menarche are assessed at the time of the interview. The combination of probit analysis and the status quo method is the best method for estimating the median age at menarche for a population, because it eliminates recall bias and solves the problem posed by individuals who have not yet reached menarche.33 Instead of recall of the age at menarche, the probit transformation is applied to the distribution of the percentage of postmenarcheal girls in each age group. The age corresponding to 50% of the probit distribution is the predicted median age at menarche for the population.
Binary logistic regression analysis was used to test the influence of 6 toxicants (mirex, p,p'-DDE, HCB, PCBs, lead, and mercury) on the likelihood (log odds) of having attained menarche (yes/no), controlling for other influences on menarche. The control variables (age, SES, and BMI) tested for inclusion in the model were retained in the model if they reached statistical significance (P < .05). Any variable not included on the basis of statistical significance was reconsidered, to estimate the effects on results and to determine whether its exclusion would alter any observed effects of toxicants. All continuous independent variables were mean-centered. Squared terms were included to test for nonlinear effects. All statistical testing was performed with SPSS software, version 11.5.0 (SPSS, Chicago, IL). An level of .1 was chosen for inclusion of nonlinear terms. An level of .05 was used for testing of all other effects. Squared terms that were maintained in the model and their corresponding variable main effects were tested at several toxicant levels.
In addition, a logistic regression model with age, SES, and only 1 toxicant was run for each toxicant. We compared the results of the single-toxicant models with the model containing all 6 toxicants simultaneously. The purpose of this comparison was to test the influence of additional toxicants in the multiple-toxicant model on the observed effect of a given toxicant.
RESULTS
Toxicant Levels
Toxicant levels for this sample of Mohawk girls are shown in Table 2. Levels were consistent with long-term exposure to a variety of toxicants in multiple media. Mercury levels in the current study were at or below background levels observed for the general population (0.1–0.8 μg/dL).34 The highest lead level was less than one half the Centers for Disease Control and Prevention action limit of 10 μg/dL. PCB levels were consistent with a pattern of both cumulative and recent exposure.22
Probit Analysis
Girls included in this analysis ranged in age from 10 through 16.9 years; 83 (59.7%) had reached menarche. With probit analysis, the median predicted age at menarche for the total sample was 12.2 years (95% confidence interval: 11.9–12.5 years). This group of Mohawk girls was comparable to a larger sample of 10- to 16-year-old American girls from the Third National Health and Nutrition Examination Survey (NHANES III),18 in terms of the distribution of menarcheal status according to age (Table 3).
Girls at or above the median blood lead level of 1.2 μg/dL had a predicted age at menarche of 12.7 years (95% confidence interval: 12.2–13.1 years), without adjustment for any other factors that can influence age at menarche. Girls below the median lead level had a predicted age at menarche of 11.8 years (95% confidence interval: 9.9–12.8 years), 10.6 months earlier than that for girls with higher lead levels.
Logistic Regression
Binary logistic regression analysis predicting menarcheal status (premenarcheal or postmenarcheal) was used to consider simultaneously the effects of multiple toxicants while controlling for other influences on menarche (Table 4). The effect associated with each predictor variable takes into account all other variables in the model. As expected, age was the strongest predictor of attainment of menarche in the binary logistic regression analysis and was positively associated with having reached menarche. Lower SES was related to higher odds of having reached menarche. In other words, same-aged girls with lower SES were more likely to have reached menarche than were girls with higher SES. Mirex, p,p'-DDE, and HCB were unrelated to menarcheal status. The effect of BMI was tested, but BMI was found not to be a significant predictor in the model, perhaps because of its association with age (r = 0.322, P < .001), which is included in the model.
A nonlinear effect was observed for lead, as indicated by the squared term in Table 4. At the geometric mean (0.49 μg/dL) and 75th percentile (1.66 μg/dL), lead was associated with a lower likelihood of having attained menarche, after adjustment for age, SES, and other toxicants. This main effect of lead was not significant below the geometric mean. A 100% increase in lead levels above the mean, from 0.49 to 0.98 μg/dL, decreased the odds of reaching menarche by 72% when all other variables were held constant at their respective means. The odds of reaching menarche decreased by 98% with a doubling of lead levels above the 75th percentile, from 1.66 to 3.32 μg/dL.
The estrogenic PCB (E-PCB) group described by Wolff and co-workers27,28 was associated with a higher likelihood of having reached menarche, after adjustment for age, SES, and other toxicants. A 100% increase in E-PCB levels above the mean, from 0.12 to 0.24 ppb, was associated with 8.4 times greater odds of having reached menarche. Using the same logistic regression analytic technique, we tested the effects of the 2 other classes of PCBs (antiestrogenic and enzyme-inducing) described by Wolff and co-workers.27,28 No relationship was observed between menarche and the PCB groups that Wolff and co-workers27,28 classed as antiestrogenic or enzyme-inducing.
A nonlinear effect of mercury was observed, as indicated by the squared term in Table 4. The main effect of mercury was not significant at the mean and was marginally significant ( = 2.57, P = .08, 2-tailed test; results not shown) at the geometric 95% percentile of mercury levels (0.28 μg/dL).
To substantiate the interpretation of coefficients in the logistic regression model containing all 6 toxicants (Table 4), single-toxicant logistic regression models that corrected only for age and SES were considered. We compared these single-toxicant models with the model shown in Table 4 and found little change in the measure of effect ( coefficients), with the exception of mercury (results not shown). Neither main nor nonlinear effects of mercury were found in the model containing only age, SES, and mercury. Also, adding BMI to the logistic regression model (Table 4) had little effect on the relationships between menarche and either lead or E-PCB, as indicated by the small changes in coefficients (lead: without BMI, = –1.29; with BMI, = –1.10; E-PCB: without BMI, = 2.13; with BMI, = 1.98).
We assessed nonadditivity in the effects of toxicants in the 6-toxicant model. Interactions among all pairs of the 6 toxicants were tested. Three statistically significant interactions were detected, namely, lead and E-PCB, lead and HCB, and lead and mirex (results not shown). In the 3 models with significant interaction terms, age, SES, E-PCB, and lead remained significant at P < .05. To estimate how the effects of toxicants were conditioned by levels of the other toxicants, predicted probability plots were compared, those based on the presented model without interactions (Table 4) and those based on models including interactions. The patterns of effects for each toxicant were indistinguishable across models.
DISCUSSION
Analysis of multichemical exposure among Akwesasne Mohawk adolescent girls provides evidence that lead and certain PCB congeners may affect attainment of menarche. These associations were observed after controlling for potential confounders. No association was observed between mirex, p,p'-DDE, or HCB and menarcheal status.
In this study, lead was associated with a later median predicted age at menarche, without controlling for other variables, as well as a reduced likelihood of having reached menarche, controlling for other toxicants, age, and SES. This finding is consistent with analyses of the NHANES III data that showed 1- to 6-month delays in attainment of menarche and Tanner breast and pubic hair stages among girls 8 to 18 years of age with only modest increases in blood lead levels from 1 to 3 μg/dL.18,19 The NHANES III effects were seen after adjustment for socioeconomic factors and body size, which suggests that lead might have delayed maturation through changes in the endocrine system, rather than as a consequence of delayed growth. Although the effect we observed was larger, it was stable across analyses, both with and without control for SES, BMI and other toxicants. One possible explanation for the larger effect size is our greater control for other toxicant levels, which could have maximized the observed effect of lead.
The plausibility of the observed relationships between higher lead levels and a lower likelihood of having reached menarche and delayed menarche is supported by animal studies showing that lead exposure delayed pubertal development and reduced reproductive organ growth.10–12 Two biological mechanisms might be at work here. Lead might alter sexual maturation through direct effects on all levels of the hypothalamic-pituitary-gonadal axis.3,13 Lead might also indirectly affect sexual maturation by restricting or delaying growth. Several cross-sectional studies found higher lead levels to be associated with decreased weight and height.35–38 Body size, in turn, has been directly related to the timing of sexual maturation.39,40
In addition to the effect of lead, a group of 4 potentially estrogenic PCB congeners was found to be associated with a greater likelihood of having attained menarche, controlling for age, SES, lead, mercury, p,p'-DDE, HCB, and mirex. Few other studies are available for comparison, and among these there is some variation in results. Among breastfed girls, perinatal polybrominated biphenyl exposure was associated with earlier pubic hair staging, and in utero polybrominated biphenyl exposure was associated with earlier menarche.14 Similarly, significantly higher concentrations of p,p'-DDE were found among girls with precocious puberty who immigrated to Belgium from developing countries.15 Among girls in a North Carolina cohort, there was some evidence that PCB exposure was associated with younger ages at attainment of pubertal stages, but results were not conclusive because of the small sample size.41 These associations of POP exposure with earlier sexual maturation are consistent with the present study. However, a study of Flemish adolescents found an association between dioxin-like compounds and delayed attainment of Tanner adult breast stage, and neither PCB levels nor polychlorinated aromatic hydrocarbon exposure was significantly correlated with the timing of menarche or pubic hair staging among girls.14,16,42 These differences in results (delay, acceleration, or no effect) might be attributable to exposure to different types and/or combinations of POPs, which can have estrogenic, androgenic, and antiestrogenic effects. In addition, some of these studies focused on 1 or 2 toxicants or classes of toxicants in their analyses, leaving open the possibility that other covarying but unmeasured toxicants modify the observed effects.
PCBs and related compounds have been associated with several other effects related to the reproductive system, including decreased implantation rate, litter size, and sexual receptivity among animals, as well as reduced fecundity, menstrual irregularity, decreased sperm motility, and hormonal changes among humans.9 Although these effects are consistent with our results, the exact mechanisms underlying the effects of POPs are unclear. Furthermore, >1 pathway may be involved, depending on the combination, concentration, and timing of POP exposure, as well as the species affected. Although the specific mechanisms of effects are not known, modification of the hypothalamic-pituitary-gonadal axis has been proposed as a likely means of action.43–46
The results for mercury reported here cannot be assessed against other research because no investigations of the relationship between mercury levels and sexual maturation among humans have been published. Animal studies indicate the possibility of endocrine and reproductive effects of mercury.47,48
The relationship observed here between lower SES and earlier menarche requires interpretation. Higher SES has been shown to be associated with earlier menarche.49,50 However, this negative relationship was observed in settings in which nutritional status is highly variable and strongly related to socioeconomic factors. Other studies that found no relationship involved well-off populations without threats to maturation from nutritional stress.51,52 The Akwesasne Mohawk sample demonstrates no evidence of nutritional deficiency that would affect menarche,53,54 and the observed relationship between menarche and SES fits the pattern observed for well-off samples.
Here our focus was not on the relationship of SES to menarche but on the role of SES in influencing the relationship of lead and PCBs to menarche. In the present study, the removal of SES from the logistic regression model had little impact on the effects of age and toxicants on menarche (on the basis of changes in coefficients) or on the pseudo-variance explained (2.8% reduction in both Cox-Snell and Nagelkerke pseudo-R2).
This study is distinguished by its ability to consider multiple common toxicants simultaneously and to include congener-specific PCB analysis. Despite relatively low toxicant levels in our sample, both lead and E-PCB were found to be associated with attainment of menarche after controlling for age, SES, and other toxicants. By testing several PCB congener groupings, we were able to determine that specifically a group of potentially estrogenic PCB congeners affected the odds of reaching menarche. Although a nonlinear effect of mercury on menarche was suggested, with the effect strengthening at higher levels, mercury did not reach statistical significance in our model. Additional tests of single-toxicant models suggested that the results observed for lead and E-PCB were robust and not influenced by the other model variables, as indicated by the stable coefficients. Body size had little impact on the associations of lead or E-PCB levels with menarcheal status in this sample. Inclusion of terms representing interactions among the toxicants did not substantially change the effects of each toxicant. The small sample size and very small number of participants at the extremes of the interaction space limit our ability to generalize from analyses of interactions.
CONCLUSIONS
We observed an association between low levels of 4 potentially estrogenic PCB congeners and lead and the attainment of menarche, controlling for SES, age, and other toxicants. These findings are consistent with the literature and are biologically plausible. However, cautious interpretation is warranted because of the sample size, the cross-sectional study design, and the possible occurrence of confounders beyond those tested, including genetic factors. In this, as in other observational studies, the results are derived from a specific range of values, and it is inappropriate to generalize the effects to levels above or below the values in this sample. At much higher or lower levels of lead and/or PCBs, different effects may occur. Findings reported here suggest that the timing of sexual maturation may be sensitive to relatively low levels of lead and certain PCB congeners. Additional investigation to determine whether such low levels may affect reproduction and disorders of the reproductive system is warranted.
ACKNOWLEDGMENTS
This work was supported by grants from the National Institute of Environmental Health Sciences (grants ES04913-09, ES07231-07, and ES10904-05).
We thank the Akwesasne Mohawk community for their cooperation and participation in this research and Dr David O. Carpenter for his leadership in the Superfund Basic Research Program at the University at Albany, State University of New York.
FOOTNOTES
Accepted Sep 29, 2004.
No conflict of interest declared.
REFERENCES
Jones KC, de Voogt P. Persistent organic pollutants (POPs): state of the science. Environ Pollut. 1999;100 :209 –221
Kavlock RJ, Daston GP, DeRosa C, et al. Research needs for the risk assessment of health and environmental effects of endocrine disruptors: a report of the U.S. EPA-sponsored workshop. Environ Health Perspect. 1996;104 :715 –740
Ronis MJ, Gandy J, Badger TM. Endocrine mechanisms underlying reproductive toxicity in the developing rat chronically exposed to dietary lead. J Toxicol Environ Health. 1998;54 :77 –99
Kelce WR, Stone CR, Laws SC, Gray LE Jr, Kemppainen JA, Wilson EM. Persistent DDT metabolite p,p'-DDE is a potent androgen receptor antagonist. Nature. 1995;375 :581 –585
Cooper RL, Kavlock RJ. Endocrine disruptors and reproductive development: a weight-of-evidence overview. J Endocrinol. 1997;152 :159 –166
Herman-Giddens ME, Slora EJ, Wasserman RC, et al. Secondary sexual characteristics and menses in young girls seen in office practice: a study from the Pediatric Research in Office Settings Network. Pediatrics. 1997;99 :505 –512
Herman-Giddens ME, Kaplowitz PB, Wasserman R. Navigating the recent articles on girls' puberty in Pediatrics: what do we know and where do we go from here Pediatrics. 2004;113 :911 –917
Parent A-S, Teilmann G, Juul A, Skakkebaek NE, Toppari J, Bourguignon J-P. The timing of normal puberty and the age limits of sexual precocity: variations around the world, secular trends, and changes after migration. Endocr Rev. 2003;24 :668 –693
Faroon OM, Keith S, Jones D, de Rosa C. Effects of polychlorinated biphenyls on development and reproduction. Toxicol Ind Health. 2001;17 :63 –93
Dearth RK, Hiney JK, Srivastava V, Burdick SB, Bratton GR, Dees WL. Effects of lead (Pb) exposure during gestation and lactation on female pubertal development in the rat. Reprod Toxicol. 2002;16 :343 –352
Corpas I, Castillo M, Marquina D, Benito MJ. Lead intoxication in gestational and lactation periods alters the development of male reproductive organs. Ecotoxicol Environ Saf. 2002;53 :259 –266
McGivern RF, Sokol RZ, Berman NG. Prenatal lead exposure in the rat during the third week of gestation: long-term behavioral, physiological, and anatomical effects associated with reproduction. Toxicol Appl Pharmacol. 1991;110 :206 –215
Ronis MJ, Badger TM, Shema SJ, Roberson PK, Shaikh F. Reproductive toxicity and growth effects in rats exposed to lead at different periods during development. Toxicol Appl Pharmacol. 1996;136 :361 –371
Blanck HM, Marcus M, Tolbert PE, et al. Age at menarche and Tanner stage in girls exposed in utero and postnatally to polybrominated biphenyl. Epidemiology. 2000;11 :641 –647
Krstevska-Konstantinova M, Charlier C, Craen M, et al. Sexual precocity after immigration from developing countries to Belgium: evidence of previous exposure to organochlorine pesticides. Hum Reprod. 2001;16 :1020 –1026
Den Hond E, Roels HA, Hoppenbrouwers K, et al. Sexual maturation in relation to polychlorinated aromatic hydrocarbons: Sharpe and Skakkebaek's hypothesis revisited. Environ Health Perspect. 2002;110 :771 –776
Guo Y-LL, Lai TJ, Ju SH, Chen YC, Hsu CC. Sexual developments and biological findings in Yu-Cheng children. Organohalogen Compounds. 1993;14 :235 –238
Wu T, Buck GM, Mendola P. Blood lead levels and sexual maturation in U.S. girls: the Third National Health and Nutrition Examination Survey, 1988–1994. Environ Health Perspect. 2003;111 :737 –741
Selevan SG, Rice DC, Hogan KA, Euling SY, Pfahles-Hutchens A, Bethel J. Blood lead concentration and delayed puberty in girls. N Engl J Med. 2003;348 :1527 –1536
Forti A, Bogdan KG, Horn E. Health Risk Assessment for the Akwesasne Mohawk Population From Exposure to Chemical Contaminants in Fish and Wildlife. Albany, NY: New York State Department of Health; 1995
Sloan RJ, Jock K. Chemical Contaminants in Fish From the St Lawrence River Drainage on Lands of the Mohawk Nation at Akwesasne and Near the General Motors Corporation/Central Foundry Division, Massena, NY Plant. Albany, NY: New York State Department of Environmental Conservation; 1990
Schell LM, Hubicki LA, DeCaprio AP, et al. Organochlorines, lead, and mercury in Akwesasne Mohawk youth. Environ Health Perspect. 2003;111 :954 –961
Fitzgerald EF, Hwang S-A, Bush B, Cook K, Worswick P. Fish consumption and breast milk PCB concentrations among Mohawk women at Akwesasne. Am J Epidemiol. 1998;148 :164 –172
DeCaprio A, Tarbell AM, Bott A, Wagemaker DL, Williams RL, O'Heir CM. Routine analysis of 101 polychlorinated biphenyl congeners in human serum by parallel dual-column gas chromatography with electron capture detection. J Anal Toxicol. 2000;24 :403 –420
Parsons P, Slavin W. A rapid Zeeman graphite-furnace atomic absorption spectrometric method for the determination of lead in blood. Spectrochim Acta B. 1993;48B :925 –939
US Environmental Protection Agency. Guidance for Data Quality Assessment: Practical Methods for Data Analysis. Washington, DC: Environmental Protection Agency; 1998. Report QA/G-9
Wolff MS, Toniolo PG. Environmental organochlorine exposure as a potential etiologic factor in breast cancer. Environ Health Perspect. 1995;103 :141 –145
Wolff MS, Camann D, Gammon M, Stellman SD. Proposed PCB congener groupings for epidemiological studies. Environ Health Perspect. 1997;105 :13 –14
Liao TF. Interpreting Probability Models: Logit, Probit, and Other Generalized Linear Models. Thousand Oaks, CA: Sage; 1994. Sage University Series on Quantitative Applications in the Social Sciences publication 07–101
Aldrich JH, Nelson FD. Linear Probability, Logit, and Probit Models. Beverly Hills, CA: Sage; 1991. Sage University Series on Quantitative Applications in the Social Sciences publication 07–045
Finney DJ. Probit Analysis: A Statistical Treatment of the Sigmoid Response Curve. 2nd ed. Cambridge, United Kingdom: Cambridge University Press; 1952
Tanner JM. Growth at Adolescence. 2nd ed. Oxford, United Kingdom: Blackwell Scientific Publications; 1962
Marshall WA, Tanner JM. Puberty. In: Falkner F, Tanner JM, eds. Human Growth: A Comprehensive Treatise. 2nd ed. New York, NY: Plenum Press; 1986; 2:171–209
Agency for Toxic Substances and Disease Registry. Toxicological Profile for Mercury. Atlanta, GA: US Department of Health and Human Services; 1999
Bellinger DC, Leviton A, Rabinowitz M, Allred E, Needleman HL, Schoenbaum S. Weight gain and maturity in fetuses exposed to low levels of lead. Environ Res. 1991;54 :151 –158
Shukla R, Dietrich KN, Bornschein RL, Berger O, Hammond PB. Lead exposure and growth in the early preschool child: a follow-up report from the Cincinnati Lead Study. Pediatrics. 1991;88 :886 –892
Frisancho AR, Ryan AS. Decreased stature associated with moderate blood lead concentrations in Mexican-American children. Am J Clin Nutr. 1991;54 :516 –519
Ballew C, Khan LK, Kaufmann R, Mokdad A, Miller DT, Gunter EW. Blood lead concentration and children's anthropometric dimensions in the Third National Health and Nutrition Examination Survey (NHANES III) 1988–1994. J Pediatr. 1999;134 :623 –630
Rubin K, Schirduan V, Gendreau P, Sarfarazi M, Mendola R, Dalsky G. Predictors of axial and peripheral bone mineral density in healthy children and adolescents, with special attention to the role of puberty. J Pediatr. 1993;123 :863 –870
Kaplowitz PB, Slora EJ, Wasserman RC, Pedlow SE, Herman-Giddens ME. Earlier onset of puberty in girls: relation to increased body mass index and race. Pediatrics. 2001;108 :347 –353
Rogan WJ, Ragan NB. Evidence of effects of environmental chemicals on the endocrine system in children. Pediatrics. 2003;112 :247 –252
Staessen JA, Nawrot T, Hond ED, et al. Renal function, cytogenetic measurements, and sexual development in adolescents in relation to environmental pollutants: a feasibility study of biomarkers. Lancet. 2001;357 :1660 –1669
Gore AC. Organochlorine pesticides directly regulate gonadotropin-releasing hormone gene expression and biosynthesis in the GT1–7 hypothalamic cell line. Mol Cell Endocrinol. 2002;192 :157 –170
Jansen HT, Cooke PS, Porcelli J, Liu T-C, Hansen LG. Estrogenic and antiestrogenic actions of PCBs in the female rat: in vitro and in vivo studies. Reprod Toxicol. 1993;7 :237 –248
Hany J, Lilienthal H, Sarasin A, et al. Developmental exposure of rats to a reconstituted PCB mixture or Aroclor 1254: effects on organ weights, aromatase activity, sex hormone levels, and sweet preference behavior. Toxicol Appl Pharmacol. 1999;158 :231 –243
Khan IA, Thomas P. Aroclor 1254-induced alterations in hypothalamic monoamine metabolism in the Atlantic croaker (Micropogonias undulatas): correlation with pituitary gonadotropin release. Neurotoxicology. 1997;18 :553 –560
Hammerschmidt CR, Sandheinrich MB, Wiener JG, Rada RG. Effects of dietary methylmercury on reproduction of fathead minnows. Environ Sci Technol. 2002;36 :877 –883
Matta MB, Linse J, Cairncross C, Francendese L, Kocan RM. Reproductive and transgenerational effects of methylmercury or Aroclor 1268 on Fundulus heteroclitus. Environ Toxicol Chem. 2001;20 :327 –335
Prabhakar AK, Sundaram KR, Ramanujacharyulu TK, Taskar AD. Influence of socio-economic factors on the age at the appearance of different puberty signs. Indian J Med Res. 1972;60 :789 –792
Lee MMC, Chang KSF, Chan MMC. Sexual maturation of Chinese girls in Hong Kong. Pediatrics. 1963;32 :389 –398
Lindgren G. Pubertal stages 1980 of Stockholm schoolchildren. Acta Paediatr. 1996;85 :1365 –1367
Wu T, Mendola P, Buck GM. Ethnic differences in the presence of secondary sex characteristics and menarche among US girls: the Third National Health and Nutrition Examination Survey, 1988–1994. Pediatrics. 2002;110 :752 –757
Gallo MV, Ravenscroft J, Denham M, Schell LM, DeCaprio A, Akwesasne Task Force on the Environment. Environmental contaminants and growth of Mohawk adolescents at Akwesasne. In: Gilli G, Schell LM, Benso L, eds. Human Growth From Conception to Maturity. London, United Kingdom: Smith-Gordon; 2002: 279–287
Ravenscroft J, Gallo MV, Denham M, Akwesasne Task Force on the Environment, Schell LM. Dietary patterns of Mohawk youth. Am J Hum Biol. 2003;16 :220 –221(Melinda Denham, MA, Lawre)