Lead, a ubiquitous environmental toxicant, is associated with a variety of adverse health effects (Bellinger 2011). Infants, children, and pregnant women are the most vulnerable to lead toxicity (National Research Council (U.S.). Committee on Measuring Lead in Critical Populations. et al. 1993). A growing body of research has shown that elevated blood lead levels in utero and during childhood are associated with impaired nervous, reproductive and cardiovascular systems (Silbergeld 1990; WHO 2010). Although blood lead levels have decreased over time, mainly due to the phase-out of leaded gasoline (Caravanos et al. 2014; Falk 2003), other routes of exposure still exist and pose a continued public health burden, especially in developing countries such as Mexico (Falk 2003; Schnur and John 2014) and many US inner cities and underserved areas (Bellinger 2016). Moreover, bone lead accumulated from years of exposure is known to both persist for many years and, in pregnant women, serve as a source of prenatal exposure due to the marked increase in bone turnover that is known to occur during pregnancy (Hu and Hernandez-Avila 2002).

Previous studies have shown that lead is readily transferred from mother to offspring by crossing the placental-fetal barrier during gestation and via breast milk during lactation (Koyashiki et al. 2010; WHO 2010). Many existing studies have investigated the association between early-life lead exposure and neurodevelopment in children (Bellinger et al. 1987; Sanders et al. 2009; Schnaas et al. 2006) with a few studies that have linked lead exposure to delayed pubertal timing. Animal studies found that prenatal and early-life exposure to lead resulted in delayed pubertal onset by decreasing serum hormones including estradiol, luteinizing hormone (LH) and insulinlike growth factor-1 (IGF-1) (Dearth et al. 2002; Dearth et al. 2004; Ronis et al. 1996). However, fewer analyses have characterized the potential role of lead in pubertal development across sensitive periods of exposure in human.

In girls, it has been shown that blood lead (3 μg/dL versus 1 μg/dL) was associated with a later age at menarche in African-Americans, delayed breast maturation and pubic hair growth in both African-Americans and Mexican-Americans (Selevan et al. 2003). Blood lead levels (> 2 μg/dL) were also found to be significantly associated with delayed female pubic hair growth and attainment of menarche in another U.S. investigation (Wu et al. 2003). Blood lead concentration higher than 5 μg/dL has been related to a delay in pubertal development for breast maturation, pubic hair growth, and menarche in 13-year-old South African girls (Naicker et al. 2010). In contrast, two studies provide no evidence for the negative associations in U.S. girls at age 9 (Wolff et al. 2008), and in Polish girls at age 7–16 (Slawinska et al. 2012). The evidence on the role of lead exposure on pubertal development in boys is rare. Two Russian investigations revealed a delayed pubertal onset of genitalia in boys aged 8–9 years (Hauser et al. 2008) and testicular volume (>3mL) in a later follow-up study of the same boys (Williams et al. 2010).

In this study, we aimed at addressing research gaps including the lack of studies that characterize multiple sensitive periods of exposure using a longitudinal design and the studies in boys by examining the prospective association of prenatal and early-life lead exposure with secondary sexual characteristics in girls and boys at age 9.8–18.0 years in Mexico City.

Our final analysis included 547 mother-offspring pairs with a total of 283 girls and 264 boys (Table 1). The mean age was 14.5 years, ranging from 9.8 to 18.0 years. In the total sample, the interquartile range (IQR) for maternal patella and tibia and cumulative early-childhood lead was 13.57 μg/g, 13.3 μg/g and 7.66 μg/dL, respectively. Maternal bone lead concentration in the patella at 1 month postpartum was appreciably higher than in the tibia among girls, but not boys (female: median patella 8.2 μg/g; tibia 7.6 μg/g). The median of cumulative early-childhood blood lead was 14.3 μg/dL and 13.8 for boys and girls, respectively. The 3 lead biomarkers were correlated (p <0.0001) (Supplemental Table 1). The Spearman correlation between maternal patella and tibia lead was moderate at 0.48; the correlation was 0.18-0.19 between maternal bone lead (tibia or patella) and cumulative early-childhood blood lead.

Among 264 male participants, 19.1% were at stage 1 (pre-pubertal) and 21.1% were at stage 5 (adult level) for pubic hair growth; 5.6% and 22.7% were at stage 1 and 5 for genitalia, respectively (Table 2). There were 176 boys (70.1%) who had matured testicular volume (≥ 20mL). Among 285 female participants, 7.3% were at stage 1 and 19.6% were at stage 5 for pubic hair growth; 4.3% and 23.3% were at stage 1 and 5 for breast maturation, respectively. Among 238 girls (84.4%) who had attained menarche, the mean (SD) age at menarche was 11.6 (SD=1.2).

We report the associations between one-IQR higher prenatal and early-life lead concentrations and pubertal development in boys (Table 3) and girls (Table 4) using separate models. No significant associations were found in boys between any given lead biomarker and puberty estimated by Tanner stage or testicular volume (Figure 1). In girls, maternal bone lead and early childhood blood lead were negatively associated with puberty. In the fully adjusted models, the odds of reaching a high stage of breast maturation versus combined low and middle stages decreased by 28% (OR: 0.72, 95% CI: 0.51, 1.00, p=0.048) per IQR increase in patella bone lead, and decreased by 30% (OR: 0.70, 95% CI: 0.53, 0.93, p=0.013) per IQR increase in cumulative 1-4 years blood lead (Figure 2). The cumulative early-life blood lead was also negatively associated with pubic hair growth with an OR=0.68 (95% CI: 0.51, 0.90, p=0.006). Maternal tibia lead level was not associated with any indicator of pubertal development in girls.

In addition, after controlling for number of siblings at birth, maternal education and marital status, we found that girls with maternal patella lead in the 3^rd tertile (range: 13.0-45.3 μg/g) had a later age at menarche compared with girls in the 1^st tertile (<3.9 μg/g) (HR=0.60, 95% CI: 0.41–0.88, p=0.008) (Table 5; Figure 3). In the adjusted models, girls with cumulative 1-4 years blood lead level in the 2^nd tertile (range: 12.1-16.1 μg/dL) had a later age at menarche compared with girls in the 1^st tertile (<12 μg/dL) (HR=0.65, 95% CI: 0.46–0.91, p=0.013).

In sensitivity analyses, additional adjustment for BMI z-score and height did not significantly change the association of pubertal development with any lead biomarker (Supplemental Table 2-4). The change of the magnitude of these associations ranged from 0% to 7.7% in girls, and from 3% to 17.2% in boys; the associations of puberty with maternal patella and cumulative early-life lead remained significant, although the association between maternal patella and breast maturation became marginally significant (p = 0.09). All the associations of lead biomarkers with puberty remained non-significant in boys. Further adjustment for cumulative early-life lead, we found that the parameter estimates of maternal patella lead in all the models were not appreciably changed (Supplemental Table 2 and 4).

This is the first epidemiological study to investigate the prospective associations of cumulative prenatal (as measured by maternal bone lead at 1 month postpartum) and cumulative early-life lead exposure (as measured by cumulative blood lead 1-4 years) with physical markers of pubertal development in Mexican boys and girls aged 9.8-18.0 years. In the present analysis, higher prenatal and early-life lead exposure were associated with a significant delay in pubertal development among girls. No associations were observed among boys. In the sensitivity analyses, further adjustment for additional potential confounders (e.g. child growth and body size) did not alter these associations.

Our findings are consistent with previous animal studies on female rats measuring the effects of early life lead exposure on pubertal development (Dearth et al. 2002; Dearth et al. 2004; Ronis et al. 1996) and a few cross-sectional analyses performed in girls similarly suggesting a negative association between childhood lead exposure and pubertal development (Naicker et al. 2010; Selevan et al. 2003; Wu et al. 2003). However, in the present analysis, no association between a given lead marker and any physical markers of puberty was found in boys, which is inconsistent with previous findings among Russian boys (Hauser et al. 2008; Williams et al. 2010). The use of different study populations, exposure periods, pubertal stages and statistical analyses may contribute to these inconsistencies. For example, the 2 Russian studies in boys examined the onset of puberty (Tanner stage>1) and testicular volume > 3mL in much younger boys (at age 8-12 years), while our study evaluated the pubertal development from stage 1-5 (majority of the boys had Tanner stage>2) and testicular volume ≥ 20mL in older boys aged 10-17 years. The Russian investigations examined lead exposure levels in middle childhood, while we examined much earlier developmental periods during pregnancy and early-life. Furthermore, the significant associations between middle childhood lead exposure and pubertal onset were only observed in the high blood lead group (≥5μg/dL) (Hauser et al. 2008; Williams et al. 2010); the significance disappeared when treating blood lead levels as continuous variables. Therefore, it is possible that the effect of lead is more profound on pubertal onset compared with advanced pubertal stages in boys or male pubertal stages or sex hormones (e.g. testosterone) can only be affected by high lead exposure (≥5μg/dL). In addition, previous studies have found that the concentrations of pubertal hormones (follicle stimulating hormone (FSH) and LH) were much higher in female fetuses than in male fetuses at mid-pregnancy (Debieve et al. 2000). It is possible that the effect of prenatal exposure to lead on sex hormones is more obvious in girls than boys, which eventually advances pubertal development in girls. Further studies evaluating all the sensitive periods are needed to confirm the inconsistent findings.

Furthermore, only patella lead was significantly associated with delayed pubertal development while tibia lead was not. It is possible that these findings are due to the smaller sample size of tibia versus patella measurements, or the difference may be caused by the structure of the two bone sites with the patella, which is made mostly of trabecular bone (versus tibia, which is made mostly of cortical bone) being more vascularized, reflecting more recent lead exposure; in other words, the accumulation of lead is slower in the tibia, and the lead in the patella may also be more bioavailable for mobilization during pregnancy (Afeiche et al. 2011; Hu et al. 1989; Hu et al. 1998).

The negative associations of pubertal development with prenatal and early-life lead exposure in girls in our study are biologically plausible. It has been suggested that early lead exposure can cause decreased pubertal hormones, which subsequently lead to delayed pubertal development (Dearth et al. 2002; Dearth et al. 2004; Ronis et al. 1996). Specifically, early lead exposure can reduce the concentrations of estradiol, LH, and IGF-1 in serum, and then cause a delay in the timing of vaginal opening and the attainment of first estrus in rodents (Dearth et al. 2002; Dearth et al. 2004; Ronis et al. 1996). In addition to animal studies, a cross-sectional analysis of U.S. girls found that blood lead levels (≥ 5 μg/dL versus < 1 μg/dL) were associated with lower serum inhibin-B levels, which was used as an indicator of follicular development (Gollenberg et al. 2010). Since inhibin B is produced from the granulosa cells in the ovaries, it is possible that lead may also have a direct effect on ovaries (Gollenberg et al. 2010). Both developmental stages (prenatal and early-life) can be considered as sensitive periods for pubertal development. It has been found that the hypothalamic-pituitary-gonadal (HPG) axis is active in mid-gestation and early infancy (Liu 2017). In the 2^nd trimester, with the increased activity of HPG axis, the concentrations of FSH and LH rise to the amount that is contributing to development of ovaries (Mueller 2013). In early infancy, FSH and LH further increase and reach the peak concentrations. During this period, sex hormones (e.g. estradiol) reach levels that are comparable to the levels at early-middle puberty, which is defined as “mini-puberty of infancy” (Kurtoglu and Bastug 2014). Hence, lead may have an impact on the development of ovaries and sex hormones in pregnancy and early-life, respectively, which may independently affect female pubertal development in later life.

This study was somewhat limited by the sample size for modeling ordinal outcomes. In addition, the mediating effects of serum sex and growth hormones were not evaluated. Given that our study sample represents low-to-middle class among Mexican children, the results may not be generalizable to other socioeconomic populations or populations with different ethnic composition. As with any investigations that report previous events, our study is subject to bias in recall due to the use of self-reported age at menarche. However, previous analyses have shown that the actual age at menarche was highly correlated with self-reported menarcheal age within 5 years of follow-up in peripubertal girls (Koprowski et al. 2001) and 30 years in middle-aged women with baseline age at 7-9 years (Must et al. 2002). The strengths of this study include the ability to examine the impact of lead at multiple life stages on pubertal development, the use of physician-assessed pubertal stages, the measurement of testicular volume as an additional pubertal marker, which has barely been used in previous studies, and the use of maternal bone lead representing cumulative measure of prenatal lead exposure. Given that lead can accumulate in bones with a long half-life ranging from 10 to 30 years, lead exposure can persist for long periods after external exposure sources have ceased (Hu 1998; Zhang et al. 2012). Mobilization of cumulative bone lead stores during pregnancy into fetal circulation contributes significantly to the fetal exposure to lead (Hu et al. 1998; Tellez-Rojo et al. 2004; Zhang et al. 2012). Therefore, maternal bone lead may be more reflective of lead exposure during sensitive periods, which have been associated with disrupted development of fetus and infants (Tellez-Rojo et al. 2004).

We found that higher prenatal and early-life exposure to lead were associated with delayed breast maturation, pubic hair growth and later age at menarche in girls. Future research including mediators such as serum hormonal markers is needed to confirm our results and to deepen our understanding of the mechanism involved in the association between lead exposure during sensitive periods and the neuroendocrine system.