Puberty is a life stage with rapid physiological changes, including growth spurt and maturation of the gonads and the brain. Variations in pubertal timing and progression have been related to several antecedents, including genetic and environmental stressors [1]. Particularly, pubertal development has been reported to be accompanied with high vulnerability of the neuroendocrine system to environmental factors, which may lead to long-term effects on subsequent development [1]. One of the environmental stressors to developing children is bisphenol A (BPA) [2, 3]. BPA is a plasticizer, which is extensively used in resin-based and polycarbonate plastic products, including dental sealants and containers for foods and beverages. BPA can leach from a variety of consumer products and lead to wide spread human exposure [4].

Animal studies have shown that BPA acts as an endocrine disruptor with both estrogenic and anti-androgenic effects [2, 5]. The endocrine disrupting properties of BPA have also been demonstrated in human studies. These studies suggest that BPA is associated with increased risk of obesity [6], polycystic ovarian syndrome [7], recurrent miscarriage [8] and male infertility [2]. As pubertal growth is a hormone-dependent process [9], it is, therefore, plausible that chemicals with endocrine disrupting effects, like BPA, may induce altered pubertal maturation tempo. BPA is suggested to disrupt the key hormones responsible for sexual maturation [10]. Studies in mice have indicated that BPA exposure is associated with accelerated vaginal opening [3, 11, 12], but delayed vaginal opening [13] and testicular descent [14] have also been reported.

Pubertal development in humans involves both adrenarche and gonadarche, two independent processes. Adrenarche is induced by adrenal androgens, which are responsible for the initiation of pubic hair growth, as well as other pubertal events including acne, oily skin, deepened voice and adult-type body odor. Gonadarche refers to the reactivation of the hypothalamic-pituitary, and is characterized by the appearance of breast tissue. Longitudinal studies have shown that the pubertal development for girls may start with breast tissue appearance (thelarche) or pubic hair growth (pubarche), or concordant maturation of breasts and appearance of pubic hair [15, 16].

A few studies have explored the effect of BPA exposure on pubertal development in girls, but most of them used menarche age as the only endpoint for pubertal development assessment [17–19]. However, as mentioned above, breast development and the presence of pubic hair are known to be the early markers of pubertal maturation, while the onset of menarche occurs approximately 2 years after breast development. In addition, earlier pubertal onset may not necessarily be followed by an earlier age of menarche [20] due to the compensatory delay, which indicates both earlier and later events of pubertal development should be included to examine the effect of exogenous endocrine disrupting chemicals like BPA.

In the present study, we examined the profile of pubertal development in relation to BPA exposure among school-aged girls using menarche and Tanner stages, through which puberty onset and progression were evaluated.

The present study was an ancillary study to a large national study of pubertal development and health of adolescents in schools in Jiading District, Shanghai, China, which had already collected anthropometric measures and information related to pubertal development. The current study added collection of urine samples. Therefore, participants were not aware of the specific hypothesis of the study. A detailed description of our study has been published elsewhere [6]. The following are brief descriptions of the study elements relevant to the present study.

The mean age of 655 participants was 12.9 ± 2.7 years, which was younger than the 88 girls who were not included in the analysis (15.8 ± 2.4 years).

The median urinary BPA concentration was 1.24 (interquartile range: LOD-4.80) μg/gCr in girls. About 60% of the urinary samples had concentrations of BPA above LOD (0.31 μg/L). The median BPA concentration level among those with detected BPA was 3.75 μg/gCr, which was used as the cut-point of higher and lower BPA exposure.

In this cross-sectional study of school-aged girls, we found that BPA exposure was associated with delayed menarche, an indicator of mid-puberty. We also found a non-significant tendency that BPA exposure was associated with earlier pubic hair onset, and delayed progression of pubic hair development.

The association between BPA exposure and delayed menarche found in the present study was consistent with a previous study based on NHANES data, which reported moderate BPA exposure was associated with delayed menarche, although the association was non-significant [19]. Timing of menarche is influenced by many factors, including genetic background, ethnicity, obesity, nutrition and physical activity. In the last few decades, more and more studies have recognized EDCs, including dioxins, phthalates, organohalogens and PCBs, as important contributors influencing the timing of menarche, although the findings have not always been consistent. In the present study, we did not observe a stronger association among girls with high BPA exposure compared with those with moderate exposure, which is consistent with the report of the nonmonotonic dose–response effect of EDCs, including BPA [30]. The onset of menarche occurs approximately 2 years after breast development. In studies carried out in Spain, the United States, and Greece, early maturing girls were found to present a compensatory delay in pubertal progression [31–33]. These observations were consistent with our finding that those with high BPA exposure were more likely to reach pubic hair onset earlier while having a later menarche age and pubic hair stage 5.

The association of BPA exposure with early pubic hair onset, although not statistically significant, is consistent with the effect observed on other estrogen-like endocrine disruptors such as phthalates [1]. It was reported that pubertal hair development may rely on adrenal androgens, chiefly dehydroepiandrosterone (DHEA) [34], while breast development is more likely to be attributed to the effect of estrogen [35]. Higher BPA has been found to be associated with increased DHEA levels in both pre-adolescent girls and adult females [36, 37], which provides a potential underlying mechanism for the present finding. The result is also in line with the finding that BPA exposure is associated with polycystic ovarian syndrome [7], an abnormality characterized by the presence of hyper androgenism [38].

Although both breast development and presence of pubic hair are known to be early markers of pubertal maturation, breast development is more likely to be confounded by excess adiposity [39], which may also contribute to the non-significant association between BPA exposure and breast development.

It has been reported that prenatal growth is sensitive to extrogenous EDCs exposure. Our study suggests that pubertal development may also be sensitive to adverse effects of BPA [2]. Perinatal BPA exposure is associated with fetal growth as well as the growth pattern after birth, including pubertal growth. The present study is limited by the fact that perinatal exposure was not collected and thus its contribution to pubertal development, as well as its interaction with peripubertal exposure, cannot be evaluated. Longitudinal studies examining both perinatal and peripubertal BPA exposure are needed to clarify the effect.

The study has several strengths. First, we utilized both menarche and Tanner Staging, a well-validated tool for the assessment of pubertal stage, which provides more information than most studies with menarche as the only endpoint. Second, we were able to control for many potential confounders including depression status, sports activity strength, and dietary pattern.

Our study also has limitations. First, the cross-sectional design limited our ability to examine a causal relationship due to difficulties with temporality. Second, the sample size in each age group was small and the power to detect a difference, if it existed, was limited. Around 12% of girls were not included in the analysis due to refusal or loss of urine samples. As the study was an ancillary study to a national survey, eligible girls were unaware of the specific hypothesis of the study, thus selection bias is less likely to be a problem. Third, we collected only one single spot urine to reflect the BPA exposure. BPA is metabolized relatively rapidly, [40]. One single urine BPA measurement may not reflect average BPA exposures. Although girls may be exposed to BPA via oral ingestion, inhalation and dermal absorption, the predominant source of exposure was likely through diet. Since dietary patterns were relatively steady, the BPA level obtained in the present study was likely representative of the average BPA exposure level. Fourth, recall error or bias may exist for reporting retrospective events such as age at menarche. Since the participants were not aware of the specific hypothesis of the study, this recall error was likely to be non-differential misclassification, thereby potentially diluting the associations [41]. Lastly, the study was also limited by the low proportion of girls with detectable BPA(around 60%), although the rate was similar to two other Chinese studies with similar LOD (0.31μg/L) [21, 42].

We observed that BPA exposure was associated with earlier pubertal onset and delayed pubertal progression, although prospective studies are needed to warrant the association.