Association of Phthalates, Parabens and Phenols Found in Personal Care Products With Pubertal Timing in Girls and Boys

Kim G. Harley; Kimberly P. Berger; Katherine Kogut; Kimberly Parra; Robert H. Lustig; Louise C. Greenspan; Antonia M. Calafat; Xiaoyun Ye; Brenda Eskenazi


Hum Reprod. 2019;34(1):109-117. 

In This Article


We found evidence that prenatal and peripubertal exposure to certain phthalates, parabens and phenols present in personal care and consumer products was associated with pubertal timing in girls, but less so in boys. Specifically, prenatal urinary concentrations of triclosan and 2,4-dichlorophenol were associated with earlier menarche and prenatal concentrations of MEP were associated with earlier pubarche in girls. Peripubertal concentrations of methyl paraben were associated with earlier thelarche, pubarche and menarche in girls, while propyl paraben was associated with earlier pubarche in girls and gonadarche in boys. We also observed an association of peripubertal 2,5-dichlorophenol with later pubarche in girls.

Only one other study has examined the association of prenatal exposure to low molecular weight phthalates with timing of puberty in children. The ELEMENT Study, a longitudinal birth cohort in Mexico City, found prenatal concentrations of MEP to be associated with earlier menarche in girls (Watkins et al., 2014, 2017a,b). While we found prenatal MEP concentrations to be associated with earlier pubarche, these findings are not necessarily consistent, since menarche is regulated by the hypothalamic–pituitary–gonadal axis while pubarche is governed by adrenal mechanisms. Like us, the ELEMENT Study found no associations with prenatal MBP and MiBP concentrations in girls, and no associations of any of the low molecular weight phthalates with pubertal timing in boys (Watkins et al., 2017a,b). To our knowledge, no other studies have examined prenatal exposure to parabens and phenols in relation to pubertal timing.

More studies have examined peripubertal exposure, particularly in girls. The Chilean Growth and Obesity Cohort Study (COGS) found peripubertal MEP concentrations to be associated with earlier menarche in overweight/obese girls (Binder et al., 2018). Peripubertal triclosan concentrations were associated with earlier menarche in the COGS Study (Binder et al., 2018) and with earlier breast development in US girls in the BCERP Study (Wolff et al., 2015). However, other studies of peripubertal exposure have found no association of MEP with any markers of puberty (Frederiksen et al., 2012; Mouritsen et al., 2013; Wolff et al., 2014, 2017) but have found associations between MnBP with later pubarche (Frederiksen et al., 2012; Wolff et al., 2014). The only other studies to examine childhood paraben concentrations in relation to female puberty (Buttke et al., 2012; Wolff et al., 2015, 2017) have not found associations with earlier puberty. Our association between 2,5-dichlorophenol concentrations and later menarche contrasts with two studies that have examined this association; in the BCERP study, urinary concentrations of 2,5-dichlorophenol in girls at ages 6–8 were associated with earlier menarche, pubarche, and thelarche (Wolff et al., 2015, 2017) and in NHANES, 2,5-dichlorophenol concentrations at ages 12–16 were associated with earlier menarche (Buttke et al., 2012). The BCERP Study also observed an association of urinary benzophenone-3 concentrations with later thelarche (Wolff et al., 2015), which was not seen in the present study. Thus, although there are suggestions of some common patterns, overall, the results of our study and the previous studies are not generally consistent.

There are several differences between our study and previous research, most notably in the timing of exposure assessment. Only one other study examined prenatal exposure and the studies examining peripubertal exposure did so at 6–16 years of age. Many studies measured Tanner staging only at one time point, examining biomarker concentrations at the same time. However, our urinary analysis methods and urinary biomarkers were generally consistent with previous studies, and concentrations of phthalate, paraben and phenol biomarkers across studies were broadly within the range of levels in NHANES.

In animal models, exposures during both the prenatal (Bateman and Patisaul, 2008) and peripubertal (Rasier et al., 2006) periods have been associated with altered timing of puberty, sometimes with different effects with different timings of exposure. None of the findings observed with prenatal concentrations in our study were also observed with peripubertal exposure. This complicates our comparison with other studies, most of which have examined only peripubertal exposure. Some of our prenatal findings (e.g. earlier puberty with MEP and triclosan) were somewhat supported by peripubertal findings in other studies, but not by peripubertal findings in our own study. We observed associations with prenatal concentrations of MEP, triclosan and 2,4-dichlorophenol, which may reflect that the in-utero period is the critical window of exposure to disrupt pubertal timing in girls. However, our findings with peripubertal exposure to methyl and propyl parabens could reflect reverse causality if, for example, girls and boys who entered puberty early were more likely to use personal care products containing these preservatives compared to children entering puberty later.

There are some limitations to this study. We lacked exact values for the highest concentrations of several phenols and instead imputed them at the highest standard. However, our findings using this method were similar to results when examining concentrations by quartile, suggesting that it was not a major issue. Additionally, urinary phthalates, parabens and phenol biomarkers typically only reflect exposure in the past 24–48 h, and one or two urinary measurements may not be sufficient to characterize usual exposure over the prenatal and peripubertal periods. Although personal care product-use habits tend to be fairly consistent, as shown by the moderate correlations between the paraben and phenol concentrations during pregnancy, multiple spot urine samples per time window would have been preferable. Finally, there is potential for confounding by other environmental exposures, such as pesticides, given that this population lives in an agricultural community. Future studies should examine multiple concurrent exposures as it is possible that some of these chemicals may act through related mechanisms.

This study is methodologically strong, with a longitudinal design, measurements of exposure biomarkers during two critical windows of development, and clinical Tanner staging conducted every 9 months. However, the study population was limited to Latino children of low socioeconomic status living in a farmworker community and may not be widely generalizable. This study contributes to a growing literature that suggests that exposure to certain endocrine disrupting chemicals may impact timing of puberty in children.