Yan Li, Yanan Xing, Xiaoqin Yin, Shaolian Zang, Ting Dong, Jiaoru Yang, Jiayin Dai, Yitao Pan, Nan Sheng, Pin Li
PFAS exposure and TH levels differ among CPP, IPP, and healthy girls, with concentration-dependent associations. PFOA and PFBS are key thyroid disruptors, and IPP girls appear particularly sensitive. Mechanistic studies are needed.
BACKGROUND: Per- and polyfluoroalkyl substances (PFASs) are persistent environmental contaminants widely detected in Chinese children. Exposure has been associated with thyroid function and altered puberty timing.
OBJECTIVES: To evaluate associations between serum PFAS concentrations and thyroid hormone (TH) levels in girls with central precocious puberty (CPP), incomplete precocious puberty (IPP), and healthy controls, and to identify key PFAS contributors using machine learning.
METHODS: In this prospective study, 943 serum samples were collected from 226 CPP, 316 IPP, and 401 healthy girls. Levels of 25 PFASs, total triiodothyronine (T3), thyroxine (T4), and reverse T3 (rT3) were measured. Multiple linear regression and random forest models assessed PFAS-TH associations. Feature importance (%IncMSE) was determined by random forest. Logistic regression evaluated odds of CPP and IPP associated with TH levels and interquartile shifts in key PFASs.
RESULTS: Among 25 PFASs, 12 had detection rates >95% and were analyzed. Perfluorooctanoic acid (PFOA) showed the highest levels, followed by perfluorooctane sulfonate. After adjustment, higher T3 was associated with increased odds of CPP (OR = 2.86) and IPP (OR = 6.38). In the CPP group, PFHpA was negatively associated with T3, while PFOA, Perfluorobutanesulfonic acid (PFBS), and Perfluorohexanesulfonic acid (PFHxS) were inversely associated with T4. In the IPP group, PFBS showed positive association with T3 but negative with T4; PFOS was also inversely associated with T4. The highest PFOA quartile linked to a 4.50% T4 decrease in CPP girls. PFBS at 0.20-0.58 ng/mL correlated with a 6.29% T4 reduction in the IPP group. Random forest explained 28.5%-35.4% of T4 variance across groups, with highest predictive power in IPP. PFOA and PFBS emerged as most influential for thyroid disruption (mean %IncMSE 21.9% and 22.2%).
CONCLUSION: PFAS exposure and TH levels differ among CPP, IPP, and healthy girls, with concentration-dependent associations. PFOA and PFBS are key thyroid disruptors, and IPP girls appear particularly sensitive. Mechanistic studies are needed.