OBJECTIVE Prenatal exposure to heavy metals, pesticides and phenols has been suggested to interfere with neurodevelopment, but the neurotoxicity of their mixtures is still unclear. We aimed to elucidate the associations of maternal urinary concentrations of selected chemical mixtures with intelligence quotient (IQ) in children. METHODS Maternal urinary concentrations of selected heavy metals, pesticide metabolites, and phenols were quantified in pregnant women who participated in the Sheyang Mini Birth Cohort Study (SMBCS) from June 2009 to January 2010. At age 7 years, child's IQ score was assessed using the Chinese version of Wechsler Intelligence Scale for Children (C-WISC) by trained pediatricians. Generalized linear regression models (GLM), Bayesian kernel machine regression (BKMR) models and elastic net regression (ENR) models were used to assess the associations of urinary concentrations individual chemicals and their mixtures with IQ scores of the 7-year-old children. RESULTS Of 326 mother-child pairs, single-chemical models indicated that prenatal urinary concentrations of lead (Pb) and bisphenol A (BPA) were significantly negatively associated with full intelligence quotient (FIQ) among children aged 7 years [β = -2.31, 95% confidence interval (CI): -4.13, -0.48; p = 0.013, sex interaction p-value = 0.076; β = -1.18, 95% CI: -2.21, -0.15; p = 0.025; sex interaction p-value = 0.296, for Pb and BPA, respectively]. Stratified analysis by sex indicated that the associations were only statistically significant in boys. In multi-chemical BKMR and ENR models, statistically significant inverse association was found between prenatal urinary Pb level and boy's FIQ scores at 7 years. Furthermore, BKMR analysis indicated that the overall mixture was associated with decreases in boy's IQ when all the chemicals' concentrations were at their 75th percentiles or higher, compared to at their 50th percentiles. ENR models revealed that maternal urinary Pb levels were statistically significantly associated with lower FIQ scores (β = -2.20, 95% CI: -4.20, -0.20; p = 0.031). CONCLUSIONS Prenatal exposure to selected chemical mixtures may affect intellectual performance at 7 years of age, particularly in boys. Pb and BPA were suspected as primary chemicals associated with child neurodevelopment.

中文翻译：

---

SMBCS研究显示7岁儿童的产前暴露于重金属，农药和酚的混合物以及智商。

目的产前暴露于重金属，农药和苯酚中会影响神经发育，但尚不清楚其混合物的神经毒性。我们旨在阐明儿童中所选化学混合物的母体尿液浓度与智商（IQ）的关系。方法对从2009年6月至2010年1月参加射阳微型出生队列研究（SMBCS）的孕妇中选定的重金属，农药代谢产物和酚的母体尿液浓度进行定量。在7岁时，使用受过训练的儿科医生的中文版韦氏儿童智力量表（C-WISC）。广义线性回归模型（GLM），贝叶斯核仁机器回归（BKMR）模型和弹性净回归（ENR）模型用于评估7岁儿童的尿液中各种化学物质及其混合物的尿液浓度与IQ得分的关系。结果在326对母婴中，单化学模型表明，7岁儿童的产前尿中铅（Pb）和双酚A（BPA）浓度与全智商（FIQ）显着负相关[β= -2.31， 95％置信区间（CI）：-4.13，-0.48；p = 0.013，性别互动p值= 0.076；β= -1.18，95％CI：-2.21，-0.15; p = 0.025；性交互作用的P值和BPA的p值分别为0.296]。按性别进行的分层分析表明，该关联仅在男孩中具有统计学意义。在多化学品BKMR和ENR模型中，产前尿中铅水平与7岁男孩的FIQ得分之间存在统计学上的显着负相关。此外，BKMR分析表明，当所有化学品的浓度都在其第75个百分位数或更高时（而不是在其第50个百分位数时），总体混合物与男孩的智商降低有关。ENR模型显示，孕妇尿中Pb水平在统计学上与FIQ得分较低相关（β= -2.20，95％CI：-4.20，-0.20； p = 0.031）。结论产前暴露于选定的化学混合物可能会影响7岁以下的孩子的智力表现，尤其是在男孩中。铅和双酚A被怀疑是与儿童神经发育有关的主要化学物质。此外，BKMR分析表明，当所有化学品的浓度都在其第75个百分位数或更高时（而不是在其第50个百分位数时），总体混合物与男孩的智商降低有关。ENR模型显示，孕妇尿中Pb水平在统计学上与FIQ得分较低相关（β= -2.20，95％CI：-4.20，-0.20； p = 0.031）。结论产前暴露于选定的化学混合物可能会影响7岁以下的孩子的智力表现，尤其是在男孩中。铅和双酚A被怀疑是与儿童神经发育有关的主要化学物质。此外，BKMR分析表明，当所有化学品的浓度均在其第75个百分位数或更高时（而不是在其第50个百分位数时），总体混合物与男孩的智商降低有关。ENR模型显示，孕妇尿中Pb水平在统计学上与FIQ得分较低相关（β= -2.20，95％CI：-4.20，-0.20； p = 0.031）。结论产前暴露于选定的化学混合物可能会影响7岁以下的孩子的智力表现，尤其是在男孩中。铅和双酚A被怀疑是与儿童神经发育有关的主要化学物质。ENR模型显示，孕妇尿中Pb水平在统计学上与FIQ得分较低相关（β= -2.20，95％CI：-4.20，-0.20； p = 0.031）。结论产前暴露于选定的化学混合物可能会影响7岁以下的孩子的智力表现，尤其是在男孩中。铅和双酚A被怀疑是与儿童神经发育有关的主要化学物质。ENR模型显示，孕妇尿中Pb水平在统计学上与FIQ得分较低相关（β= -2.20，95％CI：-4.20，-0.20； p = 0.031）。结论产前暴露于选定的化学混合物可能会影响7岁以下的孩子的智力表现，尤其是在男孩中。铅和双酚A被怀疑是与儿童神经发育有关的主要化学物质。