早期生命環境暴露與兒童神經發展：整合台灣出生世代研究之證據

Abstract

兒童早期是腦部快速發展的關鍵階段，認知行為與情緒功能的基礎在此時期逐步建立，兒童也因此對環境因子的干擾特別敏感。環境化學物質暴露可能改變兒童的神經發展軌跡,進而影響注意力、學習能力及長期健康。台灣出生世代研究（Taiwan Birth Panel Study, TBPS）及其擴展計畫 TBPS II，提供了長期追蹤、完整暴露評估與生物檢體收集的重要研究平台。然而，現有文獻多聚焦單一類別化學物質、多以產前暴露生物標記為主而忽略兒童期暴露、且神經發展結果多仰賴主觀問卷評估，較少研究能在同一世代中同時比較不同持久性化學物質並搭配客觀測驗工具。

本論文有三項研究目的：（1）探討非持久性對羥基苯甲酸酯類（paraben）暴露與兒童注意力表現之關聯,並著重性別差異;（2）探討持久性全氟/多氟烷基物質（PFAS）暴露與兒童神經行為表現之關聯,同樣聚焦性別特異性;（3）TBPS I 至 TBPS II 的方法學演進，說明暴露評估、結果測量與生物檢體分析如何支持以暴露體（exposome）與生命歷程觀點為核心的次世代研究架構。

本論文的研究方法，納入 TBPS II 世代中 6-8 歲兒童,分別測量尿液中四種 paraben（第一部分，N=446）及血清中 11 種 PFAS（第二部分，N=448），並以客觀化之 Conners' Kiddie 連續表現測驗第二版（K-CPT 2）評估兒童注意力相關表現。統計分析採用線性迴歸及分量 g 計算法（quantile g-computation, QGC）混合物分析，並依性別分層,以探討單一化學物質及混合暴露之效應。第三部分則整理 TBPS I 與 TBPS II 之世代設計沿革。

研究結果顯示，化學暴露與 K-CPT 2 表現之整體關聯並不顯著，但在兩類化學物質中皆一致觀察到性別特異性的型態:女童中,較高濃度 propylparaben 與較差的偵測力（detectability）及較多的誤反應（commission errors）有關；較高濃度 PFPeA 則與較多的遺漏反應（omission errors）有關；PFBS 與 PFHxA 之混合效應則在女童呈正向、在男童呈負向。第三部分結果顯示，TBPS 已由聚焦單一化學物質、假說驅動的 TBPS I 設計，逐步演進為以家庭為單位、exposome 導向的 TBPS II 架構，並擴充內外部暴露評估、表徵分析及基因體/微生物體研究能量。

結論上，兩類化學特性截然不同的環境污染物非持久性 paraben 與持久性 PFAS皆與兒童注意力相關神經發展表現的性別特異性差異有關，顯示性別相關的生物易感性，而非特定化學物質本身的特性，可能是兒童神經發展對環境暴露反應的重要決定因子。持續投入如 TBPS 之出生世代研究基礎建設,對於產出嚴謹、可比較且具性別特異性之科學證據，以支持未來環境兒童健康研究及公共衛生政策制定至為關鍵。

Early childhood is a critical period of rapid brain development during which the foundations of cognitive, behavioral, and emotional functioning are established. Children are therefore particularly susceptible to environmental influences during this developmental stage. Exposure to environmental chemicals may alter neurodevelopmental trajectories, thereby affecting attention, learning abilities, and long-term health. The Taiwan Birth Panel Study (TBPS) and its extension, TBPS II, provide an important research platform characterized by longitudinal follow-up, comprehensive exposure assessment, and biospecimen collection. However, existing studies have generally focused on a single class of chemicals, relied primarily on prenatal exposure biomarkers while overlooking childhood exposure, and assessed neurodevelopment mainly through subjective questionnaires. Few studies have compared chemicals with different degrees of persistence within the same cohort using objective neurobehavioral assessments. This dissertation had three objectives: (1) to investigate the association between exposure to non-persistent parabens and attention-related performance in children, with particular emphasis on sex differences; (2) to examine the association between exposure to persistent per- and polyfluoroalkyl substances (PFAS) and neurobehavioral performance in children, likewise focusing on sex-specific susceptibility; and (3) to review the methodological evolution from TBPS I to TBPS II and illustrate how advances in exposure assessment, outcome measurement, and biospecimen analysis have supported the development of a next-generation research framework centered on the exposome and life-course perspectives. Children aged 6–8 years participating in TBPS II were included. Four urinary parabens were measured in Part I (N = 446), and 11 serum PFAS were measured in Part II (N = 448). Attention-related performance was objectively assessed using the Conners Kiddie Continuous Performance Test, Second Edition (K-CPT 2). Linear regression and quantile g-computation (QGC) were applied to evaluate associations with individual chemicals and chemical mixtures. Sex-stratified analyses were further conducted to examine potential differences between boys and girls. Part III reviewed the evolution of the cohort designs of TBPS I and TBPS II. The results showed that the overall associations between chemical exposures and K-CPT 2 performance were not pronounced; nevertheless, sex-specific patterns were consistently observed across both classes of chemicals. Among girls, higher propylparaben concentrations were associated with poorer detectability and more commission errors, whereas higher perfluoropentanoic acid (PFPeA) concentrations were associated with more omission errors. In the mixture analysis, perfluorobutanesulfonic acid (PFBS) and perfluorohexanoic acid (PFHxA) exhibited positive patterns of association among girls but negative patterns among boys. The methodological review further demonstrated that the TBPS has evolved from the single-chemical, hypothesis-driven design of TBPS I into the family-based and exposome-oriented framework of TBPS II, with expanded capacities for internal and external exposure assessment, refined phenotyping, and genomic and microbiome research. In conclusion, two classes of environmental contaminants with fundamentally different chemical characteristics—non-persistent parabens and persistent PFAS were both associated with sex-specific differences in attention-related neurodevelopmental performance in children. These findings suggest that sex-related biological susceptibility, rather than the characteristics of a specific chemical alone, may be an important determinant of children’s neurodevelopmental responses to environmental exposures. Continued investment in birth-cohort infrastructure such as the TBPS is essential for generating rigorous, comparable, and sex-specific scientific evidence to advance environmental child health research and inform public health policy.