文献速递儿童青少年精神心境障碍精神分裂症与精神病性障碍

孟德尔随机化揭示童年虐待至神经精神障碍的可改变通路与表观遗传标记

Mendelian randomisation reveals modifiable pathways and epigenetic markers from childhood maltreatment to neuropsychiatric disorders.
PubMed 文献 · Br J Psychiatry · 2025-11-04

研究采用两样本孟德尔随机化分析公开全基因组关联数据,评估遗传预测的童年虐待(N=143 473)对六种神经精神障碍(最大N=500 199)风险的影响。结果显示童年虐待显著增加注意缺陷多动障碍(OR 10.09)、重度抑郁障碍(OR 1.89)和精神分裂症(OR 5.89)风险,部分效应由成瘾相关行为、认知特质和教育程度介导。研究还发现多个CpG位点DNA甲基化改变与上述障碍显著相关。

要点

  • 童年虐待遗传预测与ADHD、MDD和精神分裂症风险升高显著相关。
  • 成瘾行为、认知特质和教育程度介导了4.14-22.17%的效应。
  • ADHD、MDD和精神分裂症分别有3、4、19个CpG位点甲基化改变显著相关。

核心结论

  • 童年虐待显著增加ADHD、MDD和精神分裂症风险,部分效应由可改变的通路介导。
  • 表观遗传标记可能参与童年虐待至神经精神障碍的机制。

研究设计与样本

  • 两样本孟德尔随机化研究,使用公开全基因组关联数据。
  • 童年虐待样本N=143 473,神经精神障碍样本量最大N=500 199。
  • 采用两步孟德尔随机化确定中介比例,多变量孟德尔随机化评估直接效应,表观遗传孟德尔随机化分析DNA甲基化效应。

主要发现

  • 童年虐待与ADHD风险升高相关(OR 10.09,95%CI 4.76-21.40)。
  • 童年虐待与MDD风险升高相关(OR 1.89,95%CI 1.32-2.70)。
  • 童年虐待与精神分裂症风险升高相关(OR 5.89,95%CI 1.46-23.78)。
  • 4.14-22.17%的效应由成瘾相关行为(吸烟起始、休闲屏幕时间、物质滥用)、认知特质(执行功能、智力、风险容忍)和教育程度介导。
  • 童年虐待对ADHD直接效应OR 2.57、对精神分裂症直接效应OR 5.10,均小于总效应;对MDD直接效应OR 1.95基本不变。

关键数值与推荐

  • ADHD总效应OR 10.09(95%CI 4.76-21.40),直接效应OR 2.57。
  • MDD总效应OR 1.89(95%CI 1.32-2.70),直接效应OR 1.95。
  • 精神分裂症总效应OR 5.89(95%CI 1.46-23.78),直接效应OR 5.10。
  • 中介比例范围为4.14-22.17%。
  • ADHD、MDD和精神分裂症分别涉及3、4、19个CpG位点。

临床怎么用

  • 临床应筛查童年虐待史,尤其针对ADHD、MDD和精神分裂症高风险人群。
  • 预防策略需减少童年虐待发生,加强对高风险家庭的支持。
  • 对受害者采取响应性策略,并综合干预成瘾相关行为、认知和社会经济中介因素。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

BACKGROUND: Observational studies suggested an association between childhood maltreatment and neuropsychiatric disorders; however, mediators remain disputed. AIMS: We aimed to confirm the relationship between childhood maltreatmen and neuropsychiatric disorders, and to identify addiction-related, biological, behavioural, cognitive, socioeconomic and epigenetic mediators. METHOD: We used two-sample Mendelian randomisation and publicly available genome-wide association data to evaluate the effect of genetically predicted childhood maltreatment (N = 143 473) on the risk of six neuropsychiatric disorders (up to N = 500 199). We used two-step Mendelian randomisation to determine the proportion of the effect of childhood maltreatment on disorders that was mediated by mediators. We used multivariable Mendelian randomisation to determine the direct effect of childhood maltreatment on disorders accounting for mediators. We used epigenetic Mendelian randomisation to determine the effect of DNA methylation at childhood maltreatment-associated CpG sites on disorders. RESULTS: Childhood maltreatment was significantly associated with higher risk of attention-deficit/hyperactivity disorder (ADHD) (odds ratio 10.09, 95% CI: 4.76-21.40), major depressive disorder (MDD) (odds ratio 1.89, 95% CI: 1.32-2.70) and schizophrenia (odds ratio: 5.89, 95% CI: 1.46-23.78). We determined that 4.14-22.17% of the effect of childhood maltreatment was mediated by addiction-related behaviours (smoking initiation, leisure screen time and substance abuse), cognitive traits (executive functioning, intelligence and risk tolerance) and educational attainment. We found that the direct effects of childhood maltreatment on ADHD (odds ratio 2.57) and schizophrenia (odds ratio 5.10) were less than the total effects, while the direct effect on MDD (odds ratio 1.95) remained relatively unchanged. We found altered DNA methylation levels at 3, 4 and 19 CpG sites to be significantly associated with ADHD, MDD and schizophrenia, respectively. CONCLUSIONS: These results emphasise the need for preventative strategies to reduce childhood maltreatment prevalence, including strengthening support for high-risk families and responsive strategies to mitigate consequences for victims, with clinical screening for childhood maltreatment history and holistic approaches addressing addiction-related, cognitive and socioeconomic mediators.

来源:Br J Psychiatry,查看原文。版权归原作者所有。

引用

  1. Ng NYT, Tang JY, Zhao JV, et al. Mendelian randomisation reveals modifiable pathways and epigenetic markers from childhood maltreatment to neuropsychiatric disorders. Br J Psychiatry. 2025. PMID: 41186115. doi:10.1192/bjp.2025.10433

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。