社区老年人抑郁症状复合体的潜在网络模型
该研究基于欧洲MentDis_ICF65+研究中的2623名65-85岁社区老年人横断面数据,使用症状网络多重检测算法识别出8个抑郁症状复合体,并通过潜在网络模型验证其与数据的拟合。模型提示悲伤与焦虑、注意力与自责、自责与死亡念头之间存在相互交互,PCLT支持数据更可能由网络而非潜变量结构生成。
要点
- 社区老年人抑郁症状可归纳为8个相互作用的症状复合体。
- 悲伤/绝望与焦虑/易激惹、注意力/决策与自责、自责与死亡念头间存在交互。
- 症状复合体识别可能简化临床评估并为个体化治疗提供靶点。
核心结论
- 晚年抑郁症状组织为8个相互作用的症状复合体,数据更支持网络结构而非潜变量结构。
研究设计与样本
- 分析欧洲MentDis_ICF65+研究招募的65-85岁社区居住成人横断面数据,n=2623,平均年龄74岁,49%为女性。
- 采用年龄适配的复合国际诊断访谈评估33项抑郁症状严重程度。
- 使用症状网络多重检测算法识别症状复合体,并以潜在网络模型在探索性和验证性分析中评估拟合。
- 敏感性分析包括偏相关似然检验(PCLT)以探究数据生成结构。
主要发现
- Walktrap算法将抑郁症状组织为8个症状复合体:悲伤/绝望、快感缺失/缺乏精力、焦虑/易激惹、自责、睡眠紊乱、激越/食欲增加、注意力/决策、死亡念头。
- 潜在网络模型充分拟合人群中个体症状数据的分布。
- 模型提示悲伤与焦虑、注意力与自责、自责与死亡念头之间存在相互交互。
- PCLT结果确认症状复合体数据更可能由网络而非潜变量结构生成。
临床怎么用
- 识别晚年抑郁的症状复合体可简化临床评估。
- 症状复合体可为个体化治疗提供靶点。
- 有助于寻找晚年抑郁的生物标志物和结局预测因素。
局限性
- 原文未明确报告局限性。
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
Late-life depression has multiple, heterogeneous clinical presentations. The aim of the study was to identify higher-order homogeneous clinical features (symptom complexes), while accounting for their potential causal interactions within the network approach to psychopathology. We analyzed cross-sectional data from community-dwelling adults aged 65-85 years recruited by the European MentDis_ICF65+ study (n = 2623, mean age 74, 49% females). The severity of 33 depressive symptoms was derived from the age-adapted Composite International Diagnostic Interview. Symptom complexes were identified using multiple detection algorithms for symptom networks, and their fit to data was assessed with latent network models (LNMs) in exploratory and confirmatory analyses. Sensitivity analyses included the Partial Correlation Likelihood Test (PCLT) to investigate the data-generating structure. Depressive symptoms were organized by the Walktrap algorithm into eight symptom complexes, namely sadness/hopelessness, anhedonia/lack of energy, anxiety/irritability, self-reproach, disturbed sleep, agitation/increased appetite, concentration/decision making, and thoughts of death. An LNM adequately fit the distribution of individual symptoms' data in the population. The model suggested the presence of reciprocal interactions between the symptom complexes of sadness and anxiety, concentration and self-reproach and between self-reproach and thoughts of death. Results of the PCLT confirmed that symptom complex data were more likely generated by a network, rather than a latent-variable structure. In conclusion, late-life depressive symptoms are organized into eight interacting symptom complexes. Identification of the symptom complexes of late-life depression may streamline clinical assessment, provide targets for personalization of treatment, and aid the search for biomarkers and for predictors of outcomes of late-life depression.
引用
- Belvederi Murri M, Grassi L, Caruso R, et al. Depressive symptom complexes of community-dwelling older adults: a latent network model. Mol Psychiatry. 2021. PMID: 34642459. doi:10.1038/s41380-021-01310-y
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。