Scholarlink

<p>Major depressive disorder (MDD) is strongly associated with suicide risk. While lipid metabolism dysregulation is implicated in MDD pathophysiology, its specific relationship with suicide attempts (SA), particularly regarding sex-specific patterns, requires precise quantification. This study examined associations between lipid profiles and SA in young adults with first-episode, drug-naïve (FEDN) MDD, investigating sex differences and developing sex-stratified predictive models. We conducted a cross-sectional analysis of 1,289 young adults with FEDN-MDD, of whom 244 (18.9%) reported a history of SA. Comprehensive demographic, clinical (including Hamilton Depression Rating Scale, Hamilton Anxiety Rating Scale [HAMA], and Positive and Negative Syndrome Scale scores), and metabolic data (lipid profiles, thyroid function, glucose) were collected. Sex-specific multivariate regression and LASSO-optimized machine learning (ML) models (XGBoost, Random Forest) were used to identify risk profiles and evaluate predictive discrimination. Significant sex interactions were identified for multiple clinical and metabolic variables. Sex-specific multivariate regression (following LASSO feature selection) revealed distinct SA risk profiles. For males, suicide risk was independently driven by severe anxiety (HAMA), hypercholesterolemia (TC), and profound thyroid axis dysregulation. For females, risk factors included greater depression/anxiety severity, prolonged illness duration, and specifically reduced HDL-C. To address dataset imbalance, ROSE resampling was applied. In the independent testing sets, XGBoost achieved optimal balanced performance for males (Area Under the Curve [AUC] = 0.760, Specificity = 0.836), while Logistic Regression provided the highest discrimination and sensitivity for females (AUC = 0.809). These findings quantify clinically relevant, sex-divergent associations between metabolic dysregulation and SA in FEDN MDD. The development of sex-specific models demonstrates promise for objective risk stratification. However, their translation into clinical tools necessitates robust external validation in independent cohorts.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Association of lipid profiles with suicide attempts in young adults with first-episode drug-naive major depressive disorder: a gender-specific analysis

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月30日

查看全文
  • Boxuan Li,
  • Minghui Li,
  • Mengqian Li,
  • Xing Wang,
  • Xiao Wang,
  • Jingxuan Li,
  • Zhaoyuan Chen,
  • Ling Sun,
  • Xilian Zhang,
  • Xiangyang Zhang

摘要

Major depressive disorder (MDD) is strongly associated with suicide risk. While lipid metabolism dysregulation is implicated in MDD pathophysiology, its specific relationship with suicide attempts (SA), particularly regarding sex-specific patterns, requires precise quantification. This study examined associations between lipid profiles and SA in young adults with first-episode, drug-naïve (FEDN) MDD, investigating sex differences and developing sex-stratified predictive models. We conducted a cross-sectional analysis of 1,289 young adults with FEDN-MDD, of whom 244 (18.9%) reported a history of SA. Comprehensive demographic, clinical (including Hamilton Depression Rating Scale, Hamilton Anxiety Rating Scale [HAMA], and Positive and Negative Syndrome Scale scores), and metabolic data (lipid profiles, thyroid function, glucose) were collected. Sex-specific multivariate regression and LASSO-optimized machine learning (ML) models (XGBoost, Random Forest) were used to identify risk profiles and evaluate predictive discrimination. Significant sex interactions were identified for multiple clinical and metabolic variables. Sex-specific multivariate regression (following LASSO feature selection) revealed distinct SA risk profiles. For males, suicide risk was independently driven by severe anxiety (HAMA), hypercholesterolemia (TC), and profound thyroid axis dysregulation. For females, risk factors included greater depression/anxiety severity, prolonged illness duration, and specifically reduced HDL-C. To address dataset imbalance, ROSE resampling was applied. In the independent testing sets, XGBoost achieved optimal balanced performance for males (Area Under the Curve [AUC] = 0.760, Specificity = 0.836), while Logistic Regression provided the highest discrimination and sensitivity for females (AUC = 0.809). These findings quantify clinically relevant, sex-divergent associations between metabolic dysregulation and SA in FEDN MDD. The development of sex-specific models demonstrates promise for objective risk stratification. However, their translation into clinical tools necessitates robust external validation in independent cohorts.