Immune–metabolic signatures of suicidal ideation in adolescents with major depressive disorder: a sex-specific analysis

Suicide is a major and devastating complication in adolescents with major depressive disorder (MDD). However, it is often underreported in clinical settings, and its biological underpinnings and objective biomarkers remain unclear. This study investigated peripheral immune–metabolic alterations associated with suicidal ideation (SI) and explored sex-specific patterns. In total, 319 adolescents with MDD were enrolled, including 98 with SI and 221 without SI, as defined by the 17-item Hamilton Depression Rating Scale. Peripheral immune and metabolic biomarkers were compared between groups using sex-stratified analyses, and their correlations with SI were examined. Machine learning (ML) models incorporating Shapley Additive Explanations were developed to evaluate classification performance for SI status and identify key sex-specific biomarkers. The SI group exhibited a hypometabolic profile characterized by lower glucose and lipid levels, alongside alterations in peripheral blood inflammatory markers, including elevated white blood cell counts, neutrophil counts and proportions, and derived inflammatory indices. Sex-stratified analyses revealed divergent patterns: inflammatory alterations predominated in female patients, whereas metabolic disturbances were more prominent in male patients. Exploratory mediation analyses suggested bidirectional immune–metabolic associations in female patients, whereas associations in male patients were primarily linked to metabolic pathways. ML models demonstrated good discriminative ability, and SHAP-based feature interpretation supported the observed sex-specific immune–metabolic patterns. SI in adolescents with MDD is associated with distinct immune–metabolic profiles with pronounced sex-specific patterns, characterized by predominantly inflammatory alterations in female patients and metabolic disturbances in male patients. These findings support the potential utility of immune–metabolic biomarkers as correlates of SI and provide a foundation for future longitudinal studies evaluating their role in risk assessment.