Associations between insomnia and suicidal ideation in depressed adolescents: the mediating effect of cognitive symptoms and sex differences

Background Among adolescents diagnosed with major depressive disorder (MDD), suicidal ideation (SI) may be closely linked to cognitive symptoms and insomnia. However, the potential pathways connecting these three factors remain unclear. This study aimed to systematically investigate the mediating role of cognitive symptoms between insomnia and SI, and to further examine whether this mediating pathway differs by sex. Methods This cross-sectional study was conducted from January 2021 to September 2023, recruiting 332 adolescents diagnosed with MDD and 123 healthy controls (HCs). To assess depression severity, cognitive symptoms, insomnia, and suicidal ideation (SI), we administered the Clinical Global Impression – Severity Scale (CGI-S), Hamilton Depression Rating Scale-24 (HAMD-24), Insomnia Severity Index (ISI), and Positive and Negative Suicide Ideation inventory (PANSI), respectively. The PROCESS macro (Model 4) for SPSS was employed to perform mediation analysis, examining whether cognitive symptoms mediated the relationship between insomnia and SI. Results SI was significantly more prevalent among adolescents with MDD than among their healthy counterparts. Within the MDD group, PANSI scores were negatively correlated with sex, age, and age of onset, but positively correlated with CGI-S, ISI, and HAMD-Cog scores (all p < 0.001). These correlations were more pronounced among female patients. Mediation analysis demonstrated that cognitive symptoms, significantly mediated the relationship between insomnia severity (ISI) and suicidal ideation (PANSI). In the overall sample, the total effect, direct effect, and indirect effect values were 0.657 (95% CI: 0.452 - 0.863), 0.463 (95% CI: 0.279 - 0.647), and 0.195 (95% CI: 0.096 - 0.303), respectively. A sex-stratified analysis revealed a significant mediation effect among female patients, with total, direct, and indirect effect values of 0.734 (95% CI: 0.502 - 0.966), 0.539 (95% CI: 0.326 - 0.752), and 0.195 (95% CI: 0.088 - 0.315), respectively. Conversely, in male patients, neither the total nor the direct effects were statistically significant (confidence intervals included zero), and the indirect effect also lacked statistical significance (effect size: 0.132, 95% CI: -0.064 - 0.345), indicating that the mediating role of cognitive symptoms was not supported in this subgroup. Conclusion Adolescents with MDD exhibit elevated levels of SI, which is significantly associated with both cognitive symptoms and insomnia, particularly among females. Furthermore, cognitive symptoms could serve as a mediating factor in the link between insomnia and SI. This indirect effect was statistically significant in female patients, while results for male patients were inconclusive due to limited sample size. Consequently, in clinical practice, it is essential to understand the interconnected pathways linking insomnia, cognitive symptoms, and SI, with close attention to sex differences in the underlying mechanisms. This approach is crucial for enabling more precise risk assessment and targeted intervention.