Methods: We systematically searched electronic databases, including PubMed, EEMBASE, PsycINFO, ProQuest, and CNKI, from their inception to April 30, 2017. Only observational studies that investigated the associations between insomnia symptoms and suicidal behaviors in adolescents were included. The 2016 JBI critical appraisal instruments for appraisal quality was used to assess the study quality. Publication bias was evaluated by the Egger's test and Begg's test. All analyses were conducted using Comprehensive Meta-Analysis 2.0.
Results: A total of nine studies (seven for cross-sectional design and two for prospectively longitudinal design) involving 42,527 participants were included. For seven cross-sectional studies, we found significant associations of any insomnia symptom with suicidal ideation, plan, and attempts (pooled odds ratios [ORs] =1.68, 1.55, and 1.95; 95% confidence intervals [CIs] = 1.50-1.88, 1.40-1.72, and 1.53- 2.50). After categorizing insomnia symptoms into three types: difficulty in initiating sleep [DIS], difficulty in maintaining sleep [DMS], and early morning awakening [EMA]), we found that DIS significantly correlated to suicidal ideation, plan, and attempt, that DMS was significantly associated with suicidal ideation and plan, and that EMA significantly correlated to suicidal plan (all p < 0.05). For two longitudinal studies, there were significant associations of any insomnia symptom with suicidal ideation and attempt using fixed-effects model (pooled ORs =1.21 and 1.25; 95% CI = 1.10-1.33 and 1.08-1.44); however, the significant associations were not observed while using random -effects model. For subgroup analyses and meta-regressions, the studies conducted in the Western countries reported a stronger association between insomnia symptoms and suicidal ideation compared to those conducted in the Eastern countries (p = 0.02). Studies that adjusted for depression had comparable association of any insomnia symptom with suicidal ideation, plan, and attempts in comparison with those did not. (p = 0.20、0.54 and 0.25). The associated between insomnia symptoms and suicidal ideation elevated with increasing age (B = 0.24, p = 0.05). The association between insomnia symptoms and suicidal attempts increased with a higher percentage of women (B=0.04，p=0.003). Both the Egger's test and Begg's test proved no publication bias.
Conclusion: The current evidence suggests substantial associations between insomnia symptoms and suicidal behaviors in adolescents, particularly DIS. Depression did not moderate the relationship. From the standpoint of preventing youth suicide, healthcare providers should screen out adolescents with insomnia symptoms, and scheme effective interventions to reduce insomnia symptoms.