Sleep Disturbances and Suicidality–A Longitudinal Analysis From a Representative Community Study Over 30 Years

Wulf Rössler, Jules Angst, Vladeta Ajdacic-Gross, Helene Haker, Sofian Berrouiguet, Mariam Ujeyl, Nicholas Glozier, Michael P. Hengartner

Study objectives: Associations between sleep problems and suicidality are increasingly acknowledged, but prospective data from well-controlled long-term community studies are lacking. Methods: We analyzed data from a longitudinal cohort study with n = 591 young adults from Zurich, Switzerland, prospectively followed from 1979 (age 20/21 years) to 2008 (age 49/50 years). Twelve-month prevalence of various mental disorders, socio-environmental confounders and sleep problems were carefully assessed with semi-structured interviews at 7 assessment waves spanning overall a 30-year observation period. Interviews were conducted with the "Structured Psychopathological Interview and Rating of the Social Consequences of Psychological Disturbances for Epidemiology" (SPIKE). The 12-month prevalence of sleep problems was graded according to frequency and associated distress of reported symptoms. 12-month prevalence of suicidality was classified as either mild (transient suicidal ideation) or severe (self-harm, suicide attempts). Results: Concurrently, and fully adjusted for several covariates, including mental disorders, relative to no sleep problems there was an odds ratio (OR) of OR = 1.9 (95% confidence interval 1.4-2.5), OR = 3.3 (2.5-4.4), and OR = 1.9 (1.3-2.8) for mild, moderate and severe sleep problems in association with suicidality. There was no evidence for a prospective effect of broad sleep problems on subsequent suicidality. Mild suicidality, but not severe suicidality, prospectively predicted subsequent broad sleep problems in the fully adjusted multivariate model (adjusted OR = 1.5; 1.1-1.9). Disturbed sleep initiation, a proxy for insomnia, significantly predicted subsequent suicidality (OR = 1.5; 1.1-1.9), whereas mild suicidality, but not severe suicidality, significantly predicted subsequent insomnia (OR = 1.5; 1.1-2.0). Conclusions: Sleep problems and suicidality are longitudinally inter-related, which has important implications for clinical practice. Most importantly, the causal pathways appear to be bi-directional and independent of socio-demographics and concomitant mental disorders. More research is needed to examine the possible biopsychosocial etiological mechanisms linking suicidality to sleep problems.