Nonrestorative sleep is associated with self-injury behavior and suicidal ideation in Chinese adolescents: a longitudinal study

Previous studies showed that nonrestorative sleep (NRS) was associated with various adverse outcomes. This study aimed to explore whether NRS was associated with self-injury behavior, suicidal ideation, and suicide attempts. Adolescents attending Grades 7-12 in Nanjing, China, were invited to participate in this longitudinal school-based survey from January 11 to 23, 2024 (T1) and from May 16 to 26, 2024 (T2). The Nonrestorative Sleep Scale (NRSS), Perceived Stress Scale, Center for Epidemiologic Studies Depression-10, and Family APGAR Index were employed to assess NRS, stress, depression, and family function at T1, respectively. The self-injury behavior, suicidal ideation, and suicide attempts during the last four months were measured at T2. Multivariable logistic regressions were applied to investigate the associations of NRS with self-injury behavior, suicidal ideation, and suicide attempts. Stratified analyses were conducted to explore gender-specific effects, and the area under curve was calculated to explore the predictive ability of NRS. Totally, 322 adolescents (57.1% male, mean age 15.4 ± 1.86 years) completed the two-wave surveys. The prevalence of self-injury behavior, suicidal ideation, and suicide attempts in adolescents was 9.3%, 12.7%, and 5.9%, respectively. Reduced NRS symptoms (higher scores of NRSS) were associated with a decreased likelihood of self-injury (odds ratio, OR = 0.89, P < 0.001), suicidal ideation (OR = 0.90, P < 0.001), and suicide attempts (OR = 0.93, P = 0.015). After accounting for sociodemographics, stress, depression, and family function, reduced NRS symptoms significantly decreased the odds of self-injury behavior (adjusted odds ratio, AOR = 0.88, P = 0.002) and suicidal ideation (AOR = 0.92, P = 0.009) but not suicide attempts (AOR = 0.93, P = 0.082). Subgroup analysis showed that the associations of NRS symptoms reached statistical significance exclusively in female participants. The optimal risk-screening threshold of NRSS was 36.5. Reduced NRS symptoms might decrease the risk of self-injury and suicidal ideation. Relieving NRS symptoms with appropriate interventions would be advantageous for adolescents' health.