Abstract
OBJECTIVE: We aimed to examine the prospective associations between loneliness, social isolation, and suicide attempts (SA), and to clarify the roles of genetic and other risk factors.
METHODS: 381,824 participants from the UK Biobank without SA at baseline were analyzed. Cox proportional hazard models were conducted to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for the association. Two-sample Mendelian randomization (MR) analyses corroborated the association. The percentage of excess risk mediated quantified the contribution of adjusted risk factors to the association.
RESULTS: During a median follow-up of 13.6 years, 1281 participants (0.3%) experienced SA. The HR for SA for loneliness compared with no loneliness was 2.29 (95% CI: 1.94-2.71) after minimal adjustment for age, sex, and baseline mental/behavioral disorders. Further adjustment for socioeconomic, lifestyle, social isolation, genetic, psychological, and physical health factors (i.e., fully adjusted) attenuated the association by 77.5%, but it remained statistically significant (HR = 1.29, 95% CI: 1.08-1.54). Social isolation was associated with SA in minimally adjusted models (HR = 1.48, 95% CI: 1.29-1.69), but the association attenuated and became non-significant after full adjustment (HR = 1.12, 95% CI: 0.97-1.29). Participants with both high genetic risk and loneliness had the highest risk of SA (HR = 2.22, 95% CI: 1.74-2.83), although no significant interaction was observed. MR analyses supported a potential causal relationship of loneliness with SA, but not for social isolation.
CONCLUSION: Loneliness, rather than social isolation, is significantly associated with an increased risk of SA, independent of genetic and other established risk factors.