Risk factors for suicide in depression in Finland: first-hospitalized patients followed up to 24 years.

OBJECTIVE: To examine longitudinally risk factors for suicide in depression, and gender differences in risk factors and suicide methods. METHOD: We linked data from (i) The Finnish Hospital Discharge Register, (ii) the Census Register of Statistics Finland, and (iii) Statistics Finland's register on causes of deaths. All 56 826 first-hospitalized patients (25 188 men, 31 638 women) in Finland in 1991-2011 with a principal diagnosis of depressive disorder were followed up until death (2587 suicides) or end of the year 2014 (maximum 24 years). RESULTS: Clinical characteristics (severe depression adjusted hazard ratio [AHR] 1.19 [95% CI 1.08-1.30]; psychotic depression AHR 1.45 [1.30-1.62]; and comorbid alcohol dependence AHR 1.26 [1.13-1.41]), male gender (AHR 2.07 [1.91-2.24]), higher socioeconomic status and living alone at first hospitalization were long-term predictors of suicide deaths. Highest risk was associated with previous suicide attempts (cumulative probability 15.4% [13.7-17.3%] in men, 8.5% [7.3-9.7%] in women). Gender differences in risk factors were modest, but in lethal methods prominent. CONCLUSION: Sociodemographic and clinical characteristics at first hospitalization predict suicide in the long term. Inpatients with previous suicide attempts constitute a high-risk group. Despite some gender differences in risk factors, those in lethal methods may better explain gender disparity in risk.

Mediating role of borderline personality disorder traits in the effects of childhood maltreatment on suicidal behaviour among mood disorder patients.

BACKGROUND: Substantial evidence supports an association between childhood maltreatment and suicidal behaviour. However, few studies have examined factors mediating this relationship among patients with unipolar or bipolar mood disorders. METHODS: Depressive disorder and bipolar disorder (ICD-10-DCR) patients (n=287) from the Helsinki University Psychiatric Consortium (HUPC) Study were surveyed on self-reported childhood experiences, current depressive symptoms, borderline personality disorder traits, and lifetime suicidal behaviour. Psychiatric records served to complement the information on suicide attempts. We examined by formal mediation analyses whether (1) the effect of childhood maltreatment on suicidal behaviour is mediated through borderline personality disorder traits and (2) the mediation effect differs between lifetime suicidal ideation and lifetime suicide attempts. RESULTS: The impact of childhood maltreatment in multivariate models on either lifetime suicidal ideation or lifetime suicide attempts showed comparable total effects. In formal mediation analyses, borderline personality disorder traits mediated all of the total effect of childhood maltreatment on lifetime suicide attempts, but only one fifth of the total effect on lifetime suicidal ideation. The mediation effect was stronger for lifetime suicide attempts than for lifetime suicidal ideation (P=0.002) and independent of current depressive symptoms. CONCLUSIONS: The mechanisms of the effect of childhood maltreatment on suicidal ideation versus suicide attempts may diverge among psychiatric patients with mood disorders. Borderline personality disorder traits may contribute to these mechanisms, although the influence appears considerably stronger for suicide attempts than for suicidal ideation.