Key Takeaways
Extended Takeaways
- In this cohort, non-suicide mortality in borderline personality disorder was more common than suicide mortality, with 14.0% of borderline patients dying of other causes versus 5.9% dying by suicide over 24 years; long-term care should therefore include routine medical risk assessment rather than focusing only on self-harm.
- The leading non-suicide causes of death were cardiovascular (n = 11), substance-related complications (n = 5), cancer (n = 4), and accidents (n = 4), supporting attention to cardiometabolic risk, substance use, and injury prevention in borderline personality disorder.
- For completed suicide, the only significant baseline predictor was number of hospitalizations before index admission (HR = 1.62, P = .037), suggesting that an extensive prior inpatient course may be a more useful warning sign than many commonly cited demographic or diagnostic variables.
- Several baseline factors predicted premature non-suicide death in bivariate analyses—male sex, lower socioeconomic status, government disability, history of drug use disorder, number of prior psychiatric hospitalizations, number of psychiatric medications, and obese-range BMI (≥ 30)—which can help clinicians identify medically vulnerable patients early.
- After adjustment, only male sex (HR = 3.56, P = .003) and more prior psychiatric hospitalizations (HR = 2.93, P < .001) remained significant predictors of premature death not due to suicide, pointing to illness severity and sex-specific risk as the most robust mortality signals in this sample.
- Although borderline patients had higher observed mortality than personality-disordered comparison subjects for both suicide and non-suicide death, between-group differences were nonsignificant for time-to-suicide (HR = 4.42, P = .150) and time-to-death from other causes (HR = 2.93, P = .144), so individual risk stratification within borderline personality disorder remains clinically important.
