Different genetic liabilities to neuropsychiatric conditions in suicides with no prior suicidality.

Importance: Though suicide attempt is the most robust predictor of suicide death, few who attempt go on to die by suicide (<10%), and ∼50% of all suicide deaths occur in the absence of evidence of prior attempts. Risks in this latter group are particularly poorly understood. Objective: Data from the Utah Suicide Mortality Risk Study (USMRS) were used to study underlying polygenic liabilities among suicide deaths without evidence of prior nonfatal suicidal thoughts or behaviors (SD-N) compared to suicide deaths with prior nonfatal suicidality (SD-S). Design: We used an analysis of covariance design, comparing SD-N to SD-S and to population controls with similar genetic ancestry from the United Kingdom. Setting: We selected 12 source studies to generate descriptive quantitative polygenic scores (PGS) reflecting neuropsychiatric conditions. Analysis of covariance was used to evaluate suicide mortality subsets and controls adjusted for sex, age, and genetic ancestry effects. Participants: Suicide deaths were population-ascertained through a 25-year collaboration with the Utah State Office of the Medical Examiner. Evidence of suicidality was determined from diagnoses and clinical notes, yielding 1,364 SD-N and 1,467 SD-S deaths, compared to 20,368 controls. Main Outcomes: The tested PGS spanned 12 psychiatric, neurodevelopmental, and neurodegenerative conditions. Results: SD-N were significantly more male (82.33% vs. 67.76%) and older at death (47.26 years vs. 41.36 years) than SD-S. Controls were significantly less male than both suicide subsets (43.71%). Genetic ancestry was similar across suicide subsets and controls (% European: 96.77%, 96.81%, and 97.38%). Comparing SD-N to SD-S revealed significantly lower PGS in SD-N for: MDD (p=0.0015), neuroticism (p=0.0016), anxiety (p=0.0048), Alzheimer's (p=0.011), depressed affect (p=0.015), schizophrenia (p=0.020), PTSD (p=0.023), and bipolar disorder (p=0.028). This attenuation in SD-N was particularly pronounced for depressed affect, neuroticism, and Alzheimer's, where PGS were not different from controls. Sex-specific analyses suggested attenuation of PGS in SD-N was driven by males for MDD, anxiety, and PTSD, and by females for bipolar disorder, neuroticism, and Alzheimer's. Conclusions and Relevance: SD-N have significantly different genetic liabilities from SD-S, particularly regarding neuropsychiatric conditions. Results have far-reaching implications both for future research and for preventions for those at highest risk of mortality. KEY POINTS: Question: What are underlying genetic liabilities related to neuropsychiatric conditions in the roughly half of suicide deaths with no evidence of prior nonfatal suicidal thoughts or behaviors (SD-N), a group that has not previously been accessible for study? Findings: These suicide deaths with no prior nonfatal suicidality showed significantly attenuated underlying polygenic liabilities associated with mental health traditionally thought to be core features of suicide mortality risk, and justifies additional studies of underlying risks associated with non-psychiatric conditions and behaviors. Meaning: These differences in underlying liabilities between suicide deaths with and without prior suicidality suggest departure from the traditional mental health risks that have been the focus of suicide risk discovery, and impel new directions for future research and prevention efforts.