Highlights
Polygenic risk scores (PGS) for suicide attempts are significantly associated with suicidal behavior in both live and postmortem cohorts, with odds ratios of 1.35 and 1.34, respectively.
Higher genetic liability for suicide attempts is positively correlated with lifetime aggression and depression severity, but negatively correlated with hostility.
Genetic risk for suicide attempt does not appear to be associated with the lethality of the attempt or impulsivity, suggesting a divergence between the drive to act and the method’s lethality.
There is a significant gene-environment correlation where individuals with higher genetic liability are more likely to report childhood abuse and recent stressful life events, though these stressors do not moderate the genetic association.
The Complexity of Suicidal Behavior and the Search for Biomarkers
Suicidal behavior remains one of the most challenging public health crises globally, with a complex etiology involving a confluence of genetic, psychological, and environmental factors. While heritability estimates for suicide attempts range from 30% to 50%, identifying the specific molecular underpinnings has been difficult due to the polygenic nature of the trait. Clinicians have long sought objective markers to supplement traditional risk assessments, which often rely on subjective reporting of ideation and intent.
Traditional psychiatric research has frequently treated suicidal behavior as a secondary symptom of major depressive disorder or other mood disorders. However, emerging evidence suggests that the genetic architecture of suicide attempts may be partially independent of underlying psychiatric diagnoses. The recent study by Kim et al., published in JAMA Network Open, seeks to bridge this gap by examining the interplay between polygenic scores (PGS), internal trait variables, and external stressors.
Study Design and Methodology
This comprehensive case-control study utilized genome-wide genotyping to investigate individuals across two distinct cohorts: a live cohort and a postmortem cohort. The research was conducted across major academic centers in New York, Montreal, and Munich between 1991 and 2011, with formal data analysis concluding in late 2024.
The sample size of 1,699 individuals provided a robust foundation for analysis. The live cohort (n = 1,275) included 239 individuals with a history of suicide attempts and 1,036 control participants. The postmortem cohort (n = 424) consisted of 294 individuals who died by suicide and 130 control individuals who died of other causes. This dual-cohort approach is particularly valuable, as it allows researchers to compare genetic risk across nonfatal and fatal behaviors.
The primary exposure was the suicide-PGS, calculated based on large-scale genome-wide association studies. The researchers evaluated clinical variables including depression severity, suicidal ideation, and the number/lethality of attempts. Furthermore, internal traits such as hostility, impulsivity, and aggression were measured alongside external stressors like childhood abuse and recent life events. Statistical associations were determined using logistic regression for categorical outcomes and linear or Poisson models for continuous and count data.
Key Findings: Genetic Risk and Clinical Phenotypes
The study confirmed that suicide-PGS is a significant predictor of suicide attempts. In the live cohort, the odds ratio (OR) was 1.35 (95% CI, 1.17-1.56), while the postmortem cohort showed a nearly identical OR of 1.34 (95% CI, 1.07-1.70). These findings reinforce the validity of using PGS as a measure of genetic liability across different populations and outcomes.
Aggression vs. Hostility
One of the most nuanced findings of the study involves the relationship between genetic risk and personality traits. Among those who attempted suicide, a higher suicide-PGS was significantly associated with lifetime aggression severity (b = 0.67; 95% CI, 0.41-0.94). Interestingly, the same genetic liability was associated with lower levels of hostility (b = -0.51; 95% CI, -0.82 to -0.19). This distinction is critical for clinicians: while aggression (the tendency toward physical or verbal outbursts) may be a genetic correlate of suicide risk, hostility (a cynical or suspicious attitude) may not follow the same genetic pathway. This suggests that the genetic drive toward suicidal behavior may be more closely linked to behavioral dysregulation than to cognitive-affective bias.
Depression Severity and Recurrence
Higher suicide-PGS was also linked to increased depression severity (b = 0.20) and a higher number of lifetime depressive episodes (b = 0.11). This suggests that individuals with high genetic risk for suicide are also predisposed to a more severe and recurrent course of affective illness. However, the study found no significant association between suicide-PGS and the number of lifetime suicide attempts or the lethality of the attempts. This implies that while genetics may lower the threshold for making an attempt, the frequency and the potential fatality of those attempts may be governed by other factors, such as access to means or acute environmental triggers.
The Intersection of Nature and Nurture
The study sheds light on the complex relationship between genetic vulnerability and environmental adversity. Suicide-PGS was associated with a higher likelihood of reported childhood abuse (OR, 1.16) and recent life stress (b = 0.17). This phenomenon, known as gene-environment correlation (rGE), suggests that individuals with a higher genetic predisposition for suicidal behavior may either be more likely to experience or more likely to report environmental stressors.
Crucially, the researchers found that while these stressors were associated with suicidal behavior, they did not moderate the relationship between the PGS and suicide. In other words, the genetic risk remained a constant factor regardless of the level of environmental stress. This indicates that genetic liability and environmental trauma may act as additive risk factors rather than multiplicative ones, emphasizing the need for comprehensive risk assessments that account for both history and biology.
Expert Commentary and Clinical Implications
The findings by Kim et al. highlight the multifaceted nature of suicide risk. For the clinician, the association between suicide-PGS and aggression severity suggests that screening for aggressive traits may be a useful proxy for underlying genetic vulnerability. Furthermore, the decoupling of genetic risk from attempt lethality serves as a reminder that all suicide attempts, regardless of medical severity, should be treated with high clinical concern, as the underlying genetic drive may be equally strong in those who survive.
However, the clinical utility of PGS is currently limited by the small effect sizes (OR ~1.3). While these scores provide valuable insights for population-level research and understanding biological pathways, they are not yet ready for individual-level clinical prediction. Future research must focus on larger, more ancestrally diverse datasets to refine these scores. Additionally, the finding that hostility was negatively correlated with suicide-PGS warrants further investigation into the protective or risk-modulating roles of specific personality dimensions.
Conclusion
This case-control study demonstrates that genetic liability for suicide attempt is intricately linked to specific clinical traits like aggression and depression severity, as well as exposure to external stressors. By identifying these associations, the research moves us closer to a translational understanding of suicidal behavior. The findings underscore the importance of a biopsychosocial approach to suicide prevention, where genetic predispositions are considered alongside life history and behavioral phenotypes. As we move toward an era of precision psychiatry, such multidimensional models will be essential for identifying and protecting at-risk individuals.
References
Kim MJ, Galfalvy H, Singh T, Mann JJ. Polygenic Risk, Trait Variables, and External Stressors in Fatal and Nonfatal Suicidal Behavior. JAMA Netw Open. 2026;9(1):e2554325. doi:10.1001/jamanetworkopen.2025.54325.
