Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Jul 1.
Published in final edited form as: Am J Med Genet B Neuropsychiatr Genet. 2024 Feb 17;195(5):e32974. doi: 10.1002/ajmg.b.32974

Mediational pathways between aggregate genetic liability and non-fatal suicide attempt: A Swedish-population based cohort

Séverine Lannoy 1, Henrik Ohlsson 2, Mallory Stephenson 1, Jan Sundquist 2,3, Kristina Sundquist 2,3, Alexis C Edwards 1
PMCID: PMC11147698  NIHMSID: NIHMS1971896  PMID: 38366742

Abstract

Despite recent progress in the genetics of suicidal behavior, the pathway by which genetic liability increases suicide attempt risk is unclear. We investigated the mediational pathways from family/genetic risk for suicide attempt (FGRSSA) to suicide attempt by considering the roles of psychiatric illnesses. In a Swedish cohort, we evaluated time to suicide attempt as a function of FGRSSA and the mediational effects of alcohol use disorder, drug use disorder, attention-deficit/hyperactivity disorder, major depression, anxiety disorder, bipolar disorder, and non-affective psychosis. Analyses were conducted by sex in three age periods: 15–25 (Nfemales=850,278, Nmales=899,366), 26–35 (Nfemales=800,189, Nmales=861,774), 36–45 (Nfemales=498,285, Nmales=535,831) years. The association between FGRSSA and suicide attempt was mediated via psychiatric disorders. The highest mediation effects were observed for alcohol use disorder in males (15–25 years, HRtotal=1.60 [1.59;1.62], mediation=14.4%), drug use disorder in females (25–36 years, HRtotal=1.46 [1.44;1.49], mediation=11.2%), and major depression (25–36 years) in females (HRtotal=1.46 [1.44;1.49], mediation=7%) and males (HRtotal=1.50 [1.47;1.52], mediation=4.7%). While the direct effect of FGRSSA was higher at ages 15–25, the mediation via psychiatric disorders was more prominent in later adulthood. Our study informs about the psychiatric illnesses via which genetic liability operates to impact suicide attempt risk, with distinct contributions according to age and sex.

Keywords: Suicide attempt, Psychiatric disorders, Mediation, Cox model

Introduction

Suicidal thoughts and behaviors are worldwide public health issues. Suicide exceeds 700,000 deaths annually with approximately 20 non-fatal suicide attempts for each death (WHO, 2021). Non-fatal suicide attempt is defined when individuals “harm themselves with an intent to end their life, but do not die as a result of their actions” (Centers for Disease Control and Prevention, 2022). The etiology of suicide attempt is heterogeneous, implicating genetic liability, environmental stressors, comorbid conditions, and vary by age and sex (Turecki & Brent, 2016). The study of comorbidity indicates that almost 80% of individuals who developed a suicide plan or attempted suicide had a history of mental illness (Nock et al., 2010).

Genetic factors contribute to the etiology of suicide attempt (Brent & Melhem, 2008), with heritability estimates ranging from 0.30–0.55 (Edwards et al., 2021; Voracek & Loibl, 2007). Though familial aggregation of suicidal behavior has been described for many years, it is only recently that we have been able to quantify aggregate genetic liability. The availability of sufficiently powered genome-wide association studies (Docherty et al., 2020; Mullins et al., 2021) now allows us to compute polygenic risk scores, representing a weighted count of risk alleles for a specific disorder. Importantly, an analogous measure of risk (family genetic risk score [FGRS]) has been developed in the Swedish and Danish registries; this score represents the weighted rate of a disorder in the individual’s family and is controlled for familial environment effects (K. S. Kendler et al., 2021b; Krebs et al., 2023). These advances have resulted in several studies that provide support of the role of aggregate genetic liability for suicide attempt and death in a range of suicide-related phenotypes (Edwards, Ohlsson, Moscicki, Sundquist, et al., 2022; Kenneth S. Kendler et al., 2021; Lannoy et al., 2022; Lee, Doyle, Silberstein, et al., 2022).

Prior studies indicate genetic correlations between suicide attempt and comorbid psychiatric disorders such as alcohol use disorder, attention-deficit/hyperactivity disorder (ADHD), major depression, bipolar disorder, and psychotic disorders (Mullins et al., 2021). Using FGRS, findings indicate that genetic risk for psychiatric disorders (alcohol/drug use disorders and major depression) were related to higher risks of suicide attempt. Moreover, the effects of genetic liability for alcohol/drug use disorders on suicide attempt risk were mediated by the occurrence of these disorders (Kenneth S. Kendler et al., 2021). These findings reflect a partial overlap between the etiology of suicide attempt and psychiatric disorders, which may explain the high rate of comorbidity. However, despite recent advances in our ability to capture aggregate genetic liability, it is unclear whether the specific effect of genetic liability for suicide attempt may act through the manifestation of psychiatric illness. Recent research has identified depressive mood as a mediator that explains part of the role of genetic factors in childhood risk for suicide attempt (Lee, Doyle, Silberstein, et al., 2022). This observation expands upon the direct effect of genetic factors by showing that genetic liability for suicide attempt is related to depressive mood, which in turn increases the risk of suicidal behavior. Understanding the pathways through which genetic liability acts is critical, as it further informs the etiological bases of suicidal behavior and may help to implement more effective prevention methods.

This study aims to clarify the mediational pathways to non-fatal suicide attempt by evaluating the direct and indirect effects of aggregate genetic liability using FGRS (Kenneth S. Kendler et al., 2021). Using Swedish national registry data, we targeted a range of psychiatric phenotypes that have been previously associated with suicidal behavior (Mullins et al., 2021). We investigated the association between psychiatric disorders and aggregate genetic liability for suicide attempt (FGRSSA), namely whether these disorders mediated the association between FGRSSA and non-fatal suicide attempt. In particular, leveraging a large Swedish sample, we aimed to fill two important knowledge gaps: (i) How do genetic liability for suicide attempt and the effects of psychiatric disorders on suicide attempt risk vary across the life course? And (ii) do direct and indirect effects of aggregate genetic liability differ according to biological sex? Together, elucidation of these processes will help to better characterize the etiology of suicide attempt by considering genetic liability, comorbid psychiatric disorders, age, and sex.

Materials & Methods

For this cohort study, we collected information on individuals from Swedish population-based registers with national coverage. Registers are linked using each person’s unique personal identification number assigned at birth, which, to preserve confidentiality, was replaced with a serial number by Statistics Sweden. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. All procedures involving human subjects/patients were approved by the Regional Ethical Review Board in Lund (No. 2008/409 and later amendments). As we used registry data, no participant consent was required. Registry data are not publicly available.

Our database consisted of all individuals born in Sweden between 1970–1990 to Swedish born parents, followed until 12/31/2018. In the database, we included FGRSSA and first registration of non-fatal suicide attempt (outcome), alcohol use disorder, drug use disorder, ADHD, major depression, anxiety disorder, bipolar disorder, and non-affective psychosis (Table S1 for ICD-Codes). When individuals were registered with several diagnoses at the same time, we selected the main diagnosis, i.e., reflected the primary reason for presentation. The FGRS is calculated from morbidity risks for suicide attempt in first-through fifth degree relatives, controlling for cohabitation effects, and thus arise from phenotypes in extended pedigrees rather than from molecular genetic data (see online supplement for more details). FGRS have been developed from data derived from the Swedish registries and later in the Danish registries (Krebs et al., 2023). They show robust associations with psychiatric disorders and their clinical features (Kenneth S. Kendler et al., 2021; K. S. Kendler et al., 2021b) and have proven to be sensitive in distinguishing among major affective and psychotic disorders (K. S. Kendler et al., 2021a).

Statistical Analyses

To clarify the pathways from aggregate genetic liability (FGRSSA) to suicide attempt, we used a causal mediation approach in a survival framework (Valeri & VanderWeele, 2015a, 2015b). We targeted first suicide attempt as the outcome and investigated the effect of FGRSSA and the mediational effects of alcohol use disorder, drug use disorder, ADHD, major depression, anxiety disorder, bipolar disorder, and non-affective psychosis on the association between FGRSSA and suicide attempt. All mediators were analyzed in separate models, and analyses were conducted separately for females and males. The models were based on a Cox proportional hazards model with time to suicide attempt as a function of FGRSSA while controlling for year of birth. The effects were estimated on the hazard ratios (HR) scale and presented with 95% confidence intervals (CI). We presented the direct effect of FGRSSA on suicide attempt, the indirect effect of FGRSSA on suicide attempt (through the mediator), the total effect of FGRSSA (=direct + indirect), as well as the proportion of the total effect mediated via psychiatric disorders. In addition, we presented the direct effect of the mediator on suicide attempt, which was measured from a binary logistic model and thereby presented as odds ratios (OR).

We investigated the mediational effects in three age ranges corresponding to meaningful developmental periods (Thapar & Riglin, 2020):15 to 25 (adolescence/emerging adulthood), 26 to 35 (early adulthood), and 36 to 45 (mid-adulthood) years old. This means that the registration of suicide attempt had to occur during the ages investigated. Individuals can be included in all age groups’ analyses. However, individuals that were registered for suicide attempt during ages 15–25 were censored from further analyses. For example, in the 15–25 analyses, individuals were followed from age 15 until end of the follow-up (suicide attempt, death, emigration, 12/31/2018 or age 25). As we focused on first suicide attempt as the outcome, censoring participants with a previous attempt did not impact subsequent models. Individuals who were registered with a psychiatric diagnosis any time prior to end of follow-up were considered as exposed for the remainder of the follow-up; psychiatric disorders had to occur prior to suicide attempt. All statistical analyses were performed using SAS 9.4.

Data availability is not applicable to this article as we used the Swedish registry data, which are not publicly available. Code for the statistical analyses can be available on request.

Results

Descriptive analyses

Table 1 reports the number of individuals included in each model and the prevalence of suicide attempt and psychiatric disorders in females and males across different age groups. We targeted the first suicide attempt and psychiatric phenotypes that occurred prior to suicide attempt.

Table 1.

Descriptive statistics of a Swedish population-based sample born 1970–1990

Females Males
Age groups 15–25 26–35 36–45 15–25 26–35 36–45
N 850,278 800,189 498,285 899,366 861,774 535,831
Suicide attempt 2.4% 1.2% 0.7% 1.7% 1.5% 0.9%
Mean age at suicide attempt 19.7 (2.8) 29.3 (2.8) 39.2 (2.7) 20.6 (2.7) 29.4 (2.8) 39.1 (2.8)
Psychiatric phenotypes
Alcohol use disorder 1.1% 0.7% 0.8% 2.7% 1.7% 1.5%
Drug use disorder 1.3% 0.9% 0.6% 3.7% 2.2% 0.9%
ADHD 0.4% 0.9% 0.9% 0.6% 1.1% 0.9%
Major depression 5.1% 11.0% 9.7% 2.6% 6.1% 5.5%
Anxiety disorder 5.0% 11.5% 10.4% 2.6% 6.8% 5.9%
Bipolar disorder 0.4% 0.8% 0.7% 0.2% 0.5% 0.4%
Non-affective psychosis 0.3% 0.3% 0.2% 0.4% 0.4% 0.3%
Open in a new tab

Figure 1 depicts aggregate genetic liability for suicide attempt, i.e., the mean standardized FGRSSA among individuals with registrations for suicide attempt or psychiatric illness. In both males and females, higher FGRSSA were found in those registered with suicide attempt, followed by alcohol use disorder, drug use disorder, and ADHD. Individuals with major depression and anxiety disorders had lower FGRSSA than those with alcohol and drug use disorders. Overall, males had higher FGRSSA than females, particularly those who reported suicide attempt and alcohol use disorder.

Figure 1.

Figure 1.

Open in a new tab

Mean Familial Genetic Risk Score for suicide attempt

Note. The y-axes depict the mean (SE) Familial Genetic Risk Score (FGRS) for suicide attempt among individuals with a lifetime registration for suicide attempt or psychiatric illness. FGRS is a standardized score with a mean of 0 and an SD of 1. The x-axes illustrate subgroups with suicide attempt (SA) or different psychiatric diagnoses (AUD = alcohol use disorder, DUD = drug use disorder, ADHD = attention-deficit/hyperactivity disorder, MD = major depression, AD = anxiety disorder, BP = bipolar disorder, NAP = non-affective psychosis). The three age groups during which first suicide attempt was registered are: age 15–25, age 26–35, and age 36–45.

Mediation models in females

Results are reported in Figure 2 and show that risk of suicide attempt was related to both genetic liability and psychiatric disorders (alcohol use disorder, drug use disorder, ADHD, major depression, anxiety disorder, bipolar disorder, and non-affective psychosis) in all age groups. The effect sizes for the total effect of FGRSSA ranged from HR=1.45 to 1.54 and the direct effects of psychiatric disorders ranged from 1.52 to 15.8 (Table S2). A few mediational pathways emerged in the different age groups.

Figure 2.

Figure 2.

Open in a new tab

Estimates for the effects of aggregate genetic liability and psychiatric disorders on risk of suicide attempt across ages and sexes

Note. Direct on suicide attempt (SA) represent the direct effects of aggregate genetic risk for suicide attempt (FGRSSA) on suicide attempt risk, Indirect are the effects of FGRSSA that act through the mediator, Total is the total of the direct and indirect effects. The estimates for the direct, indirect, and total effects are represented by hazard ratios (Cox proportional hazards model). Mediators to SA represent the effect of psychiatric disorders on risk for suicide attempt, and the corresponding estimates are odds ratios. Error bars indicate the 95% confidence intervals. From left to right, results are shown for the 15–25, 26–35, 36–45 age groups and depicted in orange for females and in green for males. AUD = alcohol use disorder, DUD = drug use disorder, ADHD = attention deficit hyperactivity disorder, MD = major depression AD = anxiety disorders, BD = bipolar disorder, NAP = non-affective psychosis.

At ages 15–25, the effect of FGRSSA was mediated through alcohol use disorder (total effect=1.52 [1.51; 1.54], proportion mediated=8.9%), drug use disorder (total effect=1.52 [1.50; 1.54], proportion mediated=5.2%), and major depression (total effect=1.54 [1.52; 1.55], proportion mediated=3.8%). We did not find significant indirect effects for ADHD, anxiety disorder, bipolar disorder, and non-affective psychosis (i.e., the 95% CI overlapped with 1).

At ages 26–35, the effect of FGRSSA was mediated via alcohol use disorder (total effect=1.46 [1.44; 1.49], proportion mediated=9.3%), drug use disorder (total effect=1.46 [1.44; 1.49], proportion mediated=11.2%), major depression (total effect=1.46 [1.44, 1.49], proportion mediated=7.0%), anxiety disorder (total effect=1.46 [1.44; 1.48], proportion mediated=3.2%), and bipolar disorder (total effect=1.46 [1.43; 1.48], proportion mediated=2.6%). We found no significant indirect effects for ADHD and non-affective psychosis (Table S2).

At ages 36–45, the effect of FGRSSA was mediated through alcohol use disorder (total effect=1.45 [1.42; 1.49], proportion mediated=6.9%), drug use disorder (total effect=1.46 [1.42; 1.49], proportion mediated=9%), and major depression (total effect=1.46 [1.42; 1.50], proportion mediated=3.2%).

Our results support age-related differences in the direct, indirect, and total effects of genetic liability, which decline over time (all p-values <0.0001). The mediational pathways through psychiatric disorders also shifted across age groups (Figure S1). There was only one exception: We found no significant indirect effects for non-affective psychosis, and this did not differ across ages (p=0.327).

Mediation models in males

Results show the role of aggregate genetic liability and psychiatric disorders on risk for suicide attempt in all age groups (Figure 2). The effect sizes for the total effect of FGRSSA ranged from HR=1.39 to 1.62 and the direct effects of psychiatric disorders ranged from 1.85 to 8.54 (Table S3). We observed distinct mediational pathways as a function of age.

At ages 15–25, the effect of FGRSSA was mediated via alcohol use disorder (total effect=1.60 [1.59; 1.62], proportion mediated= 4.4%), drug use disorder (total effect=1.62 [1.60; 1.64], proportion mediated=7.2%), and major depression (total effect=1.61 [1.59; 1.63], proportion mediated=1.5%). Using confidence intervals as the indicator, we did not find indirect effects for ADHD, anxiety disorder, bipolar disorder, and non-affective psychosis.

At ages 26–35, the effect of FGRSSA was mediated through alcohol use disorder (total effect=1.50 [1.48; 1.52], proportion mediated=8.5%), drug use disorder (total effect=1.50 [1.48; 1.52], proportion mediated=8.6%), major depression (total effect=1.50 [1.47; 1.52], proportion mediated=4.7%), and anxiety disorder (total effect=1.49 [1.47; 1.52], proportion mediated=2.9%) (Table S3).

At ages 36–45, the effect of FGRSSA was mediated via alcohol use disorder (total effect=1.40 [1.36; 1.43], proportion mediated=4.1%), drug use disorder (total effect=1.39 [1.36; 1.43], proportion mediated6.1%), and major depression (total effect =1.39 [1.36; 1.43], proportion mediated=3.5%).

As for females, all the effects significantly differed across age groups (Figure S1, all p-values <0.0001), except for the indirect effect of non-affective psychosis (p=0.056).

Test of equality between females and males

In both females and males, we found significant mediational pathways through psychiatric disorders such as alcohol use disorder, drug use disorder, and major depression across all age groups. However, statistical tests indicated that these effects were different between sexes at younger ages (p<0.0001), with higher indirect effects via alcohol use disorder and drug use disorder in males and higher indirect effects via major depression in females. At ages 26–35, we did not find evidence that the indirect effect through alcohol use disorder differed by sex (p=0.102). However, the indirect effects via drug use disorder and major depression were slightly higher in females (p<0.0001), and we found an indirect effect via bipolar disorder in females but not in males (p<0.0001). At ages 36–45, the indirect effects through alcohol use disorder and drug use disorder were higher in females (p<0.0001). Sex differences are reported in Table 2.

Table 2.

Test of equality between females and males according to age groups

Direct effect of FGRS on SA Indirect effect
p-values p-values
15–25
Alcohol use disorder <0.0001 <0.0001
Drug use disorder <0.0001 <0.0001
ADHD <0.0001 0.004
Major depression <0.0001 <0.0001
Anxiety disorder <0.0001 <0.0001
Bipolar disorder <0.0001 0.007
Non-affective psychosis <0.0001 1
26–35
Alcohol use disorder 0.102 0.575
Drug use disorder 0.001 <0.0001
ADHD 0.103 0.089
Major depression 0.006 <0.0001
Anxiety disorder 0.029 0.545
Bipolar disorder 0.007 <0.0001
Non-affective psychosis 0.103 0.089
26–35
Alcohol use disorder 0.166 <0.0001
Drug use disorder 0.021 <0.0001
ADHD 0.006 0.280
Major depression 0.037 0.941
Anxiety disorder 0.015 0.851
Bipolar disorder 0.050 0.079
Non-affective psychosis 0.019 0.159
Open in a new tab

Discussion

The goal of this study was to gain insight into the mediational pathways from aggregate genetic liability (FGRSSA) to non-fatal suicide attempt by investigating mediations via various psychiatric disorders across ages and sexes. Using a survival model in a well-powered sample, we found that the association between FGRSSA and suicide attempt acted via the manifestation of alcohol use disorder, drug use disorder, and major depression in both females and males. In early adulthood (ages 26–35), we also observed a mediation through anxiety disorder in both sexes and bipolar disorder in females. Our study advances the field of suicide research by informing about the phenotypes via which genetic liability for suicide attempt operates to impact the risk of suicide attempt across the life course.

First, we replicated the previously established roles of alcohol use disorder, drug use disorder, and major depression (Kenneth S. Kendler et al., 2021; Mullins et al., 2021), while we showed both their direct and indirect effects on suicide attempt risk. The high FGRSSA observed in individuals registered with these disorders supports a possible shared genetic liability (Edwards, Ohlsson, Lannoy, et al., 2022; Mullins et al., 2021), further observed in a mediational pathway implying that genetic liability increases suicide attempt risk through the occurrence of alcohol/drug use disorder and major depression across all age ranges in both females and males. This is supported by the high FGRSSA observed in individuals registered with these disorders. It is worth noting, however, that genetic correlations between suicide attempt and these psychiatric disorders are significantly below 1 (Mullins et al., 2021), suggesting that shared genetic liability would only explain part of the mechanisms by which genetic factors increase the risk of suicidal behavior. In addition, it is possible that the higher effect of alcohol and drug use disorders compared to other psychiatric disorders in this survival model would be explained by the variable time to suicide attempt, as alcohol and drug use may facilitate the act of attempting suicide (Pompili et al., 2010) and therefore more closely precede a suicide attempt registration.

In contrast, although FGRSSA was elevated in those registered with ADHD and non-affective psychosis, we did not find evidence of mediation via these phenotypes. Our findings indicate direct effects of ADHD and non-affective psychosis on risk for suicide attempt, which is in line with previous research (Lee, Doyle, Li, et al., 2022; Liu et al., 2022; Mullins et al., 2021; Taylor et al., 2015). However, beyond these effects, FGRSSA was not mediated through the occurrence of ADHD or non-affective psychosis. We also observed specific pathways for anxiety disorder (both sexes) and bipolar disorder (females), as results indicate that these phenotypes only mediated the FGRSSA-suicide attempt association in early adulthood (age 26–35). These distinct observations underscore etiologic heterogeneity in the association between psychiatric illnesses and suicide attempt.

Second, these findings provide insight into sex differences and developmental perspectives. We observed sex differences in the direct and indirect effects of psychiatric disorders on suicide attempt risk, indicating the prominence of affective disorders in females and of substance use disorders in both sexes. While the direct effect of substance use disorders on suicide attempt risk was higher in females, the proportion mediated by these phenotypes was overall higher in males. This is consistent with previous findings (Merikangas & Almasy, 2020), while our results clearly show that the mediational pathways from genetic liability are also influenced by these sex effects.

As for age, early adulthood (ages 26–35) is the period where the mediational pathways explained the highest proportion of the FGRSSA-suicide attempt association. According to a recent meta-analysis, this age range corresponds to the median age of onset for depression, anxiety, and substance use disorders (Solmi et al., 2022), suggesting that both direct and indirect effects of psychiatric phenotypes on risk for suicide attempt were related to the onset of the disorder (Edwards, Ohlsson, Moscicki, Crump, et al., 2022). In males, we also observed that a higher proportion of the association between genetic liability and suicide attempt risk was mediated through alcohol use disorder in adolescence/emerging adulthood (ages 15–25, 14.4%), and this effect decreased considerably in later adulthood (ages 36–45, 4.1%). This could be explained by the early age of alcohol use/problems onset in males (Erol & Karpyak, 2015), which is known to be involved in risk of suicidal behavior (Lannoy et al., 2022). In contrast to males, the mediational pathways via alcohol/drug use disorders were more persistently high in later adulthood in females.

Finally, our results indicate that the direct effect of aggregate genetic liability was relevant to all age groups and the strongest during adolescence/emerging adulthood, i.e., when the heritability of suicidal behavior is documented to be the largest (Edwards et al., 2021). Overall, these findings clearly denote the importance of adopting a developmental perspective, which is slightly different in males and females, and informs about specific vulnerability windows according to the phenotype of interest.

These results should be interpreted in the context of some limitations. First, we evaluated suicide attempts in the Swedish medical registry, which captures attempts that came to the attention of health care providers. Some proportion of suicide attempts do not require medical attention, and this proportion might vary across age groups (Jollant et al., 2022). Moreover, ICD codes have been shown to underestimate the incidence of suicidal behavior in previous works (Randall et al., 2017). These results should, therefore, be confirmed by the inclusion of self-reported suicide attempts where available, although such samples would be much smaller. As this study only focused on non-fatal suicide attempt, future studies should also investigate mediational pathways from aggregate genetic liability with suicide death as an outcome. Second, prevalence and age at first ADHD registration have changed in the last decades (Xu et al., 2018), suggesting that some, previously unidentified, individuals may have received a diagnosis several years after the disorder onset. Future studies should go further in evaluating the effects of ADHD on suicide attempt risk in different age groups. Third, the direct and indirect effects of FGRSSA are considerably lower than the effects of psychiatric disorders, underscoring that while considering genetic factors is essential in etiological models, it is not sufficient to inform prevention methods. Though statistically significant in this large cohort, the HRs for indirect effects were quite small (1.01–1.06). Fourth, we performed mediation analyses in a survival framework and explored the role of various psychiatric disorders. However, to facilitate model convergence and improve interpretability, only one psychiatric disorder was included in each model. Future studies should consider these results to study the possible comorbidity between psychiatric mediators (e.g., between major depression and anxiety disorder or alcohol and drug use disorders). Finally, though we used survival models and relied on longitudinal data, it is worth noting that mediation analyses in the context of observational data do not prove causal mechanisms.

Conclusion

This study provides novel insights into how genetic liability may manifest in risk for non-fatal suicide attempt. The current findings underscore the dynamic role of psychiatric illnesses in their associations with genetic liability and suggest that targeting psychiatric disorders at specific ages could help mitigate the impact of aggregate genetic liability for suicide attempt and prove useful for risk assessment. Alcohol use disorder appears important in all age groups, with higher effects in adolescence (males) and early adulthood (females). Drug use disorder and depression are also involved in the mediational pathways from genetic liability across all age groups, while effects of anxiety and bipolar disorders are specific to early adulthood and females for the latter. Both age and sex allow for a better understanding of the pathways to suicide attempt and need to be considered in prevention and risk assessment, as well as in the clinical decisions regarding treatment of psychiatric disorders.

Supplementary Material

Fig S1
Supinfo

Acknowledgments:

This study has been supported by the National Institute of Health (grant number MH129356) and the Swedish Research Council (grant number 2020-01175).

Footnotes

Conflict of Interests: None.

Data availability statement

Registry data are not publicly available.

References

  1. Brent DA, & Melhem N (2008). Familial transmission of suicidal behavior. Psychiatric Clinics of North America, 31(2), 157–177. 10.1016/j.psc.2008.02.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Centers for Disease Control and Prevention. (2022). Preventing suicide.
  3. Docherty AR, Shabalin AA, DiBlasi E, Monson E, Mullins N, Adkins DE, Bacanu SA, Bakian AV, Crowell S, Chen D, Darlington TM, Callor WB, Christensen ED, Gray D, Keeshin B, Klein M, Anderson JS, Jerominski L, Hayward C, … Coon H (2020). Genome-Wide Association Study of Suicide Death and Polygenic Prediction of Clinical Antecedents. American Journal of Psychiatry, 177(10), 917–927. 10.1176/appi.ajp.2020.19101025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Edwards AC, Ohlsson H, Lannoy S, Stephenson M, Crump C, Sundquist J, Sundquist K, & Kendler KS (2022). Shared genetic and environmental etiology between substance use disorders and suicidal behavior. Psychological Medicine, in press. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Edwards AC, Ohlsson H, Moscicki E, Crump C, Sundquist J, Kendler KS, & Sundquist K (2022). Alcohol use disorder and non-fatal suicide attempt: findings from a Swedish National Cohort Study. Addiction, 117(1), 96–105. 10.1111/add.15621 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Edwards AC, Ohlsson H, Moscicki E, Crump C, Sundquist J, Lichtenstein P, Kendler KS, & Sundquist K (2021). On the Genetic and Environmental Relationship Between Suicide Attempt and Death by Suicide. American Journal of Psychiatry, 178(11), 1060–1069. 10.1176/appi.ajp.2020.20121705 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Edwards AC, Ohlsson H, Moscicki EK, Sundquist J, Crump C, Kendler KS, & Sundquist K (2022). Genetic differences between suicide deaths and deaths of undetermined intent. Suicide Life Threat Behav. 10.1111/sltb.12926 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Erol A, & Karpyak VM (2015). Sex and gender-related differences in alcohol use and its consequences: Contemporary knowledge and future research considerations. Drug Alcohol Depend, 156, 1–13. 10.1016/j.drugalcdep.2015.08.023 [DOI] [PubMed] [Google Scholar]
  9. Jollant F, Hawton K, Vaiva G, Chan-Chee C, du Roscoat E, & Leon C (2022). Non-presentation at hospital following a suicide attempt: a national survey. Psychol Med, 52(4), 707–714. 10.1017/S0033291720002305 [DOI] [PubMed] [Google Scholar]
  10. Kendler KS, Ohlsson H, Mościcki EK, Sundquist J, Edwards AC, & Sundquist K (2021). Genetic liability to suicide attempt, suicide death, and psychiatric and substance use disorders on the risk for suicide attempt and suicide death: a Swedish national study. Psychological Medicine, 1–10. 10.1017/s0033291721003354 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Kendler KS, Ohlsson H, Sundquist J, & Sundquist K (2021a). Family Genetic Risk Scores and the Genetic Architecture of Major Affective and Psychotic Disorders in a Swedish National Sample. JAMA Psychiatry, 78(7), 735–743. 10.1001/jamapsychiatry.2021.0336 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Kendler KS, Ohlsson H, Sundquist J, & Sundquist K (2021b). The patterns of family genetic risk scores for eleven major psychiatric and substance use disorders in a Swedish national sample. Transl Psychiatry, 11(1), 326. 10.1038/s41398-021-01454-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Krebs MD, Hellberg K-LG, Lundberg M, Appadurai V, Ohlsson H, Pedersen E, Steinbach J, Matthews J, LaBianca S, Calle X, Meijsen JJ, Ingason A, Buil A, Vilhjálmsson BJ, Flint J, Bacanu S-A, Cai N, Dahl A, Zaitlen N, … Schork AJ (2023). PA-FGRS is a novel estimator of pedigree-based genetic liability that complements genotype-based inferences into the genetic architecture of major depressive disorder. preprint. 10.1101/2023.06.23.23291611 [DOI]
  14. Lannoy S, Mars B, Heron J, & Edwards AC (2022). Suicidal ideation during adolescence: The roles of aggregate genetic liability for suicide attempts and negative life events in the past year. J Child Psychol Psychiatry, 63(10), 1164–1173. 10.1111/jcpp.13653 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Lee PH, Doyle AE, Li X, Silberstein M, Jung JY, Gollub RL, Nierenberg AA, Liu RT, Kessler RC, Perlis RH, & Fava M (2022). Genetic Association of Attention-Deficit/Hyperactivity Disorder and Major Depression With Suicidal Ideation and Attempts in Children: The Adolescent Brain Cognitive Development Study. Biol Psychiatry, 92(3), 236–245. 10.1016/j.biopsych.2021.11.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Lee PH, Doyle AE, Silberstein M, Jung JY, Liu RT, Perlis RH, Roffman J, Smoller JW, Fava M, & Kessler RC (2022). Associations Between Genetic Risk for Adult Suicide Attempt and Suicidal Behaviors in Young Children in the US. JAMA Psychiatry, 79(10), 971–980. 10.1001/jamapsychiatry.2022.2379 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Liu RT, Walsh RFL, Sheehan AE, Cheek SM, & Sanzari CM (2022). Prevalence and Correlates of Suicide and Nonsuicidal Self-injury in Children: A Systematic Review and Meta-analysis. JAMA Psychiatry, 79(7), 718–726. 10.1001/jamapsychiatry.2022.1256 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Merikangas AK, & Almasy L (2020). Using the tools of genetic epidemiology to understand sex differences in neuropsychiatric disorders. Genes Brain Behav, 19(6), e12660. 10.1111/gbb.12660 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Mullins N, Kang J, Campos AI, Coleman JRI, Edwards AC, Galfalvy H, Levey DF, Lori A, Shabalin A, Starnawska A, Su M-H, Watson HJ, Adams M, Awasthi S, Gandal M, Hafferty JD, Hishimoto A, Kim M, Okazaki S, … Ruderfer DM (2021). Dissecting the shared genetic architecture of suicide attempt, psychiatric disorders and known risk factors. Biological Psychiatry, in press. 10.1101/2020.12.01.20241281 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Nock MK, Hwang I, Sampson NA, & Kessler RC (2010). Mental disorders, comorbidity and suicidal behavior: results from the National Comorbidity Survey Replication. Mol Psychiatry, 15(8), 868–876. 10.1038/mp.2009.29 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Pompili M, Serafini G, Innamorati M, Dominici G, Ferracuti S, Kotzalidis GD, Serra G, Girardi P, Janiri L, Tatarelli R, Sher L, & Lester D (2010). Suicidal behavior and alcohol abuse. Int J Environ Res Public Health, 7(4), 1392–1431. 10.3390/ijerph7041392 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Randall JR, Roos LL, Lix LM, Katz LY, & Bolton JM (2017). Emergency department and inpatient coding for self-harm and suicide attempts: Validation using clinician assessment data. Int J Methods Psychiatr Res, 26(3). 10.1002/mpr.1559 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Solmi M, Radua J, Olivola M, Croce E, Soardo L, Salazar de Pablo G, Il Shin J, Kirkbride JB, Jones P, Kim JH, Kim JY, Carvalho AF, Seeman MV, Correll CU, & Fusar-Poli P (2022). Age at onset of mental disorders worldwide: large-scale meta-analysis of 192 epidemiological studies. Mol Psychiatry, 27(1), 281–295. 10.1038/s41380-021-01161-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Taylor PJ, Hutton P, & Wood L (2015). Are people at risk of psychosis also at risk of suicide and self-harm? A systematic review and meta-analysis. Psychol Med, 45(5), 911–926. 10.1017/S0033291714002074 [DOI] [PubMed] [Google Scholar]
  25. Thapar A, & Riglin L (2020). The importance of a developmental perspective in Psychiatry: what do recent genetic-epidemiological findings show? Mol Psychiatry, 25(8), 1631–1639. 10.1038/s41380-020-0648-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Turecki G, & Brent DA (2016). Suicide and suicidal behaviour. The Lancet, 387(10024), 1227–1239. 10.1016/s0140-6736(15)00234-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Valeri L, & VanderWeele TJ (2015a). eAppendix for: ”SAS macro for causal mediation analysis with survival data”. Epidemiology. [DOI] [PubMed] [Google Scholar]
  28. Valeri L, & VanderWeele TJ (2015b). SAS macro for causal mediation analysis with survival data. Epidemiology, 26(2), e23–24. 10.1097/EDE.0000000000000253 [DOI] [PubMed] [Google Scholar]
  29. Voracek M, & Loibl LM (2007). Genetics of suicide: a systematic review of twin studies. The Middle European Journal of Medicine, 119(15–16), 463–475. 10.1007/s00508-007-0823-2 [DOI] [PubMed] [Google Scholar]
  30. WHO. (2021). Suicide worldwide in 2019: global health estimates.
  31. Xu G, Strathearn L, Liu B, Yang B, & Bao W (2018). Twenty-Year Trends in Diagnosed Attention-Deficit/Hyperactivity Disorder Among US Children and Adolescents, 1997–2016. JAMA Netw Open, 1(4), e181471. 10.1001/jamanetworkopen.2018.1471 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Fig S1
NIHMS1971896-supplement-Fig_S1.pdf (6.6KB, pdf)
Supinfo
NIHMS1971896-supplement-Supinfo.docx (28.4KB, docx)

Data Availability Statement

Registry data are not publicly available.

RESOURCES