Genetic epidemiology of depression subgroups and suicide attempts

Major depressive disorder (MDD) is a common psychiatric disorder, affecting 20% of the global population. Individuals with MDD exhibit variable disease manifestations, aetiologies, and treatment responses. This heterogeneity hinders the identification of risk factors, pathophysiology, and treatment specifications.

Studies 1-3 in this thesis investigated the genetic heterogeneity of MDD by studying clinically-informed subgroups. We defined 23 subgroups of MDD based on 12 clinical and psychosocial features such as recurrence, age at onset, or suicidality. Using genotype data from the UK-Biobank cohort (Study 1) and pedigree data from Swedish and Danish registers (Studies 2,3), we estimated the single nucleotide polymorphism (SNP) heritability (SNP-h2) and pedigree heritability (pedigree-h2) of each subgroup, genetic correlations (rg) between the pairs of subgroups defined by specific features, and rg of these subgroups with other psychiatric disorders and traits. Furthermore, we compared psychotic and non-psychotic MDD in their polygenic risk scores (PRS) of mood and psychotic disorders.

We observed varying heritability estimates among MDD subgroups (SNP-h2 range 4.4- 19.0%, pedigree-h2 range 19.7-58.3%). The two MDD subgroups defined within each feature (e.g., single vs. recurrent episodes) were positively correlated with each other, with moderate to high genetic correlations (SNP-/pedigree-rg range from 0.33-0.90). MDD subgroups exhibited diverse patterns of genetic overlaps with other psychiatric disorders, metabolic and cognitive traits. The PRS analyses indicated that, compared to non-psychotic MDD, the psychotic subgroup had a higher genetic burden of schizophrenia, bipolar disorder type 1, a similar genetic burden of bipolar disorder type 2, and a lower genetic burden of MDD. In the UK-Biobank, genome-wide association studies (GWAS) of subgroups identified more associated loci than the GWAS of all MDD. These studies concluded that MDD subgroups with different clinical and psychosocial characteristics may have partially distinct genetic underpinnings. Future research investigating genetic risk, biological mechanisms, and treatments tailored to specific subgroups of MDD is encouraged. Suicide is a significant public health concern, not only common among individuals with MDD but also affects the general population. Individuals who have attempted suicide, with or without psychiatric disorders, are at a high risk of dying by suicide and represent an important target for suicide prevention.

Study 4 provided a comprehensive epidemiological characterisation of suicide attempts. We utilised the Swedish patient register to identify suicide attempts using the patient records related to intentional self-harm, and self-harm of undetermined intent that involves lethal methods or hospitalisation. We estimated cumulative incidence and incidence rates of the first suicide attempts, and assessed the associations with various risk factors such as demographic, comorbidities and adverse life events. We also investigated outcomes including subsequent attempts and mortalities associated with the first suicide attempts. We quantified the familial aggregation of suicide attempts, pedigree-h2 and rg with psychiatric disorders.

Our findings indicated that approximately 3% of the population experienced suicide attempts at some point in their lives, and 1 in 11 Swedish families within the selected birth cohort was affected by suicide attempts. The incidence was notably higher among females compared to males, and highest among the age group 18-24 years. Overdose or poisoning were the most common methods of attempting suicide. Higher socioeconomic statuses were associated with lower risk of suicide attempts, while a previous history of mental illnesses, somatic diseases, and adverse life events, represented strong risk factors. The suicide attempters had a fifteen-fold higher risk of death due to suicide, and a seven-fold higher risk of death due to other causes. We observed a familial aggregation of suicide attempts and an estimated heritability of 50% (95% CI 48.8-51.8%). Suicide attempts exhibited moderate or high genetic correlations with substance use disorders, MDD, bipolar disorder, and schizophrenia (rg ranging from 0.47 to 0.88). Study 4 concluded that suicide attempt carries a substantial disease burden due to its high prevalence, significant proportion of affected families, and increased mortality risks. This study provided evidence for future research investigating aetiology, consequences, and effective prevention strategies tailored to this vulnerable population.

List of scientific papers

I. Nguyen, T. D., Harder, A., Xiong, Y., Kowalec, K., Hägg, S., Cai, N., Kuja-Halkola, R., Dalman, C., Sullivan, P. F., & Lu, Y. (2022). Genetic heterogeneity and subtypes of major depression. Molecular psychiatry. 27(3), 1667–1675.
https://doi.org/10.1038/s41380-021-01413-6

II. Nguyen, T. D., Kowalec, K., Pasman, J., Larsson, H., Lichtenstein, P., Dalman, C., Sullivan, P. F., Kuja Halkola, R., & Lu, Y. (2023). Genetic Contribution to the Heterogeneity of Major Depressive Disorder: Evidence From a Sibling-Based Design Using Swedish National Registers. The American journal of psychiatry. 180(10), 714–722.
https://doi.org/10.1176/appi.ajp.20220906

III. Thuy-Dung Nguyen*, Joeri J. Meijsen*, Ralf Kuja-Halkola, Ying Xiong, Arvid Harder, Kaarina Kowalec, Joëlle Pasman, Sara Hägg, Niamh Mullins, Christina Dalman, Dorte Helenius, Richard Zetterberg, Mikael Landén, Henrik Larsson, Paul Lichtenstein, Thomas Werge, Alfonso Buil, Patrick F. Sullivan, Yi Lu. The Genetic Basis of Psychotic Major Depressive Disorder: Liability within the Mood-Psychotic Spectrum. *Contributed equally. [Manuscript]

IV. Thuy-Dung Nguyen*, Kejia Hu*, Karen Borges*, Ralf Kuja-Halkola, Agnieszka Butwicka, Isabell Brikell, Jame Crowley, Zheng Chang, Brian D’Onofrio, Christian Rück, Henrik Larsson, Paul Lichtenstein, Cynthia Bulik, Patrick F. Sullivan, Fang Fang, Yi Lu. Suicide attempts in Sweden: epidemiology, genetic epidemiology, and healthcare utilization using nationwide data. *Contributed equally. [Manuscript]