Investigating the genetic and environmental aetiologies of non-suicidal and suicidal self-harm: a twin study

Abstract

Background Self-harm is a major health concern, not only as a signal of distress but also as a strong predictor of later suicide. Self-harm can be further refined into suicidal self-harm (SSH, i.e. suicide attempt) and non-suicidal self-harm (NSSH). Understanding the aetiologies of NSSH and SSH can help inform suicide prevention strategies. Using a twin design, we investigated the phenotypic and aetiological relationships between NSSH and SSH, and their aetiological overlap with mental health problems. Methods We analysed data from the Twins Early Development Study using structural equation modelling. At age 21 years, 9063 twins (62.4% female) answered questions related to self-harm. At age 16 years, 19 self- or parent-reported mental health measures were administered, including measures of internalising and externalising problems, psychotic-like experiences and substance abuse. Results Prevalences for NSSH and SSH were 21.9% and 10.5%, respectively. Additive genetic factors explained half of the variance in NSSH (55%) and SSH (50%), with the rest explained by non-shared environmental factors. Phenotypically, NSSH and SSH were strongly correlated (r = 0.87) with their correlation explained by genetic (57%) and non-shared environmental (43%) factors. We found no evidence that NSSH and SSH differed in their phenotypic and aetiological relationships with mental health measures. Conclusion Our findings suggest no aetiological difference between NSSH and SSH. NSSH and SSH should be regarded as two different ends of a continuum, rather than as two distinct categories.

Prevalence and risk factors for suicide attempts in prison

“…Furthermore, in sensitivity analyses, we found no marked differences in risk factors for near-lethal suicide attempts and suicide attempts more generally. Collectively, these findings may reflect a shared vulnerability to self-harm irrespective of intent or lethality (Lim et al, 2021;Mars et al, 2014a), although future longitudinal studies that directly compare risk factors for multiple outcomes in the same prison population are needed to clarify this. This review highlights several modifiable factors that can be targeted by interventions and may improve suicide prevention in prisons.…”

Investigating the causal risk factors for self-harm by integrating Mendelian randomisation within twin modelling

“…Using the classical twin design, Lim et al ( 2021) found strong phenotypic (r = 0.87) and genetic (r g = 0.94) correlations between NSSH and SSH; and that the genetic and environmental relationships of 17 mental health measures with NSSH were comparable to those with SSH. In that twin study, major depressive disorder (MDD) symptoms and insomnia were identified to have the strongest phenotypic and genetic correlations with both NSSH and SSH among 17 mental health measures (Lim et al 2021). This was also supported by two recent molecular genetic studies, where MDD and insomnia had significant…”

“…By estimating only C x , rC and Cy are constrained to zero by default (see Minică et al, [2018], Scenario 6 in Table 1). Based on earlier research using the same sample that showed that depression and ADHD are ACE traits (Greven et al 2011;Waszczuk et al 2015Waszczuk et al , 2021 whereas NSSH and SSH are AE traits (Lim et al 2021), we initially adopted ACE models for the exposures and AE models for the outcomes, with rC and Cy constrained to zero by default. However, this model had a low resolution (i.e., the confidence intervals were extremely wide e.g., -37 to 477 for a causal path estimate of 0.66).…”

Polygenic contributions to suicidal thoughts and behaviors in a sample ascertained for alcohol use disorders

“…Twin and family studies have been best able to answer these questions using distinct phenotypes; for example, a recent study from Kendler et al [32] found evidence to suggest that suicide attempt and death by suicide are genetically distinct in ways beyond severity. Still, some relationships between these phenotypes are less clear, with some twin studies having found that a large portion of the variation in suicide attempt is explained by genetic factors shared with nonsuicidal self-injury [33] and that the two behaviors share a genetic correlation of 0.94 [34], while a study from Russell et al [35] was not able to identify any genetic correlation between non-. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.…”

“…[32] found evidence to suggest that suicide attempt and death by suicide are genetically distinct in ways beyond severity. Still, some relationships between these phenotypes are less clear, with some twin studies having found that a large portion of the variation in suicide attempt is explained by genetic factors shared with non-suicidal self-injury [33] and that the two behaviors share a genetic correlation of 0.94 [34], while a study from Russell et al . [35] was not able to identify any genetic correlation between non-suicidal self-injury and suicide attempt, although in a considerably smaller sample.…”