Understanding the relationship between loneliness, substance use traits and psychiatric disorders: A genetically informed approach

Loneliness is a common, yet distressing experience associated with adverse outcomes including substance use problems and psychiatric disorders. To what extent these associations reflect genetic correlations and causal relationships is currently unclear. We applied Genomic Structural Equation Modelling (GSEM) to dissect the genetic architecture between loneliness and psychiatric-behavioural traits. Included were summary statistics from 12 genome-wide association analyses, including loneliness and 11 psychiatric phenotypes (range N: 9,537 - 807,553). We first modelled latent genetic factors amongst the psychiatric traits to then investigate potential causal effects between loneliness and the identified latent factors, using multivariate genome-wide association analyses and bidirectional Mendelian randomization. We identified three latent genetic factors, encompassing neurodevelopmental/mood conditions, substance use traits and disorders with psychotic features. GSEM provided evidence of a unique association between loneliness and the neurodevelopmental/mood conditions latent factor. Mendelian randomization results were suggestive of bidirectional causal effects between loneliness and the neurodevelopmental/mood conditions factor. These results imply that a genetic predisposition to loneliness may elevate the risk of neurodevelopmental/mood conditions, and vice versa. However, results may reflect the difficulty of distiguishing between loneliness and neurodevelopmental/mood conditions, which present in similar ways. We suggest, overall, the importance of addressing loneliness in mental health prevention and policy.

Using genomic structural equation modeling to examine the genetic architecture of PTSD and life satisfaction phenotypes

Objective: Posttraumatic stress disorder (PTSD) and life satisfaction phenotypes are inversely related on a phenotypic level. Given these established relations, researchers have begun to examine possible shared genetic contributions to these outcomes, though the existing genetic literature is sparse and examines these relations via univariate methods. We sought to examine the genetic architecture of PTSD and six life satisfaction and well-being phenotypes (i.e. subjective well-being, friend satisfaction, life satisfaction, family satisfaction, work satisfaction, and financial satisfaction) using a multivariate approach.Method: We used Genomic Structural Equation Modeling (gSEM) to analyze summary-level genetic data from large-scale GWAS of the European Ancestry.Results: Findings show that a two, correlated factors model fit the data best, in which PTSD and life satisfaction phenotypes load on separate but correlated factors.Conclusions: Findings suggest that, using multivariate methods, a latent factor capturing many different positive phenotypes is genetically related to PTSD. This finding confirms and extends prior phenotypic work demonstrating that PTSD and positive phenotypes are inversely related.