Detailing the Genetic and Environmental Factors Shared Between Major Depressive Disorder and Its Comorbidity with Alcohol Use Disorder

Author

• Jonathan L. Wells, Virginia Commonwealth UniversityFollow

Author ORCID Identifier

https://orcid.org/0000-0002-0247-9313

Defense Date

2026

Document Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Epidemiology

First Advisor

James Burch

Second Advisor

Elizabeth Prom-Wormley

Third Advisor

Brion Maher

Fourth Advisor

Jill Rabinowitz

Fifth Advisor

Amanda Gentry

Abstract

Background and Objectives: Major Depressive Disorder (MDD) and Alcohol Use Disorder (AUD) are common, costly, and often comorbid. Patients with comorbid MDD-AUD often experience more severe symptoms and worse outcomes. While overlapping etiological factors have been implicated in these disorders (i.e., common genetic effects, psychosocial factors, and physiological stress indicators), the distinctions of the associations of these factors with MDD and AUD, and their comorbidity, are unclear. Research into these etiological factors is currently limited by the ability of genome-wide polygenic scores (PGS) to predict liability for these outcomes, hindering the development of treatments and interventions that could prevent the onset and improve the long-term outcomes of individuals with these disorders. Thus, the focus of this project is to characterize the etiology of these three disorders in three ways: 1) by investigating the pleiotropy in MDD and AUD (i.e. when one variant is associated with more than one trait); 2) by investigating whether the use of pleiotropy can improve the predictive ability of PGS in all three disorders; 3) by investigating the shared and distinct relationships of the aforementioned implicated factors in all three disorders.

Methods: Data from the National Longitudinal Study of Adolescent to Adult Health (Add Health) and a genome-wide association study (GWAS) of both MDD and AUD were used in this study. The GWAS were used for all genetic analyses, including the generation of PGS, which were created in Add Health. Multiple types of pleiotropically informed PGS were generated: traditionally generated PGS, PGS informed by the variant-level type of pleiotropy, and PGS informed by variant-level associations with genetic covariance. MDD was measured using a self-reported diagnosis, AUD using a self-reported diagnostic checklist, and comorbid MDD-AUD reflecting the co-occurrence of both single disorders. Psychosocial factors were measured using sum scores of adverse childhood experiences (ACEs) and perceived social support. Physiological stress indicators were measured using an allostatic load sum score and self-reported sleep duration. Models were adjusted for known covariates, including two household income variables, for which missing data were imputed. Thus, pooled logistic regressions and structural equation models (SEMs) were used to test hypothesized relationships with each outcome.

Results: When used together MDD and AUD PGS were uniquely associated with comorbid MDD-AUD and in MDD, but not in AUD, more variance was explained by PGS in comorbid MDD-AUD. A novel method for segmenting GWAS summary statistics by pleiotropy type was explored, validated, and used to create PGS. These PGS were uniquely associated with each outcome and were comparable in predictive ability to other pleiotropically informed PGS. PGS were also associated with psychosocial factors but not with physiological stress indicators. Psychosocial factors were associated with physiological stress indicators. Associations with each outcome were observed for psychosocial factors and physiological stress indicators. Lastly, common genetic effects and psychosocial factors showed similar directions of association across all three disorders but had larger magnitudes of association for comorbid MDD-AUD, whereas physiological stress indicators showed differing associations with all three outcomes.

Discussion: This study demonstrated the benefits of using pleiotropically informed PGS to not only improve the predictive ability of PGS but also improve knowledge of similar and distinct etiologies in MDD, AUD, and comorbid MDD-AUD. This study also generated novel information about the nature of the comorbid condition relative to the single disorders, providing support for conceptualizing the comorbidity as a more severe subtype of the single disorder and for distinct etiologies across the three disorders.

Rights

© The Author

Is Part Of

VCU University Archives

Is Part Of

VCU Theses and Dissertations

Date of Submission

5-5-2026