Causes and consequences of major depressive disorder: An encompassing Mendelian randomization study

Joelle A. Pasman,Jacob A. Bergstedt,Arvid Harder,Tong Gong,Ying Xiong,Sara Hagg,Fang Fang,Jorien Treur,Karmel W Choi,Patrick F Sullivan,Yi Lu

DOI: https://doi.org/10.1101/2024.05.21.24307678

2024-05-21

Abstract:Background: Major depressive disorder (MDD) is a prevalent and debilitating disorder that has been associated with a range of risk factors and outcomes. Causal pathways between MDD and other traits can be studied using genetic variants as instrumental variables. Methods: A literature review was conducted to identify 201 MDD-associated traits. For 115 traits, there were well-powered genome-wide association study (GWAS) results available that could be used to assess the genetic correlation with MDD. Of these, there were 89 meeting criteria for investigating causal associations in both directions using two-sample Mendelian randomization (TSMR). Of the traits that were not captured by GWAS, 43 could be included as outcomes of MDD using one-sample MR (OSMR). A range of methods and sensitivity tests was applied to gauge robustness of results, together with statistical power analyses to aid interpretation. Outcomes: Moderate to strong genetic overlap was found between MDD and most traits. Support for causal effects of MDD liability were found for circadian, cognitive, diet, medical disease, endocrine, functional, inflammatory, metabolic, mortality, physical activity, reproduction, risk behavior, social, socioeconomic, and suicide outcomes. Most associations were bidirectional, although there was less evidence for diet, disease, and endocrine traits causing MDD risk. Results were robust across sensitivity analyses. Interpretation: This study provides a systematic overview of traits putatively causally related to MDD, confirming previous findings as well as identifying new associations. Our results highlight the importance of MDD as a risk factor cross-cutting across medical, functional, and psychosocial domains and emphasize the need for concerted efforts at reducing this highly prevalent disorder.

What problem does this paper attempt to address?

The paper aims to explore the causal relationships between Major Depressive Disorder (MDD) and a range of related traits. Specifically:

1. **Research Background**: MDD is a common and debilitating mental disorder known to be associated with various risk factors and outcomes. To investigate the causal pathways between MDD and other traits, genetic variation can be used as an instrumental variable.

2. **Methods**: A literature review identified 201 traits associated with MDD, and 115 of these traits with available large-scale genome-wide association study (GWAS) data were analyzed. For 89 of these 115 traits, bidirectional potential causal associations were tested using the two-sample Mendelian randomization (TSMR) method. Additionally, for 43 traits without GWAS data, the one-sample Mendelian randomization (OSMR) method was used for analysis.

3. **Results**: It was found that there is moderate to strong genetic overlap between MDD and most traits. The study supports the causal effects of MDD on various outcomes including circadian rhythm, cognition, diet, disease, endocrine function, inflammation, metabolism, mortality, physical activity, reproduction, risk behavior, social, socioeconomic, and suicide. Most associations were bidirectional, but there was less evidence for diet, disease, and endocrine traits leading to MDD risk. The results remained robust in sensitivity analyses.

4. **Conclusion**: This study provides a systematic overview, confirming previous findings and identifying new associations. MDD, as a significant risk factor, has notable impacts in medical, functional, and psychosocial domains, highlighting the necessity to reduce this prevalent disorder. The TSMR analysis more strongly supports the causal effects of MDD on other traits, with less support for the effects of other traits on MDD. Overall, these results show consistency across different MR methods and, within the limits of current tools, most results demonstrate robustness against violations of MR assumptions.