Supplementary Material for: Alcohol use and mental health: how genetic information can help unravel their relationship

Background: Traditional epidemiological evidence suggests various associations exist between alcohol and mental/cognitive health outcomes. However, whether these reflect causal relationships remains unclear. Mendelian Randomization (MR) – a kind of instrumental variables analysis using genetic variants to proxy for an exposure of interest – has the potential to improve causal inference from observational data.
Summary: In the first part of this review, the challenges of causal inference in the field are discussed, and a theoretical and practical introduction to the technique of MR is given. Next, we report on literature searches performed to update a previous systematic review of MR studies evaluating alcohol–mental health relationships. Twelve relevant studies were identified and considered in the context of the 22 studies included in the previous review. While the reviewed MR literature suggests possible causal relationships/a lack thereof, for the most part, the nature of causal relationships between alcohol and mental health remains unclear.
Key messages: MR is beginning to yield valuable insights into the causal effects of (problematic) alcohol consumption on mental and cognitive health outcomes. Future studies must be mindful of the technique’s underlying assumptions and should allow for potential non-linearity in relationships. Triangulating across sensitivity methods within MR studies, as well as between MR studies and other methods for enhanced causal inference, will be crucial.