Genetic and epidemiological analyses of alcohol consumption patterns and age-related macular degeneration risk among current drinkers: a role for TNFRSF10A

Background: Alcohol intake, a major modifiable lifestyle factor, has been variably associated with age-related macular degeneration (AMD). We aim to investigate the association and potential causal relevance of alcohol consumption with AMD risk using complementary epidemiological and genetic approaches. Methods: We used nonlinear Cox proportional hazards models to evaluate incident AMD associations among 440 052 participants of European ancestry with a mean age of 56.7 years and of whom 53.7% were female. We employed two-sample Mendelian randomisation (MR), integrating summary-level genetic associations with cis-expression quantitative trait locus colocalisation between alcohol consumption (UK Biobank n = 433 353) and AMD (GWAS Catalog n = 105 248; FinnGen study n = 474 181), to refine genetic instruments for biologically relevant measures of alcohol consumption. We also used inverse-variance weighting (IVW) and conducted sensitivity analyses. In addition, we integrated publicly available single-cell RNA sequencing (scRNA-seq) data from human retina and retinal pigment epithelium/choroid tissues to assess cell-type-specific expression patterns of genes implicated in AMD. Results: Over the follow-up period, 1553 participants were diagnosed with AMD. Alcohol consumption was inversely associated with AMD risk (hazard ratio (HR) = 0.83; 95% confidence interval (CI) = 0.73 - 0.95), with the association being most pronounced for wine (HR = 0.86; 95% CI = 0.76 - 0.97). Beer consumption showed a positive association with AMD risk. These results were supported by MR analyses (IVW β = -0.13; P = 0.02) and remained consistent in sensitivity analyses. Colocalisation analysis highlighted TNFRSF10A (PP.H4 = 0.98) as a shared locus associated with both alcohol intake and AMD risk. Integration with scRNA-seq data from human samples suggested endothelial cell-enriched expression of TNFRSF10A in AMD tissues, providing functional context for the genetic prioritisation of this locus. Conclusions: Epidemiological and genetic data suggest that TNFRSF10A and its variants may be involved in AMD, with alcohol consumption patterns potentially contributing to this association. This provides new hypotheses for how lifestyle factors may influence disease risk. These findings should not be interpreted as evidence supporting alcohol consumption for AMD prevention, given the potential for residual confounding and the well-established health risks of alcohol.