Metabolomic signature of retinal ageing, polygenetic susceptibility, and major health outcomes

Background/aims To identify the metabolic underpinnings of retinal aging and examine how it is related to mortality and morbidity of common diseases. Methods The retinal age gap has been established as essential aging indicator for mortality and systemic health. We applied neural network to train the retinal age gap among the participants in UK Biobank and used nuclear magnetic resonance (NMR) to profile plasma metabolites. The metabolomic signature of retinal ageing (MSRA) was identified using an elastic network model. Multivariable Cox regressions were used to assess associations between the signature with 12 serious health conditions. The participants in Guangzhou Diabetic Eye Study (GDES) cohort were analyzed for validation. Results This study included 110 722 participants (mean age 56.5±8.1 years at baseline, 53.8% female), and 28 plasma metabolites associated with retinal ageing were identified. The MSRA revealed significant correlations with each 12 serious health conditions beyond traditional risk factors and genetic predispositions. Each SD increase in MSRA was linked to a 24%–76% higher risk of mortality, cardiovascular diseases, dementia and diabetes mellitus. MSRA showed dose–response relationships with risks of these diseases, with seven showing non-linear and five showing linear increases. Validation in the GDES further established the relation between retinal ageing-related metabolites and increased risks of cardiovascular and chronic kidney diseases (all p<0.05). Conclusions The metabolic connections between ocular and systemic health offer a novel tool for identifying individuals at high risk of premature ageing, promoting a more holistic view of human health. All data used in this study are available from UK Biobank via data access procedures (). Permission to use the UK Biobank Resource was obtained via material transfer agreement as part of Application 105658. Not applicable.