Modifiable risk factors associated with increased retinal age gap in an Australian population

Retinal age gap (RAG)-the difference between retina-predicted age and chronological age-indicates biological ageing that has been linked to the risk of mortality and chronic disease. We aimed to identify risk factors associated with higher RAG in an Australian population. This cross-sectional study included 5107 participants from the Busselton Healthy Ageing Study (BHAS), a Western Australian community-based cohort. Retinal age was estimated using a validated deep-learning model applied to fundus photographs. Multivariable linear regression models were employed to examine associations between sociodemographic, lifestyle, and clinical factors and the RAG. RCS analysis was performed to investigate potential non-linear relationships and determine threshold effects between each risk factor and the RAG. A total of 4798 BHAS participants had available retinal images for RAG estimation, with a mean age of 58.0 years (SD = 5.8). Fifty-five percent were women, and around 1.5% identified as non-Caucasians. After adjusting for age, sex, and ethnicity, systolic blood pressure (SBP) (β = 0.033, 95% confidence interval [CI]: 0.025-0.042, p < 0.001), diastolic blood pressure (β = 0.052, 95% CI: 0.038-0.066, p < 0.001), body mass index (BMI) (β = 0.065, 95% CI: 0.039-0.092, p < 0.001), alcohol consumption (β = 0.146, 95% CI: 0.078-0.215, p < 0.001), and glycated haemoglobin (HbA1c) (β = 0.461, 95% CI: 0.224-0.698, p < 0.001) were significantly, positively associated with RAG. This indicates the potential clinical use of RAG to facilitate epidemiological investigation, risk stratification, healthy ageing promotion, and reducing age-related disease burden.