Long-term incidence and progression of vision-threatening diabetic retinopathy in Asian populations

To evaluate incidence, progression and risk factors of vision-threatening diabetic retinopathy (VTDR) in two Asian populations with diabetes.

We included data from 1177 Malay and Indian adults aged 40-80 years with diabetes, who participated in baseline (2004-2009) and 12-year follow-up (2017-2023) of the Singapore Epidemiology of Eye Diseases study (SEED). Incident VTDR was defined among participants free of any DR at baseline as the development of severe non-proliferative DR, proliferative DR, or clinically significant macular oedema at follow-up, assessed from two-field fundus photography using ETDRS severity scale without optical coherence tomography (OCT). Progression was assessed in those with mild or moderate DR at baseline (n = 211). Risk factors included age, sex, ethnicity, systolic blood pressure (BP), duration of diabetes, and glycated haemoglobin (HbA1c). Associations between risk factors and outcomes were evaluated using Poisson regression models, and results presented as relative risks (RR) with 95% confidence intervals (CI).

The 12-year cumulative incidence of VTDR was 3.8% (Indians = 4.4%; Malays = 3.0%) while progression to VTDR was 17.1% (Indians = 18.4%; Malays = 14%). In multivariable models, Indian ethnicity (RR = 2.23; 95% CI = 1.17-4.27), higher systolic BP (RR = 1.02; CI = 1.00-1.03), longer duration of diabetes (RR = 1.05; CI = 1.00-1.09), and higher HbA1c (RR = 1.52; CI = 1.35-1.70) were positively associated whereas older age (RR = 0.95; CI = 0.91-0.99) was negatively associated with incident VTDR. Higher HbA1c (RR = 1.29; CI = 1.11-1.49) was associated with VTDR progression.

In this 12-year population-based study, we found a low cumulative incidence of VTDR (4%) but a relatively high progression among those with baseline DR (17%). These findings support the potential value of risk-stratified screening approaches. Individuals without DR at baseline may be considered for less frequent screening, whereas those with mild DR and/or poor glycaemic control may benefit from closer monitoring due to their higher risk of progression.