The Association Between Diabetes and Age-Related Macular Degeneration Among the Elderly in Taiwan.

OBJECTIVE To investigate the relationship between diabetes and future development of age-related macular degeneration (AMD). RESEARCH DESIGN AND METHODS Longitudinal, retrospective cohort study data for the period between 1997 and 2012 were obtained from the Longitudinal Health Insurance Database (LHID) of Taiwan. The final available 71,904 patients with diabetes and 270,213 patients without diabetes >= 50 years of age were further matched by age, sex, and Charlson comorbidity index. In the end, 54,616 study subjects in each of the diabetes and nondiabetes groups were recruited. The stratified populations of patients with diabetes with diabetic retinopathy (DR) (n = 7,119) versus those with diabetes who do not have DR (n = 7,119) and populations of patients with proliferative DR (PDR) (n = 2,134) versus those with nonproliferative DR (NPDR) (n = 2,134) were also obtained. Competing risk regression models were used to assess the adjusted hazard ratio (HR) and 99% CI. The main outcome measures were the first-ever diagnosis of AMD during the observational period. RESULTS The incidences of nonexudative AMD (HR 1.23; P = 0.108) and exudative AMD (HR 1.37; P = 0.023) were not significantly associated with cohorts of persons with diabetes compared with cohorts without diabetes. The stratified analysis showed that nonexudative AMD (HR 3.89; P = 0.001) and exudative AMD (HR 3.42; P < 0.001) were significantly correlated to diabetes with DR cohorts, compared with diabetes without DR cohorts. The incidences of nonexudative AMD (HR 0.53; P = 0.277) and exudative AMD (HR 2.27; P = 0.058) were not significantly different between PDR cohorts compared with NPDR cohorts. CONCLUSIONS This study provides large-scale, population-based evidence that diabetes with retinopathy is independently associated with an increased risk of subsequent AMD development.