Long-term Cumulative Incidence of Clinically Diagnosed Retinopathy in the Finnish Diabetes Prevention Study

Context Lifestyle intervention reduces the incidence of type 2 diabetes (T2D) in people with impaired glucose tolerance (IGT). Objective This work aimed to find out whether participation in an earlier lifestyle intervention had an effect on the occurrence of clinically diagnosed diabetic retinopathy (DR) during a median of 22 years of follow-up time. Methods The study included 505 individuals from the Finnish Diabetes Prevention Study (DPS) (mean age 55; range, 40-64 years at the onset of the study) with IGT who were originally randomly assigned to the intervention (weight loss, healthy diet, and physical activity) (N = 257) and usual care control groups (N = 248). The median follow-up was 22 years. Clinical retinopathy diagnoses were obtained from the Finnish national hospital Care Register for Health. Data on glycemic parameters, serum lipids, and blood pressure were available from both the intervention (median 4 years) and postintervention period (until year 7). Results No significant difference was found in the cumulative incidence of clinically diagnosed DR between the original intervention (N = 23, 8.9%) and control groups (N = 19, 7.7%) during the extended follow-up (odds ratio: 1.15; 95% CI, 0.61-2.21). A higher cumulative glycated hemoglobin A1c (HbA1c) was significantly associated with a higher risk of retinopathy (hazard ratio 1.4; 1.02-1.88, 95% posterior interval, adjusted for group, age, and sex). Furthermore, the incidence of retinopathy diagnosis was numerically more common among individuals who had developed diabetes during the follow-up (33/349) compared with those who had not (9/156); however, the comparison was not statistically significant (odds ratio: 1.86, 95% CI, 0.89-4.28, adjusted for group, age, and sex). Conclusion A higher cumulative HbA1c was significantly associated with a higher risk of retinopathy. No evidence was found for a beneficial effect of a 4-year lifestyle intervention on the long-term occurrence of clinical DR during a median of 22-year follow-up.