We have previously reported that patients with type 2 diabetes (T2DM) of onset < 45 years have more retinopathy even after adjustment for HbA1c and duration of diabetes. Whether there is an age threshold for this increased retinopathy risk is unknown. Furthermore, the possibility exists that these observations arise from the premature death of older patients with retinopathy resulting in a selection bias towards the young. We grouped patients with T2DM according to age of diabetes onset, further stratifying the age groups <45 years of age. The prevalence of retinopathy was examined according to duration of diabetes and updated HbA1c. To evaluate the possibility of a mortality related selection bias, we compared the survival of patients with and without retinopathy at the 15th year of diabetes. Data were retrieved from our database collected over two decades. Retinopathy was detected by fundoscopy or photography. Mortality data was obtained by linkage to National Death Index.
There was a progressive increase in retinopathy with duration of diabetes and this was more pronounced in the younger onset groups (Figure 1). This was also evident within each band of updated HbA1c. Younger age of diabetes onset remained a significant risk factor for retinopathy after adjustment of HbA1c and diabetes duration by logistic regression. Patients with younger onset diabetes did have a higher risk of death if they have retinopathy. However, this was not observed in the older onset group (Figure 2).
Our study showed a progressive increased risk of developing retinopathy with younger age onset of T2DM without a clear age threshold. As retinopathy is not associated with an excess mortality in older subjects, our findings are unlikely to be due to any selection bias. These findings suggest the presence of increased susceptibility to the development of retinopathy at a tissue level in younger people.