Abstract Diabetic kidney disease (DKD) and diabetic retinopathy (DR) are important co-existing microvascular complications of type 2 diabetes (T2D) that share common pathogenic mechanisms related to chronic hyperglycemia and microvascular damage. Globally, 20%–40% of all people with T2D have DKD, and about one third have DR. The severity of one diabetic complication often reflects the severity of the other, and both these complications often progress in parallel. Recent studies highlight a significant correlation between the presence and severity of DR and subsequent risk of chronic kidney disease (CKD), suggesting potential utility of ocular biomarkers in CKD detection. The retina provides a noninvasive window to visualize the microvasculature. Recent research has demonstrated that analysis of retinal fundus photographs, routinely collected during DR screening along with some simple noninvasive systemic parameters, and use of artificial intelligence (AI), particularly in deep-learning algorithms (DLA), can accurately detect both prevalent and future risk of CKD in people with T2D. We reviewed articles published up to August 2025 on the correlation of DR and DKD. We also reviewed original research articles on the use of AI for CKD prediction by searching PubMed, Medline, Embase, Scopus, and Google Scholar using keywords like “diabetic retinopathy,” “retinal imaging,” “diabetic kidney disease,” “chronic kidney disease,” “artificial intelligence,” and “deep learning.” This narrative analyses the correlation between DR and DKD and the role of AI systems and noninvasive retinal imaging in the prediction of CKD in T2D.
Rajalakshmi et al. (Fri,) studied this question.