Blood Protein Panel Predicts Diabetic Retinopathy Risk 15 Years Before Symptoms

Diabetic retinopathy affects over 100 million people globally and remains a leading cause of blindness, yet current methods to predict who will develop this devastating complication are inadequate. Up to 90% of diabetic retinopathy-related blindness is preventable through early detection, making better prediction tools critically important.

This groundbreaking study analyzed blood samples from 10,873 people with diabetes across two major cohorts—the UK Biobank and the Guangzhou Diabetic Eye Study. Researchers profiled nearly 3,000 proteins and followed participants for up to 15 years to identify which proteins could predict diabetic retinopathy development.

The analysis revealed 668 protein associations with diabetic retinopathy risk and identified three key proteins that emerged as the most powerful predictors: plexin B2 (involved in blood vessel development), growth differentiation factor 15 (a stress response protein), and renin (a blood pressure regulator). This three-protein panel alone achieved comparable performance to complex multi-protein models, with validation across ethnically diverse populations.

Using advanced imaging techniques, researchers confirmed that these proteins correlate with actual retinal blood vessel damage—the hallmark of diabetic retinopathy. Genetic analysis further revealed that renin appears to causally promote the disease, making it a potential therapeutic target.

The proteomic approach dramatically improved prediction accuracy beyond traditional clinical factors like blood sugar levels and diabetes duration. This could transform diabetic care by enabling physicians to identify high-risk patients years before symptoms develop, allowing for targeted interventions to preserve vision.