A New Blood Test Could Predict This Diabetes Symptom Before It Starts

For the more than 500 million people living with diabetes worldwide, one of the most feared complications is vision loss. But by the time diabetic eye disease is detected, the damage is often irreversible. A new study published in PLOS Medicine suggests that may be about to change.

What is diabetic retinal neurodegeneration?

Diabetic retinal neurodegeneration, or DRN, is the breakdown of the retina in people with diabetes, which can lead to severe visual impairment and, eventually, vision loss.

But DRN isn't just an eye problem. Scientists believe it functions as a "window" into how diabetes affects the broader nervous system. DRN has been linked to cognitive impairment, dementia, and the degradation of the nerves that run to the fingers and toes. In other words, what's happening in the retina may be signaling what's happening throughout the body's entire neural network.

The problem is that DRN is currently only detected after symptoms appear, at which point the damage is already irreversible. There's currently no standard way to catch it early.

What the study found

To close that gap, researchers have developed an AI-assisted model that can predict diabetic retinal neurodegeneration before any symptoms appear, using a simple blood test.

They analyzed blood plasma from 1,492 patients enrolled in the Guangzhou Diabetic Eye Study, all of whom had type 2 diabetes and no existing DRN at the start. Researchers examined the eyes of 1,218 of those patients through scans over a six-year period. They also validated their findings against a separate cohort of 502 people with diabetes from the UK BioBank.

The team identified 71 plasma proteins associated with DRN, involved in cellular pathways related to inflammation and cellular maintenance.

When disrupted, these biological processes may signal early retinal nerve damage long before it becomes visible.

Using machine learning, the researchers fed the 71 proteins they identified into a predictive model they named Pro-DRN. The model was able to assess a patient's risk of developing DRN before symptoms emerge, outperforming the best existing predictive model by 26%.

What this means for people with diabetes

It's worth noting that the model is based on associations between protein levels and DRN, not direct causal relationships. But the authors are still optimistic about its potential to shift how diabetic eye care is practiced. The researchers have already made Pro-DRN available online so doctors can begin using it to assess patient risk.

If a routine blood test can flag early neurodegeneration before symptoms appear, it opens the door to closer monitoring and, potentially, earlier intervention, whether that's lifestyle changes, neuroprotective treatments, or simply more frequent check-ins with a specialist.

For the hundreds of millions of people managing diabetes, that kind of early warning system could be meaningful. Right now, the standard of care is largely reactive: wait for damage, then treat it. Pro-DRN points toward a model where risk can be identified early, and care is directed toward the people most likely to benefit from it. If you're managing diabetes and want to understand what proactive bloodwork can reveal about your health, it's worth a conversation with your doctor.

The bottom line

For decades, diabetic eye care has been largely reactive: catch the damage, then treat it. Pro-DRN points towards a future where a simple blood test can flag neurological risk before vision is ever threatened. While it won't change care overnight, Pro-DRN is a meaningful step toward a world where that warning actually comes in time.