Preventing vision loss in diabetes

Diabetic retinopathy was one of the hot topics at the recent International Diabetes Federation (IDF) meeting. The session covered the latest research into retinal dysfunction and the challenges of setting up national diabetic eye screening in the Philippines. Dr Susan Aldridge reports.

Diabetic retinopathy is part of the systemic disease that is diabetes, so glycaemic control is particularly important. “The Diabetes Control and Complications Trial showed that intensive therapy reduces the onset and progression of diabetic retinopathy,” said Dr Tom Gardner, ophthalmologist at the Juvenile Diabetes Research Foundation (JDRF) Centre of Excellence, University of Michigan. “I point this out because ophthalmologists should be part of the diabetes team. Even so, the challenge persists – in the US, the number of patients with diabetic macular oedema has continued to increase since 2009 and it is the main cause of vision loss in diabetes.”

In addition, glycaemic control is less than perfect with only a minority of people with diabetes meeting targets for HbA1c, blood pressure and cholesterol. If anything, this proportion is decreasing. 

The advent of anti-vascular endothelial growth factor (VEGF) drugs has made a major impact in vision loss and given another option besides photocoagulation. However, peripheral vision still declines after both treatments. “So the anti-VEGF drug is helpful in stabilising the condition but there’s more to the story,” said Dr Gardner.

The current drugs target only VEGF but there have been attempts to target other molecules involved in vascular degeneration, such as faricimab, which is bivalent against VEGF and angiopoietin-2, and has been shown to be non-inferior to the VEGF inhibitor aflibercept. “This is not a major advance but another step forward in trying to understand vascular pathology of diabetic retinopathy,” Dr Gardner said.

It’s also very important to understand which people with diabetes are most at risk of losing their vision. The FIELD study led the way in this kind of prediction, being a retrospective of sight-threatening diabetic retinopathy analysis using an algorithm involving male sex, diabetic retinopathy at baseline, HbA1c and use of fenofibrate (which reduces risk). It shows where the individual is on the progression curve. This is particularly helpful in selecting people for clinical trials. 

Widespread retinal dysfunction

“So far, the available therapies only target the vascular part of the disease – that’s important but not the whole story,” Dr Gardner said. He gave an example of a patient whose eyes could not adapt to dim light. This was not a vascular condition but meant his rods and cones weren’t functioning properly. “We would not pick that up from a photo or examination,” he said, “so it’s important to look at the whole retina and what’s there besides the blood vessels.”

All the retinal cell types are damaged in retinopathy and this begins in the preclinical stage with vascular leakage and occlusion. There is also neuronal apoptosis – particularly in the inner layer of the retina – and astrocyte dysfunction, which is not well understood but is probably important for regulation of synapses.

Furthermore, microglial activation means that there is activation of the immune system within the retina, as well as other abnormalities that may affect electrical function and morphology.

“It is now becoming clear that the whole retina is affected in diabetic retinopathy, not just the vasculature,” Dr Gardner explained. In his lab in Michigan, they’re now looking for cells in the retina that make the macromolecules essential for function. These experiments are helping them to gain a deeper understanding of how the retina works normally to provide vision and how it’s impaired in diabetes.

Put simply, the different functions of the retina should be integrated but because of the dysfunction of so many cell types beyond the vasculature, disintegration occurs, leading to vision impairment. “The retina is a very highly metabolic, electrical tissue whose job is to provide electrical signals,” said Dr Gardner. “It is a sensory organ, so diabetic retinopathy can be seen as a type of sensory neuropathy.”

Retinal screening in the Philippines

Away from retinal research, Dr Paolo Antonio Silva, ophthalmologist and Chief of Diabetic Medicine at Joslin Diabetes Center, described a joint telemedicine project between the Joslin, Queen’s University Belfast and the Philippine Eye Research Institute. This is known as the UK-Philippines Remote Retinal Evaluation Collaboration in Health: Diabetic Retinopathy (REACH-DR). Its aim is to move retinal screening away from tertiary centres to a wider geographical area, using less expensive, more accessible and portable imaging devices.

These mobile hand-held imaging devices are currently in various stages of development – undergoing validation studies, for example – and can be integrated into smartphone or mobile computing. So far, in the early treatment of diabetic retinopathy, the quality of the retinal image using three different hand-held devices has proven similar to that in a standard retinal photo. 

These hand-held devices can be used in community screening programmes where their size, accuracy and ease of use are particularly useful. “Based on these results, we then wanted to carry out a local cost-effectiveness analysis with the Philippines Department of Health and found that the use of a hand-held camera with mydriasis was the best approach,” Dr Silva said.

The REACH team is also starting to apply artificial intelligence (AI) and deep learning (a machine learning technique that teaches computers to do what comes naturally to humans) at the point of care. This will allow for more effective triage of diabetic retinopathy cases and improvement of workflow with less central load.

However, despite the promise of AI, there is a limited number of publications on its application to diabetic retinopathy and the currently approved AI algorithms are not universally generalisable. “There is an under-representation of some populations, which affects the accuracy of AI,” said Dr Silva. This is because 172 countries don’t have the required publicly available ophthalmic datasets and this may perpetuate health inequality in the application of AI. “So there is a need to create clinically useful AI systems that are based on diverse demographics and understand why, how and when different AI systems work in different settings,” he said.

It may be early days, but the REACH Diabetic Retinopathy team is laying the foundations for retinal screening in a country with limited resources, which is good news for people with diabetes everywhere.

Any opinions expressed in this article are the responsibility of the EASD e-Learning Programme Director, Dr Eleanor D Kennedy.