The Diabetic Retinopathy Cascade: How Preclinical Disease Initiates Retinal Damage

Sixty years ago, William Beetham, MD, of Massachusetts Eye and Ear, said medicine could offer no options to prevent or slow diabetic retinopathy (DR). At that time, he was basically correct. Ophthalmologists had to accept progressive vascular degeneration.¹ 

But in 2023, we know that diabetic eye disease starts with diabetic neurodegeneration (DN), which affects all the retinal neurons and glia. The diabetic damage can affect other ocular tissues, such as cranial nerves (third and sixth), and the corneal epithelium, and lead to cataracts and glaucoma.² 

With this knowledge, physicians can rightly assume that, in the near future, emerging technologies can detect and quantify this damage before diabetic retinopathy (DR) is observable. With the ability to detect disease this early, it can possibly be arrested, or even reversed.³ 

However, although various technological advances in diagnostics have emerged in recent years, many clinics continue to rely primarily on outdated modalities, such as fundus photography, as the sole measure of the presence of diabetic eye disease. 

More Than Vasculopathy

DR affects more than a third of patients with diabetes.^4,5 Ophthalmologists can vastly expand their diagnostic capabilities using new technologies and by embracing the growing evidence about how preclinical diabetic retinopathy (PCDR), DN, and diabetic choroidopathy (DC) affect vision. We now know that diabetic eye disease is more than vasculopathy of the retinal circulation, and that it affects the choroidal circulation, neurons and support glial, with earlier damage than we measure by the diabetic retinopathy severity scale (DRSS).⁴

Previous guidelines have focused on monitoring the clinically visible damage associated with DR measured by the DRSS. Those guidelines were developed before we knew about DN and DC. Today, using ultrawide field swept source optical coherence tomography angiography (UWF-OCT-A) can help identify PCDR.⁶ With the Early Treatment Diabetic Retinopathy Study (ETDRS) now nearing its 40th birthday, it is still the standard for DR — but it soon will be enhanced by UWF as validations are completed with the addition of peripheral fields. 

Understanding the mechanism of action behind diabetic eye disease development is instrumental in understanding how to monitor and combat it. What researchers now know is that the disease is transretinal and panretinal, and causes damage well before observable DR. Mediators at a biochemical and cellular level cause the progression of diabetic eye disease, but treatments that target these various mechanisms can potentially stop them — quite a departure from Dr Beetham’s predictions.

A novel therapy in the pipeline — an apurinic/apyrimidinic endonuclease 1 and reduction-oxidation effector factor-1 (APE1/RIF-1) inhibitor — targets some of the pathways that may provide an oral treatment to slow DR. With this in mind, the importance of the early detection of choroidal and neuronal damage caused by hyperglycemia is even more crucial, as the hope for treatment gives physicians the impetus to uncover early signs of PCDR, DN, and DC.⁵

Hyperglycemia and PCDR Go Hand-in-Hand

The hyperglycemia experienced by patients at risk for diabetes can lead to basal lamina thickening in PCDR. As the hyperglycemia progresses, it is followed by a process of pericyte and endothelial loss called vasoregression. The hyperglycemia targets the retinal neurovascular bundle — composed of the lamina basalis, astrocyte endplates, pericytes, and endothelial cells that create the capillary with tight junctions that form the inner blood retinal barrier (iBRB).⁴

Hyperglycemia damages this neurovascular bundle in several ways, including by initiating vascular permeability, neuronal apoptosis, astrocyte dysfunction, and micro glial activation with macrophages. The retinal capillaries undergo accelerated apoptosis even before histological changes are found.⁶ Once this hyperglycemia disturbs the healthy retina, it initiates Müller cell dysfunction (MCD). Muller cells provide organizational structure to the neural retina.

MCD affects transmitter recycling, metabolic symbiosis, free radical scavenging, and leads to an overproduction of matrix metalloproteinases (MMPs). MMP degrades the protein occludin and weakens the iBRB, leading to vascular leakage and macular edema.⁸ 

Additionally, MCD disrupts potassium currents through downregulation of kir4.1 channels, affecting fluid equilibrium without the upregulation of aquaporin-4 channels. With continued potassium uptake and no release into the microvasculature, Müller cells begin to swell. This combined with a decrease in fluid removal, contributes to diabetic macular edema (DME). 

Advanced glycation endproducts (AGEs) are a toxic byproduct of hyperglycemia. AGE-modification of the laminin leads to a decrease in Kir4.1 channels, thereby resulting in MCD edema. AGEs are modifications of proteins, lipids, and nucleic acids from hyperglycemia.⁹ Numerous other pathways are also affected and can produce different toxic hyperglycemic byproducts which lead to oxidative stress, causing neurodegeneration, inflammation, and vascular hyperpermeability. Reactive oxygen species (ROS) damage tissues, despite vascular endothelial growth factor’s (VEGF) anti-apoptotic properties with subsequent endothelial death. Mitochondrial DNA are subject to ROS oxidative damage and this affects function and cellular health.⁹

These findings present via measurable biomarkers, such as ellipsoid zone reflectivity, and may predate the clinical start of DR defined by the ETDRS. Furthermore, the DN causes MCD, which affects the retinal structure and normal physiological processes that are included in our current understanding of DME.

How PCDR Transitions to NPDR

The damage then transitions from PCDR into NPDR, affecting the vascular plexi with the formation of microaneurysms (MA).⁷ There is an area of scientific uncertainty on the initiation of the pathogenesis of vasoregression: is it purely due to VEGF causing endothelial hypertrophy or does leukostasis contribute, causing lumen thinning via leukocyte adhesion and plugging mediated by cytokines or a combination of both mechanisms?^9-11 Prevention of leukostasis mediated vasoregression may be possible with ICAM-1 and CD18 targeting.¹¹

On a biochemical level, hyperglycemia creates methlyoxil, which affects angiopoietin-2 (Ang-2) and VEGF. These agents begin the transition from PCDR to the disease of early nonproliferative diabetic retinopathy (NPDR) with inflammatory biomarkers interleukin (IL) IL-6 and IL-1 beta, which both increase retinal capillary apoptosis. This phase of NPDR is evident on fundus exam and affects the vascular plexi.^12,13

Recently published research shows that, if NPDR is present in response to low glucose, the cells increase levels of a transcription factor, called hypoxia-inducible factor (HIF)-1α. This leads to glucose transporter protein type 1 (GLUT1) gene expression necessary to improve the neurons’ ability to utilize available glucose and preserve oxygen available for energy production.¹⁴

However, in low-oxygen environments of the damaged neurovascular bundle, this normal, physiologic response to low glucose triggers the release of excessive HIF-1α protein into the nucleus, leading to an increase of VEGF and ANGPTL4, thus hypoglycemia worsens existing NPDR.¹⁴

A Role For Ultrawide Field Imaging 

Now that you’ve got a sense of the anatomical journey toward diabetic eye damage, we’re left with the question of what can be done to better manage these patients. The standard for evaluation of DR had previously been a dilated fundus exam with ETDRS-7-standard field (SF), which images only 34% of the retina and can identify 40% DR pathology.⁵ The new standard is UWF which includes the ETDRS-7SF and 5 peripheral fields. Compared with ETDRS-7SF, UWF is superior in several ways.⁵

For one, 10% of DR lesions develop on the periphery, outside ETDRS-7SF guidelines. With UWF, physicians can image 3.2 times more retina surface area, 3.9 times more nonperfusion, 1.9 times more neovascularization, 3.8 times more PRP.^5,15 All these statistics show that UWF is particularly adept at imaging the predominantly peripheral lesions (PPL) that are so threatening to patients with diabetes. 

In fact, the peripheral field has a larger surface area, and therefore more pathology, than its corresponding central field, another driver of more rapid progression to PDR. 

The big recent revelation in diabetic eye care is the unveiling of the long-awaited Diabetic Retinopathy Clinical Research Retina Network (DRCR.net) Protocol AA, which overturned some prior thinking developed after smaller studies but still supports the application of ultra-widefield fundus fluorescein angiography (UWF-FA).^{16, 17} Smaller studies found color PPL was predictive of a 2-step worsening in DRSS, but Protocol AA found no significant association between color PPL and a 2-step worsening and surprisingly secondary analysis found no association between color PPL and progression to PDR over 4 years (Table 1).^16,17

Protocol AA still supports color UWF imaging because A) it reveals more DR lesions; and, more importantly, B) it increases graded DRSS by 12.5% in eyes when compared with ETDRS-7SF over 4 years.¹⁷

Table 1. Conclusions from the DRCR.net Protocol AA

MA and hemorrhage presence of PPL increase the risk DR progression by 3.2 fold and onset of PDR by 4.7 fold.

PPL is not an independent risk factor for PDR progression.

MA turnover predictive of progression to PDR.

FA-PPL 1.7 fold greater risk of disease worsening over 4 years.

FA-PPL on UWF-FA ETDRS 2-step worsening of diabetic retinopathy severity scale (DRSS) and need of treatment over 4 years.

Artificial intelligence (AI) may have a role in DR detection in the future. Based on recent advances in imaging and deep learning, patients may one day be able to bypass an in-office dilated fundus exam as a definitive clinically measure DR.¹⁸

Ophthalmologist As Educator

Physicians can share these important findings to motivate patients and guide their healthcare team to begin treatment of DM earlier to potentially prevent preclinical damage. Clinicians should be asking themselves if their patients’ “prediabetes” is actually “early diabetes,” and if they are optimally controlling the glycemia based on this new knowledge of glycemic damage.

Ophthalmologists have a role to play by advising those comanaging the glycemia of patients with NPDR — particularly dieticians and primary care physicians. They need to know that glucose fluctuation that is too low can be detrimental to the retina when damage is present..

In the near future, it will be possible for patients to check their glucose without a prick, giving them a vital tool for limiting the glycemic damage from glucose fluctuations. 

Ophthalmologists must emphasize that maintenance of patients’ physiological vitamin D levels is necessary to reduce the progression of hyperglycemia to DM.

You can recommend to patients with DM a daily aspirin because of its anti-inflammatory properties. There is evidence that rheumatoid arthritis patients taking daily aspirin had less DR.¹⁹

A dietary recommendation for patients with DM would be the addition of omega-3 EPA and DHA for the reduction of ROS. A recent study showed eating fatty fish twice a week decreased the likelihood of DR in T2DM.⁴

With these new findings regarding PPL monitoring and the prospect of continuous glucose monitoring, the tools to keep diabetes from robbing our patients of vision is within reach. Applying these new modalities is likely to become the new par — and a robust understanding of the interplay between hypoglycemia and retinopathy will be elemental in using these tools to manage visual health.