Diabetic retinopathy (DR), the leading cause of vision loss among working-age adults in the Western world, has traditionally been considered a microvascular disease. However, increasing evidence suggests that neurodegenerative changes may precede vascular damage.^(1–3)

Diabetic retinopathy involves neurodegenerative changes such as neuronal apoptosis and glial activation, which can occur even without visible vascular damage. These changes have been found in diabetic patients with normal eye exams, indicating that neurodegeneration may precede clinical signs. Diabetes increases retinal neuron apoptosis – especially in ganglion cells – resulting in thinning of the nerve fiber layer.^(4,5)

Diabetes can trigger apoptosis of retinal ganglion, amacrine, and Müller cells, likely due to chronic glutamate toxicity. Glutamate-induced excitotoxicity increases oxidative stress and inflammation, contributing to retinal neurodegeneration.⁽⁶⁾ Glutamate accumulation in the diabetic retina may result from increased production by glial cells, reduced glutamate oxidation, and impaired clearance due to decreased glutamine synthetase activity in Müller cells.⁽⁷⁾

High glutamate levels lead to neuronal death by overactivating AMPA and NMDA receptors, increasing intracellular calcium and causing excitotoxicity. Diabetes also reduces neuroprotective factors like pigment epithelium-derived factor (PEDF), somatostatin (SST), and erythropoietin.⁽⁷⁾

Glutamate buildup and reduced neuroprotection increase vascular endothelial growth factor (VEGF) production, promoting blood-retinal barrier breakdown. Endothelial progenitor cell (EPC) dysfunction also contributes to microangiopathy and neurodegeneration.⁽⁸⁾

Glutamate buildup in DR is caused by reduced glutamine synthetase activity, impaired glutamate oxidation, and decreased glial clearance. These changes lead to excitotoxicity, a major factor in retinal neurodegeneration.^(9–11) Neural dysfunction in DR can occur before microvascular damage, even without pericyte loss, challenging the view that vascular changes are the earliest signs of the disease.^(12–14)

Diabetes-related neurodegeneration also affects avascular tissues like the cornea. Confocal microscopy shows reduced corneal nerve fiber density in patients with diabetic neuropathy, indicating neurodegeneration independent of vascular damage.^(15,16) Neuronal changes in the retina can occur simultaneously with vascular alterations or may emerge prior to the development of microcirculatory dysfunction.⁽¹⁷⁾

The aim of this narrative review was to provide a comprehensive summary of the current evidence regarding neurodegenerative changes in DR.

We conducted a comprehensive literature review covering the period from August 15, 1998, to June 20, 2025. Searches were carried out in the MEDLINE® database, via PubMed®, and Latin American and Caribbean Health Sciences (Lilacs), via the Virtual Health Library, to identify relevant studies on the neurodegenerative manifestations of diabetes. To ensure a thorough and sensitive search strategy, we used a variety of keyword and MeSH term combinations, including: ("neurodegeneration") AND ("diabetic retinopathy"); "diabetic retinopathy".

Titles and abstracts of all retrieved citations were screened for relevance and eligibility according to predefined inclusion criteria. Two duplicate articles identified across both databases were excluded. Following this, all publications categorized as reviews, case reports, randomized controlled trials, meta-analyses, and guidelines were selected, resulting in a final pool of 236 articles. From these, 73 articles were chosen for an in-depth qualitative analysis on the topic. Details of the search strategy are presented in table 1.

The cross-sectional study from the EUROCONDOR consortium evaluated 449 patients with type 2 diabetes to investigate early neurodegenerative changes in DR and their relationship with microvascular damage. Neural function assessed by multifocal ERG (mfERG) showed that 58% of patients had dysfunction despite no visible retinopathy, with the P1 component of mfERG being a sensitive marker. Structural changes detected by OCT were correlated with functional deficits in early stages, but 32% showed no detectable changes, highlighting disease heterogeneity. Another EUROCONDOR study with 440 patients found that venular caliber increased in advanced DR stages, associated with thinning of neural retinal layers, indicating inflammation and blood-retinal barrier dysfunction. These findings reinforce DR as a neurovascular disease and suggest that non-invasive vessel measurements may serve as early biomarkers for diagnosis and personalized therapies.^(30,31)

Animal models of diabetes indicate that neural damage in DR may arise before noticeable vascular abnormalities. In mice, inner retinal thinning and electrophysiological deficits are observed prior to visible microvascular pathology. In humans, although clinical evidence is still evolving, imaging techniques like optical coherence tomography (OCT) show early retinal layer thinning associated with DR progression, while functional assessments reveal subtle neuronal dysfunctions. OCT angiography (OCT-A) can detect early microvascular disturbances, such as decreased capillary density and enlargement of the foveal avascular zone, although findings are not always consistent across studies. Overall, current evidence suggests that neural and vascular changes may develop concurrently and interactively in the initial stages of DR.^(32,33)

Patients with type 2 diabetes have a higher risk of neurodegenerative diseases such as Alzheimer's. Since the retina originates from brain tissue, it provides a non-invasive way to monitor these changes. Diabetes and Alzheimer's share mechanisms including insulin resistance, inflammation, oxidative stress, and accumulation of AGEs, affecting both the brain and retina. Retinal imaging tools like OCT and microperimetry can detect early neurodegenerative signs linked to cognitive decline. In DR, neuronal damage– marked by gliosis and apoptosis of retinal cells–occurs before vascular changes, causing inner retinal thinning visible on OCT and early functional deficits detected by mfERG, which are associated with reduced visual function and quality of life.⁽³⁴⁾

There is growing interest in using imaging methods such as OCT, OCT-A, and fluorescein angiography to detect retinal inflammation in vivo. Hyperreflective foci (HF), possibly activated microglial cells, increase with the severity of DR and diabetic macular edema (DME), correlating with retinal dysfunction. Vitreous HF and macrophage-like cells near vessels are also linked to disease progression. Specific DME patterns on OCT, such as subretinal fluid (SRF), indicate inflammation and respond better to corticosteroids (e.g., dexamethasone) than to anti-VEGF therapy. Choroidal biomarkers like choroidal HF and the choroidal vascularity index (CVI) help monitor treatment response. Macular perfusion is reduced in DME, especially in the deep capillary plexus. Despite technical limitations of OCT-A, corticosteroids appear to reduce vascular density without worsening ischemia, but further studies are needed.⁽³⁵⁾

Dysfunction of the NVU is a critical early event in DR. Visual impairment often occurs before structural damage, highlighting the importance of neuroprotection to preserve visual quality. Advanced diagnostic tools such as mfERG and Fourier-domain optical coherence tomography (FD-OCT) are effective in detecting early neurodegeneration. Simpler, cost-effective methods like microperimetry, portable ERG, and OCT-A are also promising for clinical application. Emerging therapies targeting both neural and vascular damage – including GLP-1 receptor agonists, endothelin-1 blockers, and topical medications – are being explored. Overall, early DR management is shifting towards individualized strategies focused on protecting the NVU.⁽⁵²⁾

While tight glycemic control reduces the risk of DR, overly aggressive glucose lowering may increase cardiovascular problems. Hemoglobin A1c (HbA1c) alone does not account for all DR progression, which involves additional mechanisms like hyperglycemia-induced neuronal apoptosis and glutamate excitotoxicity. Neuroprotective agents (e.g., latanoprost, memantine) and anti-inflammatory treatments (like minocycline) are promising in preclinical studies. Other strategies include protein kinase C (PKC) inhibition and localized delivery of neurotrophic factors (nerve growth factor [NGF], PEDF, BDNF, SST, topical insulin). Although animal studies are encouraging, clinical trials with functional vision outcomes are needed. DR should be treated as a complex sensory neuropathy requiring targeted ocular therapies.⁽⁵³⁾

Diabetic retinopathy initiates with early neurodegeneration before vascular damage, making NVU protection a crucial early intervention. Since multiple pathways contribute to neuronal death, combined neuroprotective treatments may be more effective. Axonal damage is mostly irreversible, limiting regeneration. Topical agents like brimonidine, SST, and citicoline with vitamin B12 are promising in preserving retinal function safely. Experimental therapies targeting glial activation and oxidative stress, such as arimoclomol, are being studied. Improved retinal imaging aids early detection and treatment guidance. More research is needed to confirm the effectiveness of these therapies.⁽⁵⁴⁾

Recent studies show that DR involves multiple regulated cell death pathways – such as pyroptosis, ferroptosis, and necroptosis – that promote inflammation and disease progression beyond classic apoptosis. Since neuronal and pericyte loss are irreversible and disrupt retinal function, treatments must target these diverse death mechanisms for effective neuro- and vasoprotection. DR is now recognized as a complex disorder affecting neurons, glial cells, immune responses, and the retinal pigment epithelium. Traditional therapies focus on vascular complications but are often invasive and costly. Emerging treatments aim to protect the NVU using neurotrophic factors (PEDF, BDNF, IGF-1, and SST), antioxidants, anti-inflammatory agents, stem cells, and inhibitors of iNOS and endothelin-1 receptors. While promising, these therapies require further research to confirm long-term safety, efficacy, and optimal delivery methods before clinical use.^(55,56)

Recent advances identify diabetic retinal neurodegeneration as a crucial component of diabetic retinal disease, involving molecular markers linked to inflammation, oxidative stress, and cell death. Neurodegeneration can occur despite stable HbA1c due to glucose fluctuations. Imaging techniques like OCT, microperimetry, and OCT angiography detect retinal structural and functional changes, highlighting the connection between neural and vascular damage. Experimental treatments include anti-inflammatory agents, antioxidants, neuroprotective drugs, stem cells, flavonoids, and RAS modulators. Diabetic retinal neurodegeneration results from oxidative stress, inflammation, AGEs, DNA damage, and programmed cell death. Although promising in preclinical studies, clinical application is difficult due to disease variability and diagnostic challenges. Multimodal therapies targeting several pathways and early neuroprotection are essential for vision preservation.^(57–65)

Diabetic retinopathy begins with early neuronal and axonal injury. Topical neuroprotective treatments like SST and brimonidine showed limited success in the EUROCONDOR trial, possibly due to short follow-up and insensitive endpoints. However, combination therapies targeting multiple cell death pathways, such as citicoline, TUDCA, and NT-4, show better neuroprotection and nerve regeneration. Future research should focus on these multimodal neuroprotective and regenerative approaches.^(57–67)

Optical coherence tomography (OCT) offers detailed, noninvasive imaging of retinal layers, revealing ganglion cell loss, neuronal apoptosis, glial activation, and inner retinal thinning. Type 2 diabetes may accelerate nerve fiber layer thinning. The close interaction between neurons, glia, and blood vessels explains the concurrent neuronal and microvascular damage – such as pericyte loss, basement membrane thickening, and capillary closure – contributing to vision loss.^(58–69)

Diabetic retinopathy is a complex neurovascular disorder where neurodegeneration often occurs before vascular damage becomes apparent. Advances in imaging enable earlier diagnosis, but clinical use of neuroprotective and anti-inflammatory treatments faces challenges such as disease variability, late diagnosis, and drug delivery issues. Combining neuroprotection, vascular support, and regeneration, guided by personalized diagnostics, offers the best chance to slow disease progression. Future efforts should focus on standardized methods, long-term trials, multimodal therapies, sensitive outcome measures, and multidisciplinary collaboration to improve patient care and preserve vision.