Diabetic retinopathy (DR) is widely considered one of the earliest and most common microvascular complications of diabetes (1, 2). Although microvascular damage is the classic hallmark of DR, there is emerging evidence to suggest that retinal neurodegeneration may precede retinopathy (3). Indeed, studies have shown that hyperglycemia affects the neural layers of the retina, leading to visual symptoms such as reduced contrast sensitivity, abnormal dark adaptation and reduced visual acuity (4, 5).

The current gold standard methods for identifying retinal neural dysfunction and damage are multi-focal electroretinogram (mfERG) and frequency domain optical coherence tomography (FD-OCT). Studies have shown altered mfERG responses (6, 7), retinal nerve fibre layer (RNFL) thinning (8, 9) and a loss of central visual field sensitivity (4, 8, 10) in patients without retinopathy (11). Another study using OCT imaging of the retina reported a high correlation between perimetric function loss and retinal neuropathy in patients with no or mild DR (12). Thus, functional testing can help understand patients’ symptoms prior to DR and may be used to counsel patients about the importance of optimizing metabolic control. It is interesting that the latest position statement by the American Diabetes Association defined diabetic retinopathy as neurovascular complications rather than microvascular complications of both type 1 and type 2 diabetes (13), and the term ‘diabetic retinopathy’ has now been replaced by ‘diabetic retinal disease’ (14).

Over the past decade, a growing body of evidence has demonstrated corneal nerve fiber loss in diabetic patients through the application of corneal confocal microscopy (CCM). CCM allows quantification of corneal c-nerve fibers, enabling early diagnosis and stratification of the severity of diabetic peripheral neuropathy (DPN) (15, 16). Several cross-sectional studies have evaluated the ability of CCM to diagnose clinical levels of DPN in comparison to a range of other diagnostic tests (17) and a stepwise deterioration in corneal nerve morphology in patients with pre-proliferative and proliferative DR has been demonstrated (18–21).

Whilst there is a recognised association between the main microvascular complications, this has only been investigated in advanced cases. Indeed, the Rochester Diabetic Neuropathy Study has shown that markers of micro vessel damage such as DR and proteinuria or microalbuminuria (MA) are the strongest predictors for the severity of DPN (22). Furthermore, a strong association has been observed between cardiac autonomic neuropathy (CAN) and proliferative DR in patients with type 1 diabetes mellitis (T1DM), suggesting an underlying etiologic link (23). Although the underlying molecular mechanisms linking DR and cardiovascular disease are still under debate, similarities in their pathophysiology have been notable (24). To our knowledge, only one published study has investigated the temporal relationship between ocular neural and vascular damage. Their results suggested progressive loss in neuroretinal layer thickness in people with no or minimal DR after four-year follow-up, independent of DR or progression of DR (25). However, that study did not assess corneal nerves. Both CCM and OCT imaging demonstrate potential for the clinical application of monitoring diabetic retinal disease alongside functional assessment. The present study aimed to evaluate functional and structural changes in the retinal nerve layer and corneal c-fibres in patients with T1DM without clinical evidence of retinopathy at the time of examination

The clinical and demographic results of all participants are presented in Table 1 . None of the patients had clinical evidence of DR or nephropathy and only mild neuropathy based on the NDS score (score of less than 3).

Diabetic microvascular complications result in considerable morbidity with both microalbuminuria and the severity of DR relating to the severity of DPN (22). The temporal relationship for the development of microvascular complications has been systematically assessed in very few studies (32, 33). Recent studies have shown that hyperglycemia can affect not only the retinal vasculature but also the neuro-retina, resulting in a loss of visual acuity (4). Furthermore, studies have also shown altered multi-focal electroretinogram (mfERG) responses in patients with diabetes (6).

It is established that during diabetes, retinal glial, neural and microvascular dysfunction is interdependent and essential for the development of DR (3). Neuronal degeneration and early retinal dysfunction have been demonstrated in various animal models of diabetes (34–36). Furthermore, postmortem studies have revealed neuronal degeneration and apoptosis in the ganglion cell layer (37). Spectral-domain OCT (38) has shown a thinning of the RNFL (39, 40) and photoreceptor (41) layers in diabetic patients without retinopathy. Further thinning of the ganglion cell layer and inner plexiform layers have also been reported in patients with mild (42) and advanced DR (9). Hence, neurodegeneration may play a key role in the development and progression of DR, and structural RNFL monitoring may help to identify diabetic patients who would benefit from neuroprotection (11). Our findings demonstrated no significant global RNFL thinning in patients with diabetes when compared to healthy controls. However, significant thinning was evident in the ST, SN and IT peripapillary sectors. Sectoral thinning has been reported previously in T1DM, prior to retinopathy, with conflicting reports of the inferior (43) and superior (44) quadrant thinning occurring first. The pathophysiology of non-uniform peripapillary thinning is not fully understood. One possible explanation for thinning of the superior and inferior quadrants prior to the temporal and nasal may be due to relative RNFL sectoral thicknesses. The thicker superior and inferior quadrants may have a higher sensitivity to oxygen deprivation due to their higher metabolic demand, thus being affected more by the metabolic stress from the disease. Of course, these findings would require further investigation.

Additionally, we found significant correlations of peripapillary sectoral thinning with HbA1c and duration of diabetes. HbA1c levels are well known to be linked to microvascular abnormalities, and intensive glycaemic control therapy reduces the risk of any retinopathy developing (45). It may be that neuro-retinal damage is more sensitive to HbA1c levels in patients with T1DM, thus associating with neuro-retinal thinning prior to DR in our study. However, we recognise that glycaemic control fluctuates in many patients, and our cross-sectional study may not have captured participants’ average HbA1c level. This may explain why our findings contradict those of a previous longitudinal study in patients with T1DM, in which no relationship between progressive thinning and consistent HbA1c levels were found over a 4 year period (25).

Research has shown that several pathways are likely involved in neuronal cell death in the retina of patients with diabetes (46). Animal models and studies of intraocular fluid in eyes with diabetes have suggested elements of oxidative stress, glutamate excitotoxicity and hyperglycemia-related production of advanced glycation end-products (AGEs) as potential molecular mechanisms underlying neuro-retinal loss (47). There has been recent interest in the specific role of Müller cells in diabetic neurodegeneration. Müller glia provide critical homeostatic and trophic support for the neuronal layers and vasculature of the retina and hyperglycemia may target these cells, thus inducing neuroinflammation and leading to ganglion cell loss. A recent study by Miller and colleagues (48) assessing diabetic mice, suggested that enhancement of the protein regulated in development and DNA damage responses 1 (REDD1) expression leads to a failure of Müller cells to properly respond to the diabetic metabolic environment.

Visual field sensitivity and visual function have also demonstrated reductions in patients with diabetes (4, 8, 12, 49) prior to, or with minimal retinopathy. In the current study, we found no significant difference between the patient population and healthy controls when assessing both types of perimetric results, thus our results do not support these previous findings (42). However, in comparison to van Dijk and colleagues, we used different methods of perimetry. We also only included patients with no DR, whereas that study included patients with minimal DR. It may be that our participants had milder neurovascular changes in comparison, thus demonstrated structural changes prior to detectable functional loss.

Previous studies have also shown a stepwise deterioration in corneal nerve morphology in patients with pre-proliferative and proliferative DR (18–20), and more recently corneal nerve damage has also been detected in patients without retinopathy (18, 20, 21, 50). Our results support these previous findings with significant reductions in all three corneal nerve parameters in patients with T1DM compared to healthy controls, as well as a significant reduction in corneal sensitivity. These findings highlight the potential use of CCM imaging for detecting early neurovascular changes in patients with diabetes. Furthermore, CNFL demonstrated a significant relationship with duration of diabetes and HbA1c.

Although we found reductions in both the RNFL and corneal nerve parameters independently, we and others (51) found no correlation between neural changes in the retina (central nervous system) and cornea (peripheral nervous system). At present, the reason for this finding is unclear. It may be that corneal and retinal neurodegeneration occur as independent pathological processes. Regions of retinal non-perfusion have previously demonstrated profound inner-retinal thinning (52) and it may be that retinal microvascular changes are linked to retinal thinning but not corneal nerve degeneration. However, the argument over which causes the other within the retina is still under debate and yet to be determined (5).

Additionally, there is evidence for defective direct insulin signalling and loss of insulin-derived neurotrophic support possibly contributing to peripheral neuropathy in patients with type 1 diabetes (53). One study reported a rapid improvement of epidermal innervation in mice paws, through doses of insulin insufficient to alter glycemia (54). Another study which applied daily insulin to the eyes of both control and streptozotocin-diabetic mice without altering blood glucose levels found that it successfully prevented depletion of nerves of the sub-basal plexus (55). This evidence suggests that the neuroprotective role of insulin can be independent of effects on hyperglycemia and local manipulation of insulin signaling may offer novel therapeutic peripheral nerve repair.

Based on our definition, 16.7-27.8% of patients were categorized as ‘abnormal’ using CCM, depending on the nerve parameter used for definition, and 32% based on RNFL thickness alone. These values suggest that several subjects who do not meet the referral criteria into the hospital eye service (HES) based on retinopathy may require closer monitoring based on ocular nerve changes. The future approach for developing new pharmacological agents for DR and other microvascular complications of diabetes could also use a combined approach for targeting neurodegeneration, inflammation, and oxidative stress (56), focusing on the potential of novel therapies for neuroprotection (57).

The findings of this preliminary study may not only apply to diabetic neuropathy but also other to systemic neurological conditions. Subjects with diabetes are known to have a higher prevalence of dementia and other neurodegenerative diseases, such as Parkinson’s disease in comparison to age-matched controls (58). Additionally, a number of pathways that are involved in these neurodegenerative diseases have been found to be altered in the retinae of patients with diabetes (59). A recent study found that testing patients with T2DM using microperimetry could identify a risk of developing Alzheimer’s disease, with retinal sensitivity relating to cognitive status (60). Thus, there may also be scope for the application of neuro-retinal imaging and corneal nerve imaging to identify patients with diabetes who are at risk of this development, due to the relationship with brain neurodegeneration.

Although we used a different type of perimetry measurement, in our study we found no significant changes in visual field results, despite RNFL thinning and corneal nerve changes. Thus, structural changes may precede functional loss in these patients, and nerve imaging detected abnormalities prior to functional assessment. Furthermore, objective measures would arguably be more sensitive and reliable than subjective tests such as perimetry.

In the present study, we demonstrate that novel in-vivo ophthalmic techniques can identify corneal and retinal neurodegeneration in patients with T1DM without clinical evidence of retinopathy. One may argue that the very fact that the patients in our study have no evidence of retinopathy, despite a long duration of diabetes, suggests that they are less prone to retinopathy and indeed the other complications. However, the demonstration of neuronal dysfunction and damage in the retina and cornea would argue against this and suggest that neuronal damage may precede vascular damage in the eye. Indeed, our study has some limitations including long duration of the disease and a small sample size. Future, prospective studies are needed to establish the exact temporal relationship between the development of retinopathy and neuropathy. This will allow to development of better targeted and more efficient therapeutic approach in the early stages of the disease.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

This research adhered to the tenets of the Declaration of Helsinki and was approved by the Manchester Research Ethics Committee and the ethics review board of the university of Manchester. The participants provided their written informed consent to participate in this study. The data analyzed in this study were originally collected at the University of Manchester between 2012 and 2015 under approval from the Manchester Research Ethics Committee and the University of Manchester Research Ethics Review Board. Ethical review and approval was not required for the present study in accordance with local legislation and institutional requirements because this study represents a secondary analysis of data previously collected with ethical approval. Written informed consent for participation was obtained from all participants at the time of original data collection.

MT designed the study, researched and analyzed the data and wrote the paper. JC, HF, MT, contributed to writing and analyzing the data. All authors contributed to the article and approved the submitted version.

This work was supported by Heidelberg Engineering UK.

Support from Prof. Andrew JM Boulton and Prof. Rayaz Malik is acknowledged for the initial study. Dr Maryam Ferdousi and Dr Ioannis Petropoulos undertook corneal confocal microscopy in a proportion of the study patients.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

A correction has been made to this article. Details can be found at: 10.3389/fendo.2026.1809367.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.