Diabetologia (2023) 66:1614–1621 https://doi.org/10.1007/s00125-023-05940-5 FOR DEBATE What else can we do to prevent diabetic retinopathy? Rafael Simó1,2,3 · Cristina Hernández1,2,3 Received: 4 January 2023 / Accepted: 31 March 2023 / Published online: 6 June 2023 © The Author(s) 2023 Abstract The classical modifiable factors associated with the onset and progression of diabetic retinopathy are the suboptimal control of blood glucose levels and hypertension, as well as dyslipidaemia. However, there are other less recognised modifiable factors that can play a relevant role, such as the presence of obesity or the abnormal distribution of adipose tissue, and others related to lifestyle such as the type of diet, vitamin intake, exercise, smoking and sunlight exposure. In this article we revisit the prevention of diabetic retinopathy based on modulating the modifiable risk factors, as well as commenting on the potential impact of glucose-lowering drugs on the condition. The emerging concept that neurodegeneration is an early event in the development of diabetic retinopathy points to neuroprotection as a potential therapeutic strategy to prevent the advanced stages of the disease. In this regard, the better phenotyping of very early stages of diabetic retinopathy and the opportunity of arresting its progression using treatments targeting the neurovas- cular unit (NVU) are discussed. Keywords Diabetic retinopathy · Modifiable risk factors · Neurovascular unit Abbreviations Introduction DME Diabetic macular oedema DPP-IV Dipeptidyl peptidase IV Diabetic retinopathy remains a leading cause of prevent- ERG Electroretinogram able visual impairment and blindness in the working age EUROCONDOR European Consortium for the Early population [1]. In addition, it is one of the most common Treatment of Diabetic Retinopathy and most feared complications of diabetes, which imposes FFA Fundus fluorescein angiography a severe burden on patients, healthcare systems and society NVU Neurovascular unit throughout the world. In addition to controlling the modifi- NPDR Non-proliferative diabetic retinopathy able risk factors, current treatment is based on laser photoco- NSAID Non-steroidal anti-inflammatory drugs agulation and intravitreal injections of glucocorticoids, anti- OCT Optical coherence tomography vascular endothelial growth factor (VEGF) drugs or other PDR Proliferative diabetic retinopathy anti-angiogenic agents [2]. These treatments are aggressive, SIDD Severe insulin-deficient diabetes with significant side effects, and are used to prevent severe VEGF Vascular endothelial growth factor visual impairment and blindness in the very advanced stages of the disease. Over the last decade, the concept that neuro- degeneration is an early event in the pathogenesis of diabetic retinopathy has strongly emerged and, therefore, strategies based on neuroprotection seem a promising new avenue for * Rafael Simó treating the early stages of the condition [2, 3]. rafael.simo@vhir.org In this article the prevention based on modulating the 1 Department of Medicine, Universitat Autònoma de modifiable risk factors is revisited, and the potential impact Barcelona, Barcelona, Spain of antidiabetic drugs on diabetic retinopathy is commented 2 Vall d’Hebron Research Institute (VHIR), Vall d’Hebron on. Finally, the opportunity of arresting diabetic retinopathy University Hospital, Barcelona, Spain progression at very early stages based on treatments aimed 3 CIBER of Diabetes and Associated Metabolic Diseases at preventing or reversing the early impairment of the neu- (CIBERDEM, ID CB15/00071), Instituto de Salud Carlos III rovascular unit (NVU) is discussed. (ISCIII), Madrid, Spain 13 Vol:.(1234567890) Diabetologia (2023) 66:1614–1621 1615 Revisiting modifiable risk factors Blood pressure The classical modifiable factors associated with the onset Several randomised controlled trials have demonstrated the and progression of diabetic retinopathy are the suboptimal benefits of blood pressure control as a major modifiable fac- control of blood glucose levels and hypertension, as well as tor for diabetic retinopathy incidence and progression [12, the presence of dyslipidaemia. However, there are other less 13]. Overall, the results of these studies suggest that, while recognised modifiable factors that can play a relevant role, uncontrolled hypertension could lead to the onset or progres- such as the abnormal distribution of adipose tissue, as well sion of diabetic retinopathy, slight or mild hypertension does as lifestyle considerations such as the type of diet, vitamin not significantly affect the condition. intake, exercise and smoking, among other factors. Regarding the differential effect of anti-hypertensive drugs on diabetic retinopathy, there is only weak clinical evi- Glycaemic control dence that ACE-i (ACE inhibitors) and angiotensin II recep- tor (ARA-II) antagonists exert additional beneficial effects The most relevant risk factor for the development of diabetic in slowing diabetic retinopathy progression independently of retinopathy is suboptimal glycaemic control. There is robust their hypotensive properties [14, 15]. Therefore, the reduc- evidence supporting the relationship between blood glucose tion of blood pressure is the key point, rather than the type levels and the development and progression of diabetic retin- of anti-hypertensive drug used to achieve this. opathy [4, 5]. In fact, the higher the reduction of H­ bA1c the As is the case with metabolic control, there is no informa- higher the effect in preventing or arresting the development/ tion yet regarding the effect of tight control of blood pressure progression of diabetic retinopathy. However, there is, as on retinal neurodysfunction or neurodegeneration. yet, no information regarding the long-term effect of blood glucose control on retinal neurodysfunction or neurodegen- Dyslipidaemia eration in terms of prevention or even potential regression. The role of glycaemic variability on diabetic retinopathy Dyslipidaemia is generally considered to be a risk factor is less clear but time in range has been associated with all for diabetic retinopathy progression, but with less signifi- the stages of diabetic retinopathy in type 2 diabetes [6]. In cance than hyperglycaemia or hypertension. It has been addition, a systematic review and meta-analysis showed that reported that statins reduce the incidence of hard exudates fasting plasma glucose variability was strongly associated and microaneurysms, and decrease vision loss [16, 17]. In with the risk of development and progression of diabetic addition, fenofibrate prevents the progression of diabetic retinopathy in type 2 diabetes [7]. Finally, H­ bA1c variability, retinopathy [18], and the Action to Control Cardiovascular which reflects longer-term glucose variability, contributes Risk in Diabetes (ACCORD)-EYE study demonstrated that to the risk of diabetic retinopathy in both type 1 and type the combination of fenofibrate and simvastatin diminished 2 diabetes [8, 9]. On this basis, it could be postulated that diabetic retinopathy by approximately 40% in comparison glucose-lowering agents with more capacity to mitigate glu- with participants taking simvastatin alone [19]. However, cose variability, such as current long-acting insulins or the these results seem unrelated to the effects on lipids, and next generation of ‘smart insulin’, could reduce the risk of could be mainly accounted for by the anti-inflammatory and diabetic retinopathy. However, specific studies addressed to antioxidant actions of fenofibrate [3, 20]. answer this question are needed. The early worsening of diabetic retinopathy due to the Obesity rapid improvement of hyperglycaemia is a topic that has been revisited in recent years [10]. This has been found in A meta-analysis of prospective studies published in 2018 both type 1 and type 2 diabetes and, in addition to reductions revealed that obesity increased the incidence of non-prolifer- in glucose levels following treatment with insulin and other ative diabetic retinopathy (NPDR), but not PDR in individu- glucose-lowering drugs, it has also been reported after diet or als with type 2 diabetes [21]. In addition, studies performed bariatric surgery, thus supporting the concept that the velocity in Asian populations indicate that isolated abdominal obe- of blood glucose reduction is more important than any par- sity (i.e. high waist circumference in the absence of obesity ticular glucose-lowering strategy. The main factors associated in terms of BMI) is associated with diabetic retinopathy with early worsening of diabetic retinopathy are the magni- [22, 23]. This finding is significant because it underlines tude of the reduction in ­HbA1c (i.e. H­ bA1c reduction >1.5% the potential role of proinflammatory cytokines produced within 3 months or >2% within 6 months) and the presence of by adipocytes and macrophages infiltrating visceral adipose pre-existing diabetic retinopathy, particularly pre-proliferative tissue on the pathogenesis of diabetic retinopathy. and proliferative diabetic retinopathy (PDR) [10, 11]. 13 1616 Diabetologia (2023) 66:1614–1621 Another obesity-related risk factor for diabetic retinopa- retinopathy. Dark fish and fortified foods such as milk or thy is neck circumference, a useful predictor of obstructive cereals are important sources of vitamin D. sleep apnoea [24], which in turn is associated with the pres- It is worth mentioning that vitamin D in the eye is locally ence and severity of diabetic retinopathy [25]. Individuals produced, activated and regulated. In fact, the eye expresses with diabetes who have a larger neck circumference are all the enzymes involved in vitamin ­D3 production [32]. This more likely not only to have diabetic retinopathy, but also is important because 1α,25-dihydroxyvitamin ­D3 inhib- to progress to more severe stages [26]. This is because the its angiogenesis and this effect is mediated by vitamin D recurrent episodes of hypoxaemia due to apnoea stimulate receptors, which are also essential during retinal vascular angiogenic factors such as VEGF. development [33]. It should be noted that the new subclassification of indi- Vitamin A is essential in visual processes and is a main viduals with adult-onset diabetes shows that Group 2, severe component of rhodopsin. So, the regular consumption of insulin-deficient diabetes (SIDD), which includes individu- foods high in vitamin A or carotenoid precursors, such as als with high ­HbA1c, impaired insulin secretion and moder- kale, spinach, broccoli, carrots or sweet potatoes, is recom- ate insulin resistance, have the highest incidence of diabetic mended. Vitamin ­B1 is also a potent free radical scaven- retinopathy [27]. This finding suggests that insulin defi- ger that prevents activation of the polyol pathway. Forti- ciency could play a key role in the pathogenesis of diabetic fied breakfast cereals are an important source of vitamin retinopathy and raises the question of whether SIDD patients ­B1. Finally, vitamin C (ascorbic acid) is also involved as a would benefit for early intensive insulin therapy for the pre- protective factor against diabetic retinopathy development. vention of diabetic retinopathy. Vitamin C has antioxidant and anti-angiogenic actions and improves endothelial function. In 2010, we reported that patients with PDR presented lower intravitreous levels of Lifestyle ascorbic acid than non-diabetic individuals [34], which was confirmed more recently by other authors [35]. This is Type of diet and vitamin intake Fish consumption seems because ascorbic acid and glucose compete to use GLUT-1 important in preventing or arresting diabetic retinopathy. to reach the retina and, therefore, the higher the glucose In middle-aged and older individuals with type 2 diabetes, levels, the lower the uptake of ascorbic acid by the retina. intake of at least 500 mg/day of dietary long-chain n-3 poly- So, apart from recommending a regular consumption of unsaturated fatty acids, easily achievable with two weekly foods with high content of vitamin C, such as citrus fruits, servings of oily fish, is associated with a reduction in relative good glycaemic control is also essential to maintain correct risk (48%) of sight-threatening diabetic retinopathy [28]. In intraretinal levels of ascorbic acid. addition, dark-meat fish (e.g. salmon, mackerel, swordfish, sardines or bluefish) consumed weekly (85–141 g) vs never Smoking Another factor related to lifestyle is tobacco smok- was related to an almost 70% reduced likelihood of develop- ing. There is controversial evidence regarding smoking as ing diabetic retinopathy [29]. a risk factor for the development or progression of diabetic retinopathy. Notably, a recent study found an association A Mediterranean diet is also associated with prevention of between smoking habit and deleterious findings in angio- diabetic retinopathy development. The PREvención con optical coherence tomography (OCT) in patients with no DIeta MEDiterránea (PREDIMED) study showed that a visible diabetic retinopathy, so in very early stages of the Mediterranean diet enhanced with extra virgin olive oil condition [36]. Although other studies have had conflicting reduced the risk of incident diabetic retinopathy by 40% results, this should not alter the message about the impor- [30]. Extra virgin olive oil may be particularly beneficial tance of smoking cessation. because it is extracted from the olive fruit without using heat or chemical solvents. This mechanical process retains the Physical activity As regards exercise, it has been highest amount of natural phenolic compounds, which have reported that regular physical activity reduces the risk antioxidant and anti-inflammatory properties. of developing diabetic retinopathy [37]. However, it An important component of diet is the vitamin content. should be underlined that individuals with PDR should It has been consistently reported that adequate vitamin D avoid high-intensity aerobic and resistance exercise, status prevents the development of diabetic retinopathy [29]. and physical activities comprising Valsalva manoeu- In addition, several reports have found that individuals with vres (which increase systolic blood pressure). This is diabetes present with lower levels of vitamin D than the to reduce the risk of vitreous haemorrhage and retinal non-diabetic population [31]. Therefore, it is important to detachment. Specific sports such as boxing, high altitude monitor the vitamin D levels in patients with diabetes, not mountaineering and scuba diving should also be avoided only for the prevention of osteoporosis but also diabetic in individuals with PDR. 13 Diabetologia (2023) 66:1614–1621 1617 Sunlight exposure Another potential risk factor for relies on the complex impairment of all the components developing diabetic retinopathy is related to the del- of the NVU: macro- and microglia, neurons and vascular eterious effect of ultraviolet B radiation on retinal components (endothelial cells and pericytes). Among these pigment epithelium. Although a moderate exposure cells, the glial activation plays an essential role in linking to sunlight is recommended to maintain circulating neuronal damage with early vascular impairment, which vitamin D levels, a recent study has shown that sun- comprises the breakdown of the blood–retinal barrier and, light exposure for ≥5 h a day is significantly asso- consequently, vascular leakage [39]. ciated with an increased risk of diabetic retinopathy [38]. The underlying mechanism includes the accu- mulation of reactive oxygen species that can damage Experimental evidence retinal blood vessels causing retinal capillary apop- tosis, hypoxia and neovascularisation. This evidence One of the more logical strategies for treating diabetes- suggests that reduction in sunlight exposure could be induced retinal neurodegeneration is ‘replacement’ treat- a preventive strategy against diabetic retinopathy in ment of the neurotrophic factors that are downregulated people with diabetes, and opens up the possibility of in the early stages of diabetic retinopathy, such as pigment testing sunglasses as a tool for this purpose in appro- epithelial derived factor, somatostatin and glucagon-like priate clinical trials. peptide 1 [3]. Treatment with these neurotrophic factors has prevented the development of diabetic retinopathy in Impact of glucose‑lowering drugs on diabetic experimental models. Dipeptidyl peptidase IV (DPP-IV) retinopathy inhibitors have also provided beneficial effects on NVU [40]. Although the enhancement of intraretinal levels of The strong relationship between the reduction of ­HbA1c and GLP-1 seems a relevant mechanism accounting for the the beneficial effects on diabetic retinopathy has obscured beneficial effects of DPP-IV inhibitors, the activation of the necessity of performing clinical trials investigating the other pathways related to DPP-IV inhibition cannot be specific effects of glucose-lowering drugs per se on dia- ruled out. In addition, the lower cost and higher stability betic retinopathy, independently of their effectiveness in of DPP-IV inhibitors in comparison with GLP-1, could reducing blood glucose levels. Therefore, we do not have mean they are excellent candidates for clinical develop- robust information on this issue. This is very different from ment. Presynaptic proteins, which are crucial for neuro- the available information that we have regarding the effects transmission and synaptic homeostasis, as well as pro- of glucose-lowering drugs on cardiorenal events, which teins involved in axonal transport, are also downregulated have been widely studied. For this reason, studies aimed at in the diabetic retina [41, 42]. Therefore, a replacement examining the specific role of glucose-lowering drugs on treatment or therapeutic strategies addressed to avoid the the development and progression of diabetic retinopathy diabetes-induced intraretinal reduction of these neuro- are needed. To the best of our knowledge only one clini- transmitters could exert beneficial effects. However, fur- cal trial is currently ongoing to examine this issue. This is ther research in this field is needed. the FOCUS study (NCT03811561) a Phase III randomised Another successful strategy has been to block the clinical trial which will look at the long-term effects of endothelin-1 receptors ETB-R and ETA-R. By blocking semaglutide on diabetic eye disease when compared with these receptors, bosentan administered using eye drops placebo. A total of 1500 participants with type 2 diabetes exerted a beneficial effect on both neurons (blockade will be included and the estimated study completion date of ETB-R) and microvasculature (blockade of ETA-R), is November 2027. thus preventing retinal neurodegeneration and vascular leakage [43]. New insights and perspectives on diabetic Clinical perspectives retinopathy prevention based on targeting the NVU To the best of our knowledge, there are no drugs recom- mended by any scientific society aimed at targeting the NVU The concept that neurodegeneration is an early event in the for treating the early stages of diabetic retinopathy. The use- development of diabetic retinopathy, which antedates and fulness of corticosteroids or non-steroidal anti-inflammatory participates in its pathogenesis, has led to neuroprotection drugs (NSAIDs) by topical route for treating diabetic macu- as a potential therapeutic strategy to arrest the progression lar oedema (DME) in humans have been reported [44, 45]. of the condition [3]. As is the case with the brain, retinal However, a lack of effect in reducing retinal thickness after neurodegeneration is not an isolated neuronal process and 1 year of topical administration of the NSAID nepafenac 13 1618 Diabetologia (2023) 66:1614–1621 has also been reported in patients with non-central-involved treatment gap that currently exists for early-stage diabetic DME [46]. retinopathy. Fenofibrate and calcium dobesilate (two drugs admin- However, the long-term systemic administration of istered orally) have shown to be effective and safe for the drugs has two main problems. First, they need to be able treatment of diabetic retinopathy in several clinical trials to cross the blood–retinal barrier, which could be a limit- [3], but they are not formally recommended in clinical ing factor to reach the retina at pharmacological concen- guidelines. It is worth mentioning that three large ran- trations. Second, systemic adverse effects and potential domised clinical trials aimed at evaluating the effect of pharmacological interferences with other drugs used for fenofibrate in arresting the progression of diabetic retin- the treatment of diabetes and its comorbidities is also a opathy are ongoing in the USA (NCT04661358), Australia drawback that needs to be considered. In addition, the (NCT01320345) and UK (Scotland; NCT03439345). use of repeated intravitreal injections seems a strategy These studies will provide new evidence on the useful- disproportionally aggressive for treating the early stages ness and safety of fenofibrate for treating the early stages of diabetic retinopathy. For these reasons, topical treat- of diabetic retinopathy and could be helpful to cover the ment (eye drops) targeting the NVU has emerged as a No DR NPDR PDR No visible lesions in the fundoscopic examination Hallmarks of NVU assessment NVU impairment Neuroretinal assessment NVU impairment Glial activation (GFAP IF) ERG Microperimetry OCT Microvascular assessment Selection of treatment depending on the results Neuron apoptosis Vascular leakage (TUNEL) (Evans Blue) Angio-OCT Ultra-wide field FFA Fig. 1  In the early stages of diabetic retinopathy, even before micro- show us the degree of neural and/or microvascular damage, thus vascular abnormalities can be detected in funduscopic examination, allowing a better phenotyping. This information would be important NVU impairment is already present in a significant proportion of for predicting the risk of cardiovascular events and cognitive decline, patients. The main hallmarks of this diabetes-induced NVU impair- as well as to select the most appropriate treatment for the early stages ment are: glial activation (also named reactive gliosis), neuron apop- of diabetic retinopathy, when such treatments are available. DR, dia- tosis and vascular leakage due to the disruption of the blood–retinal betic retinopathy; GFAP, glial fibrillary acidic protein; IF, immuno- barrier. At this stage, neuronal and microvascular assessment would fluorescence 13 Diabetologia (2023) 66:1614–1621 1619 new strategy for treating the early stages of the condition clear that the development and implementation of this [2, 3]. However, only somatostatin has so far been tested new and more complex screening will depend on the in a randomised clinical trial (the European Consortium appearance in the market of effective and safe drugs for for the Early Treatment of Diabetic Retinopathy [EURO- treating early-stage diabetic retinopathy. In this regard, CONDOR] study). Although the results of this study were clinical trials of drugs with experimental effectiveness in positive in terms of preventing the progression of neuro- preventing both neurodegeneration and vascular leakage, dysfunction, there was no impact on microvascular dam- such as GLP-1, DPP-IV inhibitors or endothelin block- age [47]. This could be attributed to the high proportion ers, are needed [2, 3]. In the meantime, given the high of participants included in the study with no or very mild risk of cardiovascular disease and dementia that is pre- microvascular disease, the excellent metabolic control sent among the type 2 diabetic population with diabetic throughout the study and the short follow-up (2 years). retinopathy, the better phenotyping of this condition will An important lesson from the EUROCONDOR study permit us to identify individuals more prone to develop- is that a significant proportion (35%) of individuals ing either cardiovascular events or cognitive impairment with type 2 diabetes have early microvascular disease [48, 49]. In addition, since retinal neuron loss is related without detectable neurodysfunction [47]. Thus, neu- to deficient sensory capacity and vision-related quality rodysfunction or neurodegeneration are not always the of life, periodic assessments of neurodegeneration/neu- first abnormalities that occur in early stages of diabetic rodysfunction in the diabetic population seems warranted retinopathy. This finding underlines the need to incor- [39]. porate the assessment of NVU integrity for identifying In summar y, apar t from controlling modifiable those patients in whom neuroprotective treatment might risk factors, the early identification of patients with be of higher benefit (Fig. 1). Therefore, methods for the diabetes with NVU impairment could revolutionise assessment of neurodysfunction, such as microperimetry the current management of diabetic retinopathy (see or flicker electroretinogram (ERG) handheld recording Summary Text box). At present, a transoceanic task- device (RETeval), or methods such as OCT allowing the force funded by JDRF and The Mary Tyler Moore assessment of structural damage (e.g. neuroretinal thin- & S. Robert Levine Charitable Foundation has been ning), should be included in the screening of diabetic created to change the evaluation and grading of dia- retinopathy [40]. In addition, angio-OCT and ultra-wide betic retinopathy based on this new evidence [50]. field fundus fluorescein angiography (FFA) will improve This initiative would be the beginning of a new era in our understanding of microvascular impairment. These the diagnosis and treatment of diabetic retinopathy, examinations could be performed even when there are thus allowing better phenotyping and, consequently, no visible lesions in fundoscopic examination. It seems a more personalised and cost-effectiveness treatment. Summary Control of blood glucose levels and hypertension are the main modifiable risk factors to reduce the development and progression of diabetic retinopathy The rapid reduction of HbA1c could lead to an early worsening of diabetic retinopathy and should be avoided in those patients with advanced NPDR and PDR Obesity and visceral adipose tissue are emerging risk factors Regular physical activity, dark-meat- and oily-fish consumption, a Mediterranean diet and regular consumption of vitamin D, A, B1 and C are important lifestyle pillars to prevent or arrest diabetic retinopathy NVU impairment, which can antedate visible lesions in fundoscopic examination, merits evaluation to ensure better phenotyping of diabetic retinopathy during the screening process Treatments aimed at targeting NVU impairment in the early stages of diabetic retinopathy are a challenge to be met 13 1620 Diabetologia (2023) 66:1614–1621 Funding Open Access Funding provided by Universitat Autonoma de subjects with type 2 diabetes. Sci Rep 11:4731. https://d​ oi.o​ rg/1​ 0.​ Barcelona. 1038/​s41598-​021-​84150-8 10. Akil H, Burgess J, Nevitt S et al (2022) Early worsening of retin- Authors’ relationships and activities D-Sight, a Vall d’Hebron Research opathy in type 1 and type 2 diabetes after rapid improvement in Institute spin-off, holds intellectual property related to the use of ocu- glycaemic control: a systematic review. Diabetes Ther 13:1–23. lar DPP-4 inhibitors to treat diabetic retinopathy. The authors declare https://​doi.​org/​10.​1007/​s13300-​021-​01190-z that there are no other relationships or activities that might bias, or be 11. Bain SC, Klufas MA, Ho A, Matthews DR (2019) Worsening of perceived to bias, their work. diabetic retinopathy with rapid improvement in systemic glucose control: a review. Diabetes Obes Metab 21(3):454–466. https://​ doi.​org/​10.​1111/​dom.​13538 Contribution statement All authors were responsible for drafting the 12. UK Prospective Diabetes Study Group (1998) Tight blood pres- article and revising it critically for important intellectual content. All sure control and risk of macrovascular and microvascular compli- authors approved the version to be published. cations in type 2 diabetes: UKPDS 38. BMJ 317(7160):703–13. https://​doi.​org/​10.​1136/​bmj.​317.​7160.​703 Open Access This article is licensed under a Creative Commons Attri- 13. Do DV, Wang X, Vedula SS et al (2015) Blood pressure control for bution 4.0 International License, which permits use, sharing, adapta- diabetic retinopathy. Cochrane Database Syst Rev 1:CD006127. tion, distribution and reproduction in any medium or format, as long https://​doi.​org/​10.​1002/​14651​858.​CD006​127.​pub2 as you give appropriate credit to the original author(s) and the source, 14. Chaturvedi N, Porta M, Klein R et al (2008) Effect of candesartan provide a link to the Creative Commons licence, and indicate if changes on prevention (DIRECT-Prevent 1) and progression (DIRECT- were made. The images or other third party material in this article are Protect 1) of retinopathy in type 1 diabetes: randomised, placebo- included in the article’s Creative Commons licence, unless indicated controlled trials. Lancet 372:1394–1402. https://​doi.​org/​10.​1016/​ otherwise in a credit line to the material. If material is not included in S0140-​6736(08)​61412-9 the article’s Creative Commons licence and your intended use is not 15. Sjølie AK, Klein R, Porta M et al (2008) Effect of candesartan permitted by statutory regulation or exceeds the permitted use, you will on progression and regression of retinopathy in type 2 diabetes need to obtain permission directly from the copyright holder. To view a (DIRECT-Protect 2): a randomised placebo controlled trial. Lancet copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. 372:1385–1393. https://d​ oi.o​ rg/1​ 0.1​ 016/S ​ 0140-6​ 736(08)6​ 1411-7 16. Chung YR, Park SW, Choi SY et al (2017) Association of sta- tin use and hypertriglyceridemia with diabetic macular edema in patients with type 2 diabetes and diabetic retinopathy. Cardiovasc Diabetol 16(1):4. https://​doi.​org/​10.​1186/​s12933-​016-​0486-2 References 17. Kang EY, Chen T, Garg SJ et al (2019) Association of statin therapy with prevention of vision-threatening diabetic retinopa- 1. International Diabetes Federation (2022) Diabetes Atlas. 10th thy. JAMA Ophthalmol 137(4):363–371. https://​doi.​org/​10.​1001/​ ed. International Diabetes Federation; Brussels, Belgium. Avail- jamao​phtha​lmol.​2018.​6399 able from https://​diabe​tesat​las.​org/. Accessed 23 Dec 2022 18. Keech AC, Mitchell P, Summanen PA, FIELD study investigator 2. Simó R, Hernández C (2022) New insights into treating et al (2007) Effect of fenofibrate on the need for laser treatment early and advanced stage diabetic retinopathy. Int J Mol Sci for diabetic retinopathy (FIELD study): a randomised controlled 23(15):8513. https://​doi.​org/​10.​3390/​ijms2​31585​13 trial. Lancet 370(9600):1687–97. https://​doi.​org/​10.​1016/​S0140-​ 3. Simó R, Simó-Servat O, Bogdanov P, Hernández C (2021) Neu- 6736(07)​61607-9 rovascular unit: a new target for treating early stages of diabetic 19. ACCORD Study Group; ACCORD Eye Study Group, Chew EY retinopathy. Pharmaceutics 13(8):1320. https://​doi.​org/​10.​3390/​ et al (2010) Effects of medical therapies on retinopathy progres- pharm​aceut​ics13​081320 sion in type 2 diabetes. N Engl J Med 363(3):233–44. https://​doi.​ 4. Stratton I, Kohner E, Aldington S et al (2001) UKPDS 50: risk org/​10.​1056/​NEJMo​a1001​288 factors for incidence and progression of retinopathy in type II 20. Simó R, Hernández C (2007) Fenofibrate for diabetic retinopa- diabetes over 6 years from diagnosis. Diabetologia 44:156–163. thy. Lancet 370(9600):1667–8. https://​doi.​org/​10.​1016/​S0140-​ https://​doi.​org/​10.​1007/​s0012​50051​594 6736(07)​61608-0 5. Aiello LP, DCCT/EDIC Research Group (2014) Diabetic retinopathy 21. Zhu W, Wu Y, Meng YF, Xing Q, Tao JJ, Lu J (2018) Association and other ocular findings in the diabetes control and complications of obesity and risk of diabetic retinopathy in diabetes patients: a trial/epidemiology of diabetes interventions and complications study. meta-analysis of prospective cohort studies. Medicine (Baltimore) Diabetes Care 37(1):17–23. https://​doi.​org/​10.​2337/​dc13-​2251 97(32):e11807. https://​doi.​org/​10.​1097/​MD.​00000​00000​011807 6. Lu J, Ma X, Zhou J et al (2018) Association of time in range, as 22. Raman R, Rani PK, Gnanamoorthy P, Sudhir RR, Kumaramani- assessed by continuous glucose monitoring, with diabetic retinop- kavel G, Sharma T (2010) Association of obesity with diabetic athy in type 2 diabetes. Diabetes Care 41(11):2370–2376. https://​ retinopathy: Sankara Nethralaya Diabetic Retinopathy Epide- doi.​org/​10.​2337/​dc18-​1131 miology and Molecular Genetics Study (SN-DREAMS Report 7. Zhao Q, Zhou F, Zhang Y, Zhou X, Ying C (2019) Fasting plasma no. 8). Acta Diabetol 47(3):209–15. https://​doi.​org/​10.​1007/​ glucose variability levels and risk of adverse outcomes among s00592-​009-​0113-8 patients with type 2 diabetes: a systematic review and meta-anal- 23. Li W, Gong X, Wang W et al (2022) Association of different ysis. Diabetes Res Clin Pract 148:23–31. https://d​ oi.o​ rg/1​ 0.1​ 016/j.​ kinds of obesity with diabetic retinopathy in patients with type 2 diabr​es.​2018.​12.​010 diabetes. BMJ Open 12:e056332. https://​doi.​org/​10.​1136/​bmjop​ 8. Hietala K, Wadén J, Forsblom C, FinnDiane Study Group en-​2021-​056332 et al (2013) HbA1c variability is associated with an increased 24. Ho AW, Moul DE, Krishna J (2016) Neck circumference-height risk of retinopathy requiring laser treatment in type 1 dia- ratio as a predictor of sleep related breathing disorder in children betes. Diabetologia 56(4):737–45. https://​d oi.​o rg/​1 0.​1 007/​ and adults. J Clin Sleep Med 12(3):311–317. https://​doi.​org/​10.​ s00125-​012-​2816-6 5664/​jcsm.​5572 9. Kim HU, Park SP, Kim YK (2021) Long-term HbA1c variability and the development and progression of diabetic retinopathy in 13 Diabetologia (2023) 66:1614–1621 1621 25. Nakayama LF, Tempaku PF, Bergamo VC et al (2021) Obstruc- 38. Lee HJ, Kim CO, Lee DC (2020) Association between daily sun- tive sleep apnea and the retina: a review. J Clin Sleep Med light exposure duration and diabetic retinopathy in Korean adults 17(9):1947–1952. https://​doi.​org/​10.​5664/​jcsm.​9312 with diabetes: a nationwide population-based cross-sectional 26. Dirani M, Xie J, Fenwick E et al (2011) Are obesity and study. PLoS One 15(8):e0237149. https://​doi.​org/​10.​1371/​journ​ anthropometry risk factors for diabetic retinopathy? The al.​pone.​02371​49 diabetes management project. Invest Ophthalmol Vis Sci 39. Simó R, Stitt AW, Gardner TW (2018) Neurodegeneration in dia- 52(7):4416–21. https://​doi.​org/​10.​1167/​iovs.​11-​7208 betic retinopathy: does it really matter? Diabetologia 61(9):1902– 27. Ahlqvist E, Prasad RB, Groop L (2020) Subtypes of type 1912. https://​doi.​org/​10.​1007/​s00125-​018-​4692-1 2 diabetes determined from clinical parameters. Diabetes 40. Hernández C, Bogdanov P, Solà-Adell C et al (2017) Topical 69(10):2086–2093. https://​doi.​org/​10.​2337/​dbi20-​0001 administration of DPP-IV inhibitors prevents retinal neurodegen- 28. Sala-Vila A, Díaz-López A, Valls-Pedret C et al (2016) Dietary marine eration in experimental diabetes. Diabetologia 60:2285–2298. ω-3 fatty acids and incident sight-threatening retinopathy in middle- https://​doi.​org/​10.​1007/​s00125-​017-​4388-y aged and older individuals with type 2 diabetes: prospective investiga- 41. Ly A, Scheerer MF, Zukunft CS et al (2014) Retinal proteome tion from the PREDIMED trial. JAMA Ophthalmol 134(10):1142– alterations in a mouse model of type 2 diabetes. Diabetologia 1149. https://​doi.​org/​10.​1001/​jamao​phtha​lmol.​2016.​2906 57(1):192–203. https://​doi.​org/​10.​1007/​s00125-​013-​3070-2 29. Millen AE, Sahli MW, Nie J et al (2016) Adequate vitamin D 42. Sundstrom JM, Hernández C, Weber SR et al (2018) Prot- status is associated with the reduced odds of prevalent diabetic eomic analysis of early diabetic retinopathy reveals mediators retinopathy in African Americans and Caucasians. Cardiovasc of neurodegenerative brain diseases. Invest Ophthalmol Vis Sci Diabetol 15(1):128. https://​doi.​org/​10.​1186/​s12933-​016-​0434-1 59(6):2264–2274. https://​doi.​org/​10.​1167/​iovs.​17-​23678 30. Díaz-López A, Babio N, Martínez-González MA et al (2015) 43. Bogdanov P, Simó-Servat O, Sampedro J et al (2018) Topical Mediterranean diet, retinopathy, nephropathy, and microvascular administration of Bosentan prevents retinal neurodegeneration in diabetes complications: a post hoc analysis of a randomized trial. experimental diabetes. Int J Mol Sci 19:3578. https://​doi.​org/​10.​ Diabetes Care 38(11):2134–41. https://d​ oi.o​ rg/1​ 0.2​ 337/d​ c15-1​ 117 3390/​ijms1​91135​78 31. Scragg R, Sowers M, Bell C, Third National Health and Nutri- 44. Ohira A, Hara K, Jóhannesson G et al (2015) Topical dexa- tion Examination Survey (2004) Serum 25-hydroxyvitamin D, methasone γ-cyclodextrin nanoparticle eye drops increase vis- diabetes, and ethnicity in the Third National Health and Nutrition ual acuity and decrease macular thickness in diabetic macular Examination Survey. Diabetes Care 27(12):2813–8. https://​doi.​ oedema. Acta Ophthalmol 93(7):610–615. https://​doi.​org/​10.​ org/​10.​2337/​diaca​re.​27.​12.​2813 1111/​aos.​12803 32. Alsalem JA, Patel D, Susarla R et al (2014) Characterization of vita- 45. Semeraro F, Russo A, Gambicorti E et al (2015) Efficacy and vitre- min D production by human ocular barrier cells. Invest Ophthalmol ous levels of topical NSAIDs. Expert Opin Drug Deliv 12(11):1767– Vis Sci 55(4):2140–7. https://​doi.​org/​10.​1167/​iovs.​13-​13019 1782. https://​doi.​org/​10.​1517/​17425​247.​2015.​10687​56 33. Jamali N, Wang S, Darjatmoko SR, Sorenson CM, Sheibani N 46. Friedman SM, Almukhtar TH, Baker CW et al (2015) Topical (2017) Vitamin D receptor expression is essential during retinal nepafenec in eyes with noncentral diabetic macular edema. Retina vascular development and attenuation of neovascularization by 1, 35(5):944–956. https://​doi.​org/​10.​1097/​IAE.​00000​00000​000403 25(OH)2D3. PLoS One 12(12):e0190131. https://d​ oi.o​ rg/1​ 0.1​ 371/​ 47. Simó R, Hernández C, Porta M et al (2019) Effects of topically journ​al.​pone.​01901​31 administered neuroprotective drugs in early stages of diabetic 34. Barba I, Garcia-Ramírez M, Hernández C et al (2010) Metabolic retinopathy: results of the EUROCONDOR Clinical Trial. Dia- fingerprints of proliferative diabetic retinopathy: an 1H-NMR-based betes 68(2):457–463. https://​doi.​org/​10.​2337/​db18-​0682 metabonomic approach using vitreous humor. Invest Ophthalmol 48. Vujosevic S, Aldington SJ, Silva P et al (2020) Screening for Vis Sci 51(9):4416–21. https://​doi.​org/​10.​1167/​iovs.​10-​5348 diabetic retinopathy: new perspectives and challenges. Lancet Dia- 35. Park SW, Ghim W, Oh S et al (2019) Association of vitreous vita- betes Endocrinol 8(4):337–347. https://​doi.​org/​10.​1016/​S2213-​ min C depletion with diabetic macular ischemia in proliferative 8587(19)​30411-5 diabetic retinopathy. PLoS One 14(6):e0218433. https://​doi.​org/​ 49. Simó R, Stehouwer CDA, Avogaro A (2020) Diabetic retinopathy: 10.​1371/​journ​al.​pone.​02184​33 looking beyond the eyes. Diabetologia 63(8):1662–1664. https://​ 36. Liu DW, Haq Z, Yang D, Stewart JM (2021) Association between doi.​org/​10.​1007/​s00125-​020-​05195-4 smoking history and optical coherence tomography angiography 50. Levine SR, Sapieha P, Dutta S, Sun JK, Gardner TW (2022) It is findings in diabetic patients without diabetic retinopathy. PLoS One time for a moonshot to find “Cures” for diabetic retinal disease. 16(7):e0253928. https://​doi.​org/​10.​1371/​journ​al.​pone.​02539​28 Prog Retin Eye Res 90:101051. https://​doi.​org/​10.​1016/j.​prete​ 37. Yan X, Han X, Wu C, Shang X, Zhang L, He M (2021) Effect of yeres.​2022.​101051 physical activity on reducing the risk of diabetic retinopathy progres- sion: 10-year prospective findings from the 45 and Up Study. PLoS Publisher's note Springer Nature remains neutral with regard to One 16(1):e0239214. https://​doi.​org/​10.​1371/​journ​al.​pone.​02392​14 jurisdictional claims in published maps and institutional affiliations. 13