Associations between carotid macrovascular and retinal microvascular alterations in diabetes: Evidence synthesis and implications for risk assessment Authors: Gentian Bajraktari, Fisnik Jashari, Per Wester, Dren Boshnjaku, Edita Pllana, Ardian Sermaxhaj, Pranvera Ibrahimi Published: 2026 ABSTRACT Diabetes mellitus is characterized by progressive vascular injury affecting both the microcirculation and large arteries. Diabetic retinopathy is a major microvascular complication and has been associated in observational studies with several carotid vascular phenotypes, including increased carotid intima-media thickness, focal plaque burden, arterial stiffness, plaque vulnerability markers, and, in selected cohorts, carotid stenosis. However, these carotid endpoints differ in measurement protocol, biological meaning, and prognostic implications, and most available evidence is cross-sectional. Retinal imaging modalities, including optical coherence tomography and optical coherence tomography angiography, provide quantitative measures such as vessel density, foveal avascular-zone size, retinal nerve fiber layer thickness, and choroidal thickness that may parallel systemic vascular injury, but validated thresholds for carotid disease detection have not been established. Shared mechanisms, including chronic hyperglycemia, endothelial dysfunction, inflammation, oxidative stress, dyslipidemia, and hypoxia-driven angiogenic signaling, may contribute to both retinal and carotid abnormalities; insulin resistance is particularly relevant to type 2 diabetes and metabolic syndrome, whereas diabetes duration and glycemic exposure are central across diabetes types. This narrative review therefore summarizes the associations between retinal microvascular changes and specific carotid vascular changes, while emphasizing that many findings may be influenced by shared risk factors and that most studies show association rather than causation or proven clinical utility. Current evidence supports multidisciplinary awareness and comprehensive cardiovascular risk-factor management in people with diabetes, but it does not yet justify retinal findings or OCTA metrics as validated triggers for screening asymptomatic patients with carotid ultrasonography. Prospective studies are needed to show whether combined retinal and carotid assessment improves diagnosis, risk prediction, cost-effectiveness, and patient outcomes before it can be recommended as a routine screening strategy. Research InsightsWhat is currently known about this topic?Diabetic retinopathy and carotid vascular abnormalities frequently coexist in diabetes.Retinal and carotid abnormalities share many common risk factors.Retinal imaging offers a unique opportunity to observe microvascular damage directly, while carotid ultrasound reflects parallel macrovascular injury. What is the key research question?What is the evidence linking retinal microvascular abnormalities to specific carotid vascular endpoints in patients with diabetes? What is new?This review provides an endpoint-specific synthesis of the associations between retinal microvascular abnormalities and carotid intima-media thickness, plaque, arterial stiffness, and stenosis in diabetes.It integrates the shared metabolic, inflammatory, endothelial, oxidative, and angiogenic mechanisms that may contribute to abnormalities in both vascular beds.Combined retinal and carotid assessment is presented as a hypothesis requiring prospective validation rather than as an established strategy for vascular risk prediction or clinical decision-making. How might this study influence clinical practice?Retinal and carotid vascular findings may offer complementary perspectives on systemic vascular involvement in diabetes and may help inform comprehensive cardiovascular risk assessment; however, further prospective data are needed to clarify their combined clinical utility. Research Insights What is currently known about this topic? Diabetic retinopathy and carotid vascular abnormalities frequently coexist in diabetes. Retinal and carotid abnormalities share many common risk factors. Retinal imaging offers a unique opportunity to observe microvascular damage directly, while carotid ultrasound reflects parallel macrovascular injury. What is the key research question? What is the evidence linking retinal microvascular abnormalities to specific carotid vascular endpoints in patients with diabetes? What is new? This review provides an endpoint-specific synthesis of the associations between retinal microvascular abnormalities and carotid intima-media thickness, plaque, arterial stiffness, and stenosis in diabetes. It integrates the shared metabolic, inflammatory, endothelial, oxidative, and angiogenic mechanisms that may contribute to abnormalities in both vascular beds. Combined retinal and carotid assessment is presented as a hypothesis requiring prospective validation rather than as an established strategy for vascular risk prediction or clinical decision-making. How might this study influence clinical practice? Retinal and carotid vascular findings may offer complementary perspectives on systemic vascular involvement in diabetes and may help inform comprehensive cardiovascular risk assessment; however, further prospective data are needed to clarify their combined clinical utility. Introduction Diabetes mellitus is a chronic systemic disease characterized by persistent hyperglycemia that leads to progressive vascular injury. One of its most prevalent consequences is diabetic retinopathy, affecting nearly one-third of individuals with diabetes and remaining a leading cause of vision impairment worldwide [1]. Concurrently, involvement of large arteries, particularly the carotid arteries, represents a major macrovascular complication that substantially increases the risk of stroke and cardiovascular mortality. Historically, diabetic microvascular complications—such as retinopathy, nephropathy, and neuropathy—were considered distinct from macrovascular conditions including coronary artery disease, peripheral arterial disease, and cerebrovascular disease, with each group believed to arise through independent biological pathways [2]. Accumulating evidence now challenges this traditional dichotomy. Retinal microvascular abnormalities have been increasingly recognized as potential markers of generalized vascular injury, including early atherosclerotic changes in large arteries. Individuals with more advanced diabetic retinopathy, even in the absence of overt cardiovascular disease, frequently demonstrate subclinical atherosclerosis, manifested by increased carotid intima–media thickness and early plaque formation [3]. These findings suggest that retinal vascular alterations may reflect a broader systemic vascular burden. Given that retinal imaging is noninvasive, widely available, and routinely incorporated into diabetic care, retinal microvascular changes may provide an accessible means of identifying individuals at increased risk of systemic vascular disease. Conversely, detection of carotid atherosclerosis in patients with diabetes may prompt targeted evaluation for concomitant retinal involvement, reflecting parallel injury across vascular beds. This nonsystematic narrative overview synthesizes evidence on associations between retinal microvascular changes and carotid macrovascular phenotypes in type 1 and type 2 diabetes. The review emphasizes endpoint-specific interpretation, shared but non-specific mechanisms of vascular injury, residual confounding, and the research gaps that must be addressed before retinal or carotid imaging combinations can be used as validated screening or treatment-stratification tools. Methods This nonsystematic narrative review was conducted to summarize current evidence on the association between retinal microvascular changes and carotid macrovascular disease in individuals with diabetes mellitus. The focus was on studies evaluating retinal vascular alterations identified by fundus photography, optical coherence tomography, and optical coherence tomography angiography, as well as carotid artery changes assessed by ultrasonography, including carotid intima–media thickness, atherosclerotic plaque, and carotid stenosis. PubMed, Embase, and the Cochrane Library were searched from database inception to 18 November 2025. The core search concepts were combined using Boolean operators as follows: (diabetes OR diabetic OR "type 1 diabetes" OR "type 2 diabetes") AND ("diabetic retinopathy" OR retina* OR "retinal microvasculature" OR OCT OR OCTA OR "optical coherence tomography" OR "optical coherence tomography angiography") AND ("carotid intima-media thickness" OR CIMT OR "carotid plaque" OR "carotid atherosclerosis" OR "carotid stenosis" OR "arterial stiffness" OR "intraplaque neovascularization" OR "plaque vulnerability"). Searches were limited to human studies and English-language publications involving adult populations when these filters were available. Reference lists of selected articles and relevant reviews were screened manually. Because the article is a narrative overview, formal PRISMA flow counts, duplicate removal, duplicate independent screening, duplicate data extraction, and formal risk-of-bias scoring were not performed. Observations and comments Carotid artery changes in diabetes Diabetes affects the carotid arteries through several complementary but distinct vascular phenotypes, including increased carotid intima-media thickness, plaque formation, arterial stiffness, luminal stenosis, and plaque vulnerability. Together, these abnormalities reflect the broad macrovascular impact of diabetes, ranging from early arterial wall remodeling to clinically relevant atherosclerotic disease. Chronic hyperglycemia, insulin resistance, dyslipidemia, oxidative stress, endothelial dysfunction, inflammation, and accumulation of advanced glycation end-products contribute to these changes across different stages of vascular injury [4]. Carotid intima-media thickness (CIMT): most commonly measured by B-mode ultrasound in the common carotid artery, reflects structural remodeling of the arterial wall and is widely used as a marker of subclinical vascular disease. Increased CIMT has been associated with future cardiovascular events and remains one of the most frequently studied carotid ultrasound parameters in diabetes [5]. Patients with diabetes often show greater CIMT than non-diabetic individuals, even before clinically apparent cardiovascular disease develops. A meta-analysis showed that newly diagnosed patients with type 2 diabetes have significantly greater CIMT compared with controls, indicating that carotid wall thickening may be present early in the course of the disease [6]. Carotid plaque: represents a more focal and advanced atherosclerotic phenotype. Plaque presence, number, and total plaque burden provide clinically relevant information because plaques reflect established atherosclerotic lesion formation. In individuals with diabetes without clinically evident cardiovascular disease, carotid plaque prevalence is high, with reported estimates ranging from 44% to 72%, depending on diabetes type, disease duration, age, and population characteristics [7]. In type 2 diabetes, plaque prevalence rates of 60–72% have been reported at diagnosis or in patients without previous cardiovascular events [8]. In type 1 diabetes, plaque prevalence is generally lower but remains clinically important, ranging from 24% to 49% in cohorts without established cardiovascular or renal disease [9,10]. Higher plaque prevalence has also been observed in patients with diabetic retinopathy, supporting the concept that retinal microvascular disease and carotid atherosclerosis often develop in parallel as part of systemic diabetic vascular injury [11]. Arterial stiffness and carotid elasticity: Beyond structural atherosclerosis, diabetes is associated with functional changes in the arterial wall. Increased arterial stiffness and reduced carotid elasticity are commonly assessed using parameters such as pulse wave velocity and augmentation index [12,13]. These measurements capture impaired vascular compliance and increased pulsatile load, which may contribute to both cardiovascular risk and microvascular damage. In diabetes, arterial stiffness may arise from glycation of vascular wall proteins, endothelial dysfunction, inflammation, and medial calcification. Therefore, stiffness provides additional information about vascular aging and hemodynamic stress that complements CIMT and plaque assessment. Carotid stenosis and occlusive disease: Carotid stenosis represents the obstructive consequence of progressive atherosclerotic plaque growth. Advanced carotid stenosis or occlusion can reduce ocular perfusion and contribute to ocular ischemic syndrome, a rare but vision-threatening manifestation of carotid occlusive disease [14]. Even less severe carotid narrowing may influence retinal or choroidal circulation in selected patients with diabetes, especially in the presence of advanced retinopathy or other vascular risk factors [15,16]. Thus, carotid stenosis is clinically important not only as a marker of stroke risk but also as a potential contributor to ocular hypoperfusion. Plaque vulnerability: refers to biological features that increase the likelihood of plaque progression, rupture, embolization, or clinical events. These include intraplaque neovascularization, inflammation, lipid-rich necrotic core, intraplaque hemorrhage, and surface disruption. Diabetes may promote plaque vulnerability through endothelial dysfunction, oxidative stress, inflammatory activation, and pro-angiogenic signaling within the plaque microenvironment. In this context, intraplaque neovascularization is particularly relevant because it reflects active plaque remodeling and may represent a bridge between metabolic inflammation, angiogenesis, and plaque instability. Overall, diabetes promotes a spectrum of carotid vascular abnormalities. CIMT reflects early arterial wall remodeling, plaque burden reflects focal atherosclerotic lesion formation, arterial stiffness reflects impaired vascular compliance, stenosis reflects obstructive disease, and plaque vulnerability reflects biological instability. Considering these carotid phenotypes separately allows a more precise understanding of how diabetic macrovascular disease relates to retinal microvascular injury and supports a broader cross-vascular view of diabetic vascular complications. Progressive atherosclerosis can lead to significant carotid artery stenosis or even occlusion. Severe carotid stenosis (70–99%) drastically reduces ocular blood perfusion and can precipitate the ocular ischemic syndrome (OIS) – a rare but vision-threatening condition (discussed in a later section) [14]. Even moderate subclinical stenosis in diabetics may have consequences for ocular circulation, subtly affecting retinal perfusion or choroidal blood flow [15,16]. Retinal microvascular changes in diabetes The diabetic retina undergoes progressive microangiopathic changes driven by chronic hyperglycemia, hypertension, and other metabolic insults [17,18]. Non-proliferative diabetic retinopathy (NPDR) is characterized by microaneurysms, intraretinal hemorrhages, hard exudates, and cotton wool spots, reflecting capillary leakage and occlusion [19]. As NPDR progresses from mild to severe stages, cumulative microvascular damage occurs. Proliferative diabetic retinopathy (PDR) represents the advanced stage, defined by retinal neovascularization secondary to ischemia [20]. If untreated, PDR may lead to vitreous hemorrhage and tractional retinal detachment [21]. Beyond these classic lesions, diabetes induces neuroretinal thinning and microcirculatory alterations detectable by modern imaging. Optical coherence tomography (OCT) reveals thinning of the retinal nerve fiber layer (RNFL) and ganglion cell layer, even in patients without advanced retinopathy—likely due to neurodegeneration and inner retinal ischemia. Optical coherence tomography angiography (OCTA), a noninvasive modality, maps retinal capillary networks and demonstrates reduced vessel density and enlargement of the foveal avascular zone (FAZ) in diabetic eyes, correlating with DR severity [22]. Capillary dropout in the macula is evident even in NPDR and worsens with disease progression. Notably, studies in patients with carotid artery stenosis or diabetes have reported associations between carotid disease and lower retinal vessel density or larger FAZ, suggesting that OCTA-derived microvascular metrics may reflect systemic vascular insufficiency in selected cohorts [23]. Collectively, hallmark retinal changes in diabetes include capillary nonperfusion, increased vascular permeability, ischemia-driven neovascularization, and neural tissue loss—features that can be quantitatively monitored using fundus photography, OCT, and OCTA. Correlation between carotid and retinal changes Epidemiological and cross-sectional evidence A substantial body of observational evidence supports an association between diabetic retinopathy (DR) and several carotid vascular endpoints. Most studies are cross-sectional and compare carotid ultrasound markers in patients with and without DR. Reported associations vary by carotid phenotype and include increased CIMT, higher plaque burden, arterial stiffness, and, in selected cohorts, carotid stenosis [24]. For example, Liao et al. reported a mean difference in CCA-IMT of 0.059 mm (95% CI: 0.026-0.091, p < 0.001) between patients with and without microvascular complications, including retinopathy [25]. Similarly, Drinkwater et al. found that patients with DR had higher CIMT and lower retinal vessel density on OCTA than those without DR [26]. Several studies report adjusted associations between DR and carotid endpoints, although cross-sectional findings should be interpreted as associations rather than evidence that DR independently causes or predicts carotid disease [27]. Saif et al. reported that NPDR and PDR were associated with increased CIMT in type 2 diabetes after adjustment for age, HbA1c, and blood pressure [24]. In some studies, more severe DR has been associated with thicker CIMT or greater plaque burden, but the strength and consistency of these findings differ across carotid endpoints [25]. A systematic review of 16 studies reported that all but one found an association between at least one carotid ultrasound parameter and DR presence or grade [28]. An overview of key clinical studies examining the association between retinal microvascular abnormalities and carotid macrovascular disease in diabetes is summarized in Table 1.Table 1Evidence linking diabetic retinal findings with carotid vascular endpoints.Table 1Study (Author, Years)Population/study designCarotid parameter(s)Retinal parameter(s)Main correlation finding/Adjusted effect estimateCovariates included in modelMajor limitations/overlap/interpretationRema et al., 2004 [29]Urban Indian cohort, T2D (N = 590)Cross-sectional population-based urban Indian T2D cohort; diabetes duration about 4 ± 5 years.Cross-sectional.Right carotid ultrasound. CIMT (common carotid IMT); arterial stiffness was measured by assessing the augmentation index.DR presence/severity (any DR) retinal photographs with ETDRS-based DR severity assessment.DR was associated with greater mean CIMT and arterial stiffness; severity gradient reported. This is cross-sectional association, not longitudinal prediction.Age, HbA1c, diabetes duration and microalbuminuria.Right carotid only; no temporal inference; limited covariate adjustment; plaque endpoint not assessed.Distiller et al., 2006 [30]Long-duration T1D (N = 148, South Africa)Cross-sectional clinic cohort of long-duration T1D; median diabetes duration 26 years (range 18-59). Cross-sectional.CIMT (common carotid); carotid plaqueCCA-CIMT measured >1 cm proximal to flow divider; plaque defined as focal thickness ≥ 1.5 mm.DR presence/severity (many long-term T1Ds had PDR) any DR assessed by ophthalmologist; detailed grading system not fully reportedCCA-CIMT was not significantly associated with DR after adjustment; carotid plaque was associated with DR.Carotid plaque: adjusted OR 3.65 (95% CI 1.11-12.02), p = 0.033. CCA-CIMT/IMT-risk associations were not significant or exact adjusted CI not reported.Plaque model: age, hypertension, smoking. IMT models also included diabetes duration, BMI, HDL and/or TG:HDL ratio, HbA1c.Highly selected long-duration T1D sample; normal renal function/no proteinuria; limited generalizability; cross-sectional.Lacroix et al., 2006 [31]Asymptomatic T2D patients (N = 300, France)Cross-sectional vascular-screening cohort; asymptomatic T2D; mean diabetes duration 11 ± 9.2 years. Cross-sectional.Carotid ultrasound; stenosis categorized by velocity criteria, including clinically relevant stenosis thresholds.DR presence (any retinopathy) any DR assessed by ophthalmologist; detailed grading system not fully reported.DR was associated with carotid stenosis, including more severe stenosis, after adjustment. Any stenosis: adjusted OR 2.38 (95% CI 1.06-5.33). Stenosis ≥60%: adjusted OR 3.62 (95% CI 1.12-11.73).age >70, hypertension, BMI, CHD history, family history of diabetes. Severe stenosis: sex, ABI, ischemic neurological disorder/cervical bruit.Referral/screening selection bias; not a general T2D cohort; cross-sectional; carotid endpoint limited to stenosis.Malecki et al., 2008 [32]Clinic-based T2D (N = 182, Poland)Cross-sectional clinic-based T2D study; diabetes duration and full selection criteria were not available in the accessible extraction.CIMT; endothelial functionCIMT and endothelial function reported; exact ultrasound protocol not reported in the accessible extraction.DR presence (any DR) any DR assessed by ophthalmologist; detailed grading system not fully reportedIMT was higher in patients with DR than those without DR(0·87 mm vs. 0·79 mm, respectively, P = 0·0001. In multiple regression analysis, DR was among the predictors of increased IMT (P = 0.016)Age, HbA1c, diabetes duration, hypertension, smokingIncomplete nephropathy data limited adjustment in the multivariable model; proteinuria severity was not reported.Araszkiewicz et al., 2011 [33]Adult T1D patients (N = 87, Poland)Cross-sectional hospital T1D cohort; median diabetes duration 10 years [9-14].CIMT; central augmentation index (AI) right CCA-CIMT by high-resolution ultrasound/automatic software; wave-reflection indices also assessed.DR presence (45% with DR; mostly NPDR)Two-field retinal photographs assessed by ophthalmologist; AAO classification.Although unadjusted CIMT differed by DR status, adjusted CCA-CIMT association with DR was not significant.CCA-CIMT and DR: adjusted OR 1.00 (95% CI 0.99-1.01), p = 0.169 per 1-μm increase.Age, sex, diabetes duration, albuminuria, blood pressure, postprandial glucose, HbA1c, central and peripheral augmentation indices.Small sample; right carotid only; exclusion of higher-risk carotid/CVD cases; cross-sectional.Son et al., 2011 [34]Newly diagnosed T2D (N = 142, Korea)Cross-sectional study in newly diagnosed T2D; duration = newly diagnosed; excluded known CVD/cerebrovascular disease. Cross-sectional.CIMT (common carotid) bilateral CCA-CIMT 5-10 mm proximal to bulb; subclinical atherosclerosis defined as CIMT >0.9 mm and/or plaque.DR presence (any DR at diagnosis): any DR at diagnosis from two-field retinal photographs assessed by ophthalmologist.DR was associated with a composite carotid endpoint at diagnosis; this remains cross-sectional, not longitudinal prediction.Composite CIMT >0.9 mm and/or plaque: adjusted OR 6.57 (95% CI 1.68-25.71), p = 0.007.Age, sex, smoking, hypertension, BMI, diabetic nephropathy, HbA1c, fasting glucose, HDL and LDL.Small newly diagnosed sample; wide CI; CVD exclusions; temporal direction cannot be established.Yun et al., 2011 [35]Community T2D cohort (N = 605, Korea)Cross-sectional community/public-health T2D cohort; mean diabetes duration 8.7 ± 7.7 years. Cross-sectional.CIMT; carotid plaque; arterial stiffness (baPWV)CCA-CIMT tertiles, carotid plaque and baPWV; structural carotid endpoints are interpreted separately from arterial stiffness.DR presence retinal imaging graded with ETDRS criteria.Positive association was mainly for arterial stiffness (baPWV). CIMT tertiles and plaque were null after adjustment.CCA-CIMT T2 vs T1: OR 1.16 (95% CI 0.67-2.02); T3 vs T1: OR 1.06 (0.59-1.90), p = 0.844. Plaque: OR 1.20 (0.75-1.91).Age, sex, diabetes duration, HbA1c, total cholesterol, triglycerides, HDL, eGFR, BMI and hypertension history.Null structural carotid findings are shown alongside positive arterial-stiffness findingsCardoso et al., 2012 [36]Rio de Janeiro T2D Cohort (N = 441, Brazil)Cross-sectional baseline analysis of Rio de Janeiro T2D cohort; mean diabetes duration 10.2 ± 8.6 years.CIMT (CCA, bifurcation, ICA); carotid plaquesCIMT at CCA, bifurcation and ICA; plaque score by Mannheim criteria; single vascular radiologist.DR status (any DR vs none at baseline) any DR assessed by ophthalmologist.Plaque burden was associated with DR; adjusted IMT associations were null or not significant.Plaque score >2: adjusted OR 1.70 (95% CI 1.02-2.84), p = 0.043. CCA/BIF/ICA IMT: adjusted associations not significant or CI NR.Plaque score: age, sex, smoking, antihypertensive use, aortic PWV. IMT models included age, nighttime pulse pressure plus endpoint-specific variables.Tertiary high-risk cohort; cross-sectional. Potential cohort overlap with Cardoso et al., 2019; do not count as independent cohorts.Jung et al., 2013 [37]Clinic T2D patients (N = 131, S. Korea)Cross-sectional hospital T2D study; mean diabetes duration 7.4 ± 6.8 years.Carotid atherosclerosis (plaque via ultrasound)CCA-CIMT measured 1 cm proximal to bulb bilaterally; endpoints CCA-CIMT ≥ 1 mm and >2 plaques.DR presence (any DR); any DR assessed by ophthalmologist.DR was associated with elevated CCA-CIMT, while the plaque estimate was very imprecise and not significant.CCA-CIMT ≥ 1 mm: adjusted OR 3.8 (95% CI 1.4-10.2). >2 plaques: OR 5.7 (95% CI 0.6-51.3), not significant.Age, diabetes duration, smoking, hypertension, HbA1c, cardiac autonomic neuropathy, baPWV, statin use, ACE-I/ARB use and eGFR.Retrospective design; small sample; wide CI for plaque endpoint; CVD exclusions; cross-sectional.Alonso et al., 2015 [38]Clinic T2D patients (N = 312, Spain)Cross-sectional clinic/eye-disease program T2D cohort; diabetes duration no DR 6 [2.5-10] vs DR 11 [6-20] years.Carotid plaque count; total plaque area; CIMTSemiautomated CIMT at ICA, bifurcation and CCA; plaque count and total plaque area by Mannheim criteria.DR status (no DR vs any DR) multifield stereoscopic retinal photographs assessed by ophthalmologist; DR severity assessed.Association strongest for plaque burden; CCA/BIF CIMT and DR-severity analyses included null findings.Any plaque: adjusted OR 1.71 (95% CI 1.03-2.85), p = 0.0366. ≥ 2 plaques: OR 3.17 (95% CI 1.75-5.75), p < 0.0001. Mean ICA-CIMT p = 0.0176; CI NRAge and endpoint-specific covariates: sex, smoking, hypertension and/or dyslipidemia depending on outcome.Clinic selection; CVD/renal exclusions; endpoint-specific adjustment makes comparisons difficult; cross-sectional.Liu et al., 2015 [39]Community T2D screening (N = 1,607, China)Cross-sectional community diabetes-management program; diabetes duration DR 8 [3-13] vs no DR 6 [2-10] years.CIMT (common carotid)CCA-CIMT >1 mm, plaque and composite subclinical atherosclerosis; max left/right 1.5 cm proximal to bifurcation.DR presence (any DR) retinal images graded with International Clinical Diabetic Retinopathy severity scale.DR was associated with thickened CCA-CIMT, plaque and composite subclinical carotid atherosclerosis after adjustment.CCA-CIMT >1 mm: OR 1.84 (95% CI 1.02-3.31). Plaque: OR 1.87 (1.03-3.39). Composite endpoint: OR 1.93 (1.03-3.60).Age, sex, alcohol, LDL; models also smoking, hypertension, duration, HbA1c, diabetes medications/insulin, antihypertensives, obesity, lipids, eGFR, uACR and GGT.Known CVD excluded; single sonographer/ophthalmologist; residual confounding; cross-sectional.Carbonell et al., 2018 [40]Adult T1D cohort (N = 340, Spain)Cross-sectional adult T1D outpatient cohort; diabetes duration 20 [14-29] years.CIMT; carotid plaques (presence and count)Carotid CIMT and plaque presence/count; plaque by Mannheim criteria.DR severity (none, mild, moderate, severe)DR categorized as none, mild NPDR and advanced DR using ETDRS-based criteria.Advanced DR, rather than mild DR alone, was most clearly associated with plaque burden; ocular blood-flow findings were not part of this study.Any plaque: mild DR OR 1.35 (95% CI 0.66-2.76), null; advanced DR OR 2.66 (1.03-6.95). ≥ 2 plaques: advanced DR OR 4.71 (1.48-15.04). Any DR overall for plaque: OR 1.64 (0.85-3.17), null.Age, sex, diabetes duration, smoking, diastolic BP, dyslipidemia, uACR, BMI, pulse pressure and LDL.Advanced DR subgroup small; CVD/renal exclusions; cross-sectional. Distinguish mild/null findings from advanced-positive findings.Ichinohasama et al., 2019 [15].Hospital T2D patients with no or mild DR (N = 129, Japan)Cross-sectional hospital T2D cohort limited to no or mild DR; duration no DR 8.0 ± 6.2 vs mild NPDR 12.5 ± 7.8 years.CIMT; ocular blood flow (laser speckle flowgraphy) maximum right CCA-CIMT; ocular blood flow by laser speckle flowgraphy.DR grade (mild NPDR vs no DR) no DR vs mild NPDR; ETDRS criteria assessed by ophthalmologist.Association was limited to mild NPDR vs no DR and was cross-sectional; ocular ischemia was also assessed.Mild NPDR vs no DR: adjusted OR 8.65 (95% CI 1.95-38.40), p = 0.005 per 1-mm higher CCA-CIMT.Age, sex, diabetes duration, HbA1c, diastolic BP, heart rate, creatinine, central macular thickness and optic-nerve-head mean blur rate.Only no/mild DR; right-eye/right-carotid measurements; wide CI; recruitment/exclusions limit generalizability.Cardoso et al., 2019 [41]T2D outpatients, Brazil (N = 478), median follow-up ∼10.8 yearsProspective Rio de Janeiro T2D cohort; median diabetes duration 8 [3-15] years; median follow-up 10.8 years.CIMT (CCA, CB, ICA); carotid plaques; plaque score baseline CIMT at CCA, bifurcation and ICA; carotid plaque and plaque score; Mannheim criteria; reliability reported.Incident DR (development or progression over follow-up) annual ophthalmologist review; incident/worsening DR defined as new DR or ≥ 2-grade progression.In longitudinal adjusted analyses, baseline CIMT/plaque did not significantly predict incident or worsening DR.Highest vs lowest tertile HRs: CCA-CIMT 0.99 (95% CI 0.59-1.64); BIF-CIMT 1.23 (0.73-2.06); ICA-CIMT 1.17 (0.71-1.93). Plaque score ≥ 3: HR 1.69 (0.88-3.24).Age, sex, diabetes duration, BMI, smoking, physical activity, clinic SBP, antihypertensives, insulin, statins, macrovascular disease, baseline DR, mean HbA1c, HDL and LDL.Only longitudinal study among older studies and adjusted findings were null. Tertiary high-risk cohort; possible overlap with Cardoso et al., 2012 baseline cohort.Yang et al., 2023 [42]Mixed T1 and T2 DM cohort with OCTA and cardio/vascular ultrasound (N ~ 307)Cross-sectional diagnostic/modeling study with mixed diabetes types; diabetes duration and selection details were not reported in the accessible extraction.Carotid ultrasound (plaque of CCA among macrovascular measures) carotid plaque/atherosclerosis included among macrovascular ultrasound measures; exact definition was not available in the accessible extraction.OCTA metrics (3×3 & 6×6): FAZ area; vessel density; blood flow density (central/inner/whole regions)OCTA 3 x 3 and 6 x 6 metrics including FAZ, vessel density and blood-flow density.IRBFD improved DR model AUC from 0.836 to 0.887; CRBFD improved AUC from 0.836 to 0.851; DPVD improved carotid-atherosclerosis model AUC from 0.781 to 0.802; FAZ 3 x 3 in T2D improved AUC from 0.766 to 0.797. No conventional adjusted OR/HR with 95% CI reported.Model predictors/covariates NR in accessible abstract; not reported in the accessible extraction.Not labeled as an independent predictor because model coefficients/95% CIs were not reported. Diagnostic modeling is interpreted separately from etiologic association and longitudinal prediction.Castelblanco et al., 2025 [43]Adults with T1D from Catalonia, Spain (N = 242)Cross-sectional/post-hoc T1D study; N = 242, plaque n = 67 and no plaque n = 175; derived from a previous Catalan T1D study.Subclinical carotid plaque presence by ultrasound carotid ultrasound for subclinical plaque presence; plaque size/stenosis not the primary detailed endpoint.OCT-derived RNFL, GCL/IPL; EDI-OCT choroidal thickness (sub-foveal, nasal, temporal)OCT RNFL, GCL/IPL and EDI-OCT choroidal thickness; DR included as predictor in ML models.RNFL and GCL/IPL were not different by plaque status; subfoveal/temporal choroidal thickness was lower with plaque. ML feature importance is not an adjusted OR/HR.No conventional adjusted OR/HR with 95% CI for DR/plaque reported. ML models: accuracy 82-85%, specificity 92-98%, balanced accuracy 0.60 left eye and 0.71 right eye.ML features included age, hypertension, dyslipidemia, smoking, DR and ophthalmological OCT measures; not a standard multivariable OR/HR model.Small plaque-positive subgroup; modest sensitivity/balanced accuracy; hypertension may confound choroidal findings; potential overlap with Carbonell/Catalan T1D cohorts [40].Abbreviations: T1D, type 1 diabetes; T2D, type 2 diabetes; CIMT, carotid intima–media thickness; IMT, intima–media thickness; DR, diabetic retinopathy; PDR, proliferative diabetic retinopathy; CCA, common carotid artery; NPDR, nonproliferative diabetic retinopathy; baPWV, brachial–ankle pulse wave velocity; ICA, internal carotid artery; CB, carotid bifurcation; OCTA, optical coherence tomography angiography; FAZ, foveal avascular zone; DPVD, deep plexus vessel density; RNFL, retinal nerve fiber layer; GCL, ganglion cell layer; IPL, inner plexiform layer; EDI-OCT, enhanced depth imaging optical coherence tomography. Table 1Evidence linking diabetic retinal findings with carotid vascular endpoints.Table 1Study (Author, Years)Population/study designCarotid parameter(s)Retinal parameter(s)Main correlation finding/Adjusted effect estimateCovariates included in modelMajor limitations/overlap/interpretationRema et al., 2004 [29]Urban Indian cohort, T2D (N = 590)Cross-sectional population-based urban Indian T2D cohort; diabetes duration about 4 ± 5 years.Cross-sectional.Right carotid ultrasound. CIMT (common carotid IMT); arterial stiffness was measured by assessing the augmentation index.DR presence/severity (any DR) retinal photographs with ETDRS-based DR severity assessment.DR was associated with greater mean CIMT and arterial stiffness; severity gradient reported. This is cross-sectional association, not longitudinal prediction.Age, HbA1c, diabetes duration and microalbuminuria.Right carotid only; no temporal inference; limited covariate adjustment; plaque endpoint not assessed.Distiller et al., 2006 [30]Long-duration T1D (N = 148, South Africa)Cross-sectional clinic cohort of long-duration T1D; median diabetes duration 26 years (range 18-59). Cross-sectional.CIMT (common carotid); carotid plaqueCCA-CIMT measured >1 cm proximal to flow divider; plaque defined as focal thickness ≥ 1.5 mm.DR presence/severity (many long-term T1Ds had PDR) any DR assessed by ophthalmologist; detailed grading system not fully reportedCCA-CIMT was not significantly associated with DR after adjustment; carotid plaque was associated with DR.Carotid plaque: adjusted OR 3.65 (95% CI 1.11-12.02), p = 0.033. CCA-CIMT/IMT-risk associations were not significant or exact adjusted CI not reported.Plaque model: age, hypertension, smoking. IMT models also included diabetes duration, BMI, HDL and/or TG:HDL ratio, HbA1c.Highly selected long-duration T1D sample; normal renal function/no proteinuria; limited generalizability; cross-sectional.Lacroix et al., 2006 [31]Asymptomatic T2D patients (N = 300, France)Cross-sectional vascular-screening cohort; asymptomatic T2D; mean diabetes duration 11 ± 9.2 years. Cross-sectional.Carotid ultrasound; stenosis categorized by velocity criteria, including clinically relevant stenosis thresholds.DR presence (any retinopathy) any DR assessed by ophthalmologist; detailed grading system not fully reported.DR was associated with carotid stenosis, including more severe stenosis, after adjustment. Any stenosis: adjusted OR 2.38 (95% CI 1.06-5.33). Stenosis ≥60%: adjusted OR 3.62 (95% CI 1.12-11.73).age >70, hypertension, BMI, CHD history, family history of diabetes. Severe stenosis: sex, ABI, ischemic neurological disorder/cervical bruit.Referral/screening selection bias; not a general T2D cohort; cross-sectional; carotid endpoint limited to stenosis.Malecki et al., 2008 [32]Clinic-based T2D (N = 182, Poland)Cross-sectional clinic-based T2D study; diabetes duration and full selection criteria were not available in the accessible extraction.CIMT; endothelial functionCIMT and endothelial function reported; exact ultrasound protocol not reported in the accessible extraction.DR presence (any DR) any DR assessed by ophthalmologist; detailed grading system not fully reportedIMT was higher in patients with DR than those without DR(0·87 mm vs. 0·79 mm, respectively, P = 0·0001. In multiple regression analysis, DR was among the predictors of increased IMT (P = 0.016)Age, HbA1c, diabetes duration, hypertension, smokingIncomplete nephropathy data limited adjustment in the multivariable model; proteinuria severity was not reported.Araszkiewicz et al., 2011 [33]Adult T1D patients (N = 87, Poland)Cross-sectional hospital T1D cohort; median diabetes duration 10 years [9-14].CIMT; central augmentation index (AI) right CCA-CIMT by high-resolution ultrasound/automatic software; wave-reflection indices also assessed.DR presence (45% with DR; mostly NPDR)Two-field retinal photographs assessed by ophthalmologist; AAO classification.Although unadjusted CIMT differed by DR status, adjusted CCA-CIMT association with DR was not significant.CCA-CIMT and DR: adjusted OR 1.00 (95% CI 0.99-1.01), p = 0.169 per 1-μm increase.Age, sex, diabetes duration, albuminuria, blood pressure, postprandial glucose, HbA1c, central and peripheral augmentation indices.Small sample; right carotid only; exclusion of higher-risk carotid/CVD cases; cross-sectional.Son et al., 2011 [34]Newly diagnosed T2D (N = 142, Korea)Cross-sectional study in newly diagnosed T2D; duration = newly diagnosed; excluded known CVD/cerebrovascular disease. Cross-sectional.CIMT (common carotid) bilateral CCA-CIMT 5-10 mm proximal to bulb; subclinical atherosclerosis defined as CIMT >0.9 mm and/or plaque.DR presence (any DR at diagnosis): any DR at diagnosis from two-field retinal photographs assessed by ophthalmologist.DR was associated with a composite carotid endpoint at diagnosis; this remains cross-sectional, not longitudinal prediction.Composite CIMT >0.9 mm and/or plaque: adjusted OR 6.57 (95% CI 1.68-25.71), p = 0.007.Age, sex, smoking, hypertension, BMI, diabetic nephropathy, HbA1c, fasting glucose, HDL and LDL.Small newly diagnosed sample; wide CI; CVD exclusions; temporal direction cannot be established.Yun et al., 2011 [35]Community T2D cohort (N = 605, Korea)Cross-sectional community/public-health T2D cohort; mean diabetes duration 8.7 ± 7.7 years. Cross-sectional.CIMT; carotid plaque; arterial stiffness (baPWV)CCA-CIMT tertiles, carotid plaque and baPWV; structural carotid endpoints are interpreted separately from arterial stiffness.DR presence retinal imaging graded with ETDRS criteria.Positive association was mainly for arterial stiffness (baPWV). CIMT tertiles and plaque were null after adjustment.CCA-CIMT T2 vs T1: OR 1.16 (95% CI 0.67-2.02); T3 vs T1: OR 1.06 (0.59-1.90), p = 0.844. Plaque: OR 1.20 (0.75-1.91).Age, sex, diabetes duration, HbA1c, total cholesterol, triglycerides, HDL, eGFR, BMI and hypertension history.Null structural carotid findings are shown alongside positive arterial-stiffness findingsCardoso et al., 2012 [36]Rio de Janeiro T2D Cohort (N = 441, Brazil)Cross-sectional baseline analysis of Rio de Janeiro T2D cohort; mean diabetes duration 10.2 ± 8.6 years.CIMT (CCA, bifurcation, ICA); carotid plaquesCIMT at CCA, bifurcation and ICA; plaque score by Mannheim criteria; single vascular radiologist.DR status (any DR vs none at baseline) any DR assessed by ophthalmologist.Plaque burden was associated with DR; adjusted IMT associations were null or not significant.Plaque score >2: adjusted OR 1.70 (95% CI 1.02-2.84), p = 0.043. CCA/BIF/ICA IMT: adjusted associations not significant or CI NR.Plaque score: age, sex, smoking, antihypertensive use, aortic PWV. IMT models included age, nighttime pulse pressure plus endpoint-specific variables.Tertiary high-risk cohort; cross-sectional. Potential cohort overlap with Cardoso et al., 2019; do not count as independent cohorts.Jung et al., 2013 [37]Clinic T2D patients (N = 131, S. Korea)Cross-sectional hospital T2D study; mean diabetes duration 7.4 ± 6.8 years.Carotid atherosclerosis (plaque via ultrasound)CCA-CIMT measured 1 cm proximal to bulb bilaterally; endpoints CCA-CIMT ≥ 1 mm and >2 plaques.DR presence (any DR); any DR assessed by ophthalmologist.DR was associated with elevated CCA-CIMT, while the plaque estimate was very imprecise and not significant.CCA-CIMT ≥ 1 mm: adjusted OR 3.8 (95% CI 1.4-10.2). >2 plaques: OR 5.7 (95% CI 0.6-51.3), not significant.Age, diabetes duration, smoking, hypertension, HbA1c, cardiac autonomic neuropathy, baPWV, statin use, ACE-I/ARB use and eGFR.Retrospective design; small sample; wide CI for plaque endpoint; CVD exclusions; cross-sectional.Alonso et al., 2015 [38]Clinic T2D patients (N = 312, Spain)Cross-sectional clinic/eye-disease program T2D cohort; diabetes duration no DR 6 [2.5-10] vs DR 11 [6-20] years.Carotid plaque count; total plaque area; CIMTSemiautomated CIMT at ICA, bifurcation and CCA; plaque count and total plaque area by Mannheim criteria.DR status (no DR vs any DR) multifield stereoscopic retinal photographs assessed by ophthalmologist; DR severity assessed.Association strongest for plaque burden; CCA/BIF CIMT and DR-severity analyses included null findings.Any plaque: adjusted OR 1.71 (95% CI 1.03-2.85), p = 0.0366. ≥ 2 plaques: OR 3.17 (95% CI 1.75-5.75), p < 0.0001. Mean ICA-CIMT p = 0.0176; CI NRAge and endpoint-specific covariates: sex, smoking, hypertension and/or dyslipidemia depending on outcome.Clinic selection; CVD/renal exclusions; endpoint-specific adjustment makes comparisons difficult; cross-sectional.Liu et al., 2015 [39]Community T2D screening (N = 1,607, China)Cross-sectional community diabetes-management program; diabetes duration DR 8 [3-13] vs no DR 6 [2-10] years.CIMT (common carotid)CCA-CIMT >1 mm, plaque and composite subclinical atherosclerosis; max left/right 1.5 cm proximal to bifurcation.DR presence (any DR) retinal images graded with International Clinical Diabetic Retinopathy severity scale.DR was associated with thickened CCA-CIMT, plaque and composite subclinical carotid atherosclerosis after adjustment.CCA-CIMT >1 mm: OR 1.84 (95% CI 1.02-3.31). Plaque: OR 1.87 (1.03-3.39). Composite endpoint: OR 1.93 (1.03-3.60).Age, sex, alcohol, LDL; models also smoking, hypertension, duration, HbA1c, diabetes medications/insulin, antihypertensives, obesity, lipids, eGFR, uACR and GGT.Known CVD excluded; single sonographer/ophthalmologist; residual confounding; cross-sectional.Carbonell et al., 2018 [40]Adult T1D cohort (N = 340, Spain)Cross-sectional adult T1D outpatient cohort; diabetes duration 20 [14-29] years.CIMT; carotid plaques (presence and count)Carotid CIMT and plaque presence/count; plaque by Mannheim criteria.DR severity (none, mild, moderate, severe)DR categorized as none, mild NPDR and advanced DR using ETDRS-based criteria.Advanced DR, rather than mild DR alone, was most clearly associated with plaque burden; ocular blood-flow findings were not part of this study.Any plaque: mild DR OR 1.35 (95% CI 0.66-2.76), null; advanced DR OR 2.66 (1.03-6.95). ≥ 2 plaques: advanced DR OR 4.71 (1.48-15.04). Any DR overall for plaque: OR 1.64 (0.85-3.17), null.Age, sex, diabetes duration, smoking, diastolic BP, dyslipidemia, uACR, BMI, pulse pressure and LDL.Advanced DR subgroup small; CVD/renal exclusions; cross-sectional. Distinguish mild/null findings from advanced-positive findings.Ichinohasama et al., 2019 [15].Hospital T2D patients with no or mild DR (N = 129, Japan)Cross-sectional hospital T2D cohort limited to no or mild DR; duration no DR 8.0 ± 6.2 vs mild NPDR 12.5 ± 7.8 years.CIMT; ocular blood flow (laser speckle flowgraphy) maximum right CCA-CIMT; ocular blood flow by laser speckle flowgraphy.DR grade (mild NPDR vs no DR) no DR vs mild NPDR; ETDRS criteria assessed by ophthalmologist.Association was limited to mild NPDR vs no DR and was cross-sectional; ocular ischemia was also assessed.Mild NPDR vs no DR: adjusted OR 8.65 (95% CI 1.95-38.40), p = 0.005 per 1-mm higher CCA-CIMT.Age, sex, diabetes duration, HbA1c, diastolic BP, heart rate, creatinine, central macular thickness and optic-nerve-head mean blur rate.Only no/mild DR; right-eye/right-carotid measurements; wide CI; recruitment/exclusions limit generalizability.Cardoso et al., 2019 [41]T2D outpatients, Brazil (N = 478), median follow-up ∼10.8 yearsProspective Rio de Janeiro T2D cohort; median diabetes duration 8 [3-15] years; median follow-up 10.8 years.CIMT (CCA, CB, ICA); carotid plaques; plaque score baseline CIMT at CCA, bifurcation and ICA; carotid plaque and plaque score; Mannheim criteria; reliability reported.Incident DR (development or progression over follow-up) annual ophthalmologist review; incident/worsening DR defined as new DR or ≥ 2-grade progression.In longitudinal adjusted analyses, baseline CIMT/plaque did not significantly predict incident or worsening DR.Highest vs lowest tertile HRs: CCA-CIMT 0.99 (95% CI 0.59-1.64); BIF-CIMT 1.23 (0.73-2.06); ICA-CIMT 1.17 (0.71-1.93). Plaque score ≥ 3: HR 1.69 (0.88-3.24).Age, sex, diabetes duration, BMI, smoking, physical activity, clinic SBP, antihypertensives, insulin, statins, macrovascular disease, baseline DR, mean HbA1c, HDL and LDL.Only longitudinal study among older studies and adjusted findings were null. Tertiary high-risk cohort; possible overlap with Cardoso et al., 2012 baseline cohort.Yang et al., 2023 [42]Mixed T1 and T2 DM cohort with OCTA and cardio/vascular ultrasound (N ~ 307)Cross-sectional diagnostic/modeling study with mixed diabetes types; diabetes duration and selection details were not reported in the accessible extraction.Carotid ultrasound (plaque of CCA among macrovascular measures) carotid plaque/atherosclerosis included among macrovascular ultrasound measures; exact definition was not available in the accessible extraction.OCTA metrics (3×3 & 6×6): FAZ area; vessel density; blood flow density (central/inner/whole regions)OCTA 3 x 3 and 6 x 6 metrics including FAZ, vessel density and blood-flow density.IRBFD improved DR model AUC from 0.836 to 0.887; CRBFD improved AUC from 0.836 to 0.851; DPVD improved carotid-atherosclerosis model AUC from 0.781 to 0.802; FAZ 3 x 3 in T2D improved AUC from 0.766 to 0.797. No conventional adjusted OR/HR with 95% CI reported.Model predictors/covariates NR in accessible abstract; not reported in the accessible extraction.Not labeled as an independent predictor because model coefficients/95% CIs were not reported. Diagnostic modeling is interpreted separately from etiologic association and longitudinal prediction.Castelblanco et al., 2025 [43]Adults with T1D from Catalonia, Spain (N = 242)Cross-sectional/post-hoc T1D study; N = 242, plaque n = 67 and no plaque n = 175; derived from a previous Catalan T1D study.Subclinical carotid plaque presence by ultrasound carotid ultrasound for subclinical plaque presence; plaque size/stenosis not the primary detailed endpoint.OCT-derived RNFL, GCL/IPL; EDI-OCT choroidal thickness (sub-foveal, nasal, temporal)OCT RNFL, GCL/IPL and EDI-OCT choroidal thickness; DR included as predictor in ML models.RNFL and GCL/IPL were not different by plaque status; subfoveal/temporal choroidal thickness was lower with plaque. ML feature importance is not an adjusted OR/HR.No conventional adjusted OR/HR with 95% CI for DR/plaque reported. ML models: accuracy 82-85%, specificity 92-98%, balanced accuracy 0.60 left eye and 0.71 right eye.ML features included age, hypertension, dyslipidemia, smoking, DR and ophthalmological OCT measures; not a standard multivariable OR/HR model.Small plaque-positive subgroup; modest sensitivity/balanced accuracy; hypertension may confound choroidal findings; potential overlap with Carbonell/Catalan T1D cohorts [40].Abbreviations: T1D, type 1 diabetes; T2D, type 2 diabetes; CIMT, carotid intima–media thickness; IMT, intima–media thickness; DR, diabetic retinopathy; PDR, proliferative diabetic retinopathy; CCA, common carotid artery; NPDR, nonproliferative diabetic retinopathy; baPWV, brachial–ankle pulse wave velocity; ICA, internal carotid artery; CB, carotid bifurcation; OCTA, optical coherence tomography angiography; FAZ, foveal avascular zone; DPVD, deep plexus vessel density; RNFL, retinal nerve fiber layer; GCL, ganglion cell layer; IPL, inner plexiform layer; EDI-OCT, enhanced depth imaging optical coherence tomography. Evidence linking diabetic retinal findings with carotid vascular endpoints. Abbreviations: T1D, type 1 diabetes; T2D, type 2 diabetes; CIMT, carotid intima–media thickness; IMT, intima–media thickness; DR, diabetic retinopathy; PDR, proliferative diabetic retinopathy; CCA, common carotid artery; NPDR, nonproliferative diabetic retinopathy; baPWV, brachial–ankle pulse wave velocity; ICA, internal carotid artery; CB, carotid bifurcation; OCTA, optical coherence tomography angiography; FAZ, foveal avascular zone; DPVD, deep plexus vessel density; RNFL, retinal nerve fiber layer; GCL, ganglion cell layer; IPL, inner plexiform layer; EDI-OCT, enhanced depth imaging optical coherence tomography. Type 2 diabetes In type 2 diabetes (T2D), often accompanied by metabolic syndrome, the link between DR and macrovascular disease is well established [44,45]. T2D patients with DR are more likely to have carotid plaques and higher CIMT than those without DR. Alonso et al. reported that in T2D patients without overt cardiovascular or renal disease, DR was independently associated with a higher carotid plaque burden [38]. Diabetics with DR had increased odds (>3) of having ≥2 carotid plaques and higher internal carotid IMT, underscoring retinopathy as a marker of generalized atherosclerosis [38,46]. Clinically, the presence of DR signals elevated risk for silent carotid and coronary disease [47]. Longitudinal data from cohorts such as ACCORD confirm that retinopathy predicts cardiovascular events; in ACCORD, severe DR was associated with higher incidence of myocardial infarction and stroke during follow-up [48]. Type 1 diabetes In type 1 diabetes (T1D), microvascular complications typically occur after longer disease duration, often without traditional cardiovascular risk factors [49]. Nevertheless, similar correlations exist. Carbonell et al. found that T1D patients with advanced DR had markedly higher prevalence of carotid plaques and greater CIMT than those without DR [40]. Even adolescents with T1D can exhibit early retinal microvascular changes that are associated with increased carotid intima-media thickness, supporting a parallel progression of micro- and macrovascular disease in this population [50,51]. Proposed mechanisms include shared pathophysiology—chronic inflammation and endothelial dysfunction— alongside shared risk factors [52]. By adulthood, T1D patients with proliferative DR show higher rates of coronary calcification and stroke, supporting the “common soil” hypothesis linking microangiopathy and macroangiopathy [53,54]. Advanced retinal imaging correlates Beyond clinical DR grading, quantitative retinal imaging has been evaluated in relation to carotid disease. Studies in patients with carotid artery stenosis or diabetes have reported associations between carotid disease and lower retinal vessel density, enlarged FAZ, and thinner RNFL on OCT/OCTA [26,55]. Thinning of the retinal nerve fiber layer (RNFL), including in the temporal quadrant, is associated with carotid artery stenosis and may reflect chronic retinal ischemia or optic nerve hypoperfusion, as demonstrated by significant correlations between RNFL thinning and extracranial carotid stenosis in large community-based and meta-analytic studies [56,57]. Another study reported that carotid atherosclerosis, defined by increased IMT or plaques, was associated with lower OCTA capillary density in T2D [58]. However, no validated OCTA threshold currently exists for identifying carotid stenosis or determining the need for carotid ultrasonography. These findings highlight the need for further studies to determine whether specific ocular imaging markers or OCTA thresholds can identify patients who may benefit from carotid ultrasonography or other macrovascular arterial screening. Fig. 1 illustrates representative carotid ultrasound, fundus photography, and optical coherence tomography angiography findings across different stages of diabetic vascular involvement.Fig. 1Representative Carotid Ultrasound, Fundus Photography, and Optical Coherence Tomography Angiography Findings in Diabetes(A) Normal carotid arterial wall on ultrasound with normal fundus appearance and preserved retinal capillary network on optical coherence tomography angiography.(B) Increased carotid intima–media thickness associated with nonproliferative diabetic retinopathy changes.(C) Echolucent internal carotid artery plaque with hemodynamically significant stenosis, characterized by elevated peak systolic and end-diastolic velocities, associated with proliferative diabetic retinopathy.Note: The carotid ultrasound, fundus photography, and OCT/OCTA images shown in this figure were obtained from anonymized patients during routine clinical practice by the authors' clinical group. Panels A–C represent different illustrative clinical cases and do not originate from the same patient. None of the images were reproduced or adapted from previously published sources. Representative Carotid Ultrasound, Fundus Photography, and Optical Coherence Tomography Angiography Findings in Diabetes (A) Normal carotid arterial wall on ultrasound with normal fundus appearance and preserved retinal capillary network on optical coherence tomography angiography. (B) Increased carotid intima–media thickness associated with nonproliferative diabetic retinopathy changes. (C) Echolucent internal carotid artery plaque with hemodynamically significant stenosis, characterized by elevated peak systolic and end-diastolic velocities, associated with proliferative diabetic retinopathy. Note: The carotid ultrasound, fundus photography, and OCT/OCTA images shown in this figure were obtained from anonymized patients during routine clinical practice by the authors' clinical group. Panels A–C represent different illustrative clinical cases and do not originate from the same patient. None of the images were reproduced or adapted from previously published sources. Metabolic and vascular mechanisms linking retinal and carotid injury in diabetes Insulin resistance (IR), a hallmark of type 2 diabetes and metabolic syndrome, represents an important pathophysiological contributor to both microvascular and macrovascular complications. However, the mechanisms linking retinal microvascular injury and carotid macrovascular disease differ partly between type 1 and type 2 diabetes. In both diabetes types, chronic hyperglycemia, longer diabetes duration, oxidative stress, endothelial dysfunction, inflammation, and advanced glycation contribute to vascular injury. In type 2 diabetes, these processes are further amplified by insulin resistance, obesity, dyslipidemia, and metabolic syndrome [59,60]. The “common soil” hypothesis proposes that shared metabolic and inflammatory disturbances create a vascular environment that favors both atherosclerosis and microangiopathy, including diabetic retinopathy (DR) [59,60]. In type 2 diabetes, IR may contribute to the parallel development of carotid atherosclerosis and DR by promoting endothelial dysfunction, chronic low-grade inflammation, dyslipidemia, and metabolic abnormalities across different vascular beds [61,62]. In type 1 diabetes, the link between retinal and carotid disease is more likely driven by cumulative glycemic exposure, diabetes duration, endothelial dysfunction, renal disease, inflammation, and other cardiovascular risk factors rather than insulin resistance alone. Clinically, interventions that improve metabolic health and reduce vascular risk—such as lifestyle modification, blood pressure control, lipid-lowering therapy, glycemic management, and selected glucose-lowering agents—may reduce both microvascular and macrovascular complications [63,64]. However, these benefits should not be attributed only to improvement in insulin sensitivity. The Steno-2 study showed that intensive, target-driven multifactorial intervention, including lifestyle modification and pharmacologic therapy for hyperglycemia, hypertension, dyslipidemia, and microalbuminuria, led to a 53% reduction in major cardiovascular events and a 58% reduction in retinopathy risk over 8 years compared with conventional care [65]. These findings support early and comprehensive vascular risk-factor management rather than a single-mechanism approach focused only on IR [66]. One important component of this vascular injury is hyperglycemia-induced mitochondrial oxidative stress. Fig. 2 illustrates how glucose overload increases mitochondrial electron transport chain activity and reactive oxygen species generation, contributing to endothelial dysfunction and metabolic injury. The broader metabolic, endothelial, inflammatory, oxidative, and angiogenic pathways that may contribute to parallel retinal and carotid vascular abnormalities are summarized in Fig. 3.Fig. 2Hyperglycemia-induced mitochondrial reactive oxygen species production in diabetes.Hyperglycemia increases mitochondrial electron transport chain flux, leading to excess superoxide generation and hydrogen peroxide formation, which contribute to insulin resistance and endothelial dysfunction.Abbreviations: ATP, adenosine triphosphate; TCA, tricarboxylic acid cycle; FADH, flavin adenine dinucleotide (reduced form); FAD, flavin adenine dinucleotide; NADH, nicotinamide adenine dinucleotide (reduced form); NAD+, nicotinamide adenine dinucleotide (oxidized form); SOD, superoxide dismutase; H2O2, hydrogen peroxide; O2−, superoxide anion; H+, proton; ADP, adenosine diphosphate; SMC, smooth muscle cells.Fig. 3Pathophysiologic Pathways Linking Retinal Microvascular and Carotid Macrovascular Disease in Diabetes.Chronic hyperglycemia promotes endothelial dysfunction, inflammation, oxidative stress, and impaired nitric oxide bioavailability, which may contribute in parallel to retinal microvascular injury and carotid atherosclerosis. Hypoxia-driven vascular endothelial growth factor upregulation may contribute to neovascularization in proliferative diabetic retinopathy and intraplaque angiogenesis in carotid plaques.Abbreviations: ATP, adenosine triphosphate; ROS, reactive oxygen species; NO, nitric oxide; RNFL, retinal nerve fiber layer; GCL, ganglion cell layer; OCTA, optical coherence tomography angiography; VD, vessel density; SCP, superficial capillary plexus; DCP, deep capillary plexus; FAZ, foveal avascular zone; PWV, pulse wave velocity; VEGF, vascular endothelial growth factor; TIA, transient ischemic attack. Hyperglycemia-induced mitochondrial reactive oxygen species production in diabetes. Hyperglycemia increases mitochondrial electron transport chain flux, leading to excess superoxide generation and hydrogen peroxide formation, which contribute to insulin resistance and endothelial dysfunction. Abbreviations: ATP, adenosine triphosphate; TCA, tricarboxylic acid cycle; FADH, flavin adenine dinucleotide (reduced form); FAD, flavin adenine dinucleotide; NADH, nicotinamide adenine dinucleotide (reduced form); NAD+, nicotinamide adenine dinucleotide (oxidized form); SOD, superoxide dismutase; H2O2, hydrogen peroxide; O2−, superoxide anion; H+, proton; ADP, adenosine diphosphate; SMC, smooth muscle cells. Pathophysiologic Pathways Linking Retinal Microvascular and Carotid Macrovascular Disease in Diabetes. Chronic hyperglycemia promotes endothelial dysfunction, inflammation, oxidative stress, and impaired nitric oxide bioavailability, which may contribute in parallel to retinal microvascular injury and carotid atherosclerosis. Hypoxia-driven vascular endothelial growth factor upregulation may contribute to neovascularization in proliferative diabetic retinopathy and intraplaque angiogenesis in carotid plaques. Abbreviations: ATP, adenosine triphosphate; ROS, reactive oxygen species; NO, nitric oxide; RNFL, retinal nerve fiber layer; GCL, ganglion cell layer; OCTA, optical coherence tomography angiography; VD, vessel density; SCP, superficial capillary plexus; DCP, deep capillary plexus; FAZ, foveal avascular zone; PWV, pulse wave velocity; VEGF, vascular endothelial growth factor; TIA, transient ischemic attack. Endothelial Dysfunction: Under normal conditions, insulin signaling through the PI3K/Akt pathway promotes nitric oxide (NO) synthesis, which maintains vasodilation and has anti-atherogenic effects. In type 2 diabetes and insulin-resistant states, impaired PI3K/Akt signaling combined with compensatory MAPK pathway activation reduces NO bioavailability and increases oxidative stress [67]. In both type 1 and type 2 diabetes, chronic hyperglycemia and oxidative stress also contribute to endothelial dysfunction. This imbalance promotes vasoconstriction, platelet activation, inflammation, and impaired vascular repair. In large arteries, these processes contribute to arterial stiffness, plaque formation, and atherosclerotic progression, while in retinal capillaries they promote increased permeability, impaired autoregulation, and blood-retina barrier disruption [68,69]. These shared endothelial abnormalities provide one plausible biological link between carotid and retinal vascular disease, although they do not prove a causal relationship between the two vascular beds [70,71]. Pro-Inflammatory State: Chronic low-grade inflammation is another important mechanism connecting diabetic microvascular and macrovascular injury. In type 2 diabetes, insulin resistance and adipose tissue dysfunction increase the release of free fatty acids and pro-inflammatory cytokines such as TNF-α and IL-6, which activate endothelial cells and macrophages [72,73]. In type 1 diabetes, inflammation may be more closely related to chronic hyperglycemia, immune activation, oxidative stress, renal disease, and long disease duration. In the retina, inflammatory mediators promote leukostasis, capillary degeneration, and progression of DR. In large arteries, similar inflammatory pathways contribute to plaque formation, vascular remodeling, and plaque vulnerability. Systemic inflammatory markers such as C-reactive protein (CRP) have been associated with both DR severity and carotid intima-media thickness, although these associations may be influenced by shared cardiovascular risk factors [[74], [75], [76]]. Dyslipidemia and Advanced Glycation: Atherogenic dyslipidemia is particularly characteristic of type 2 diabetes and insulin-resistant states, with elevated triglycerides, increased small dense LDL particles, and reduced HDL cholesterol [77,78]. These lipid abnormalities accelerate atherosclerosis and may also contribute to microvascular dysfunction. In both type 1 and type 2 diabetes, chronic hyperglycemia promotes the formation of advanced glycation end-products (AGEs), which modify lipoproteins, increase oxidative stress, and alter extracellular matrix structure. AGEs can damage both large arteries and retinal microvasculature by promoting endothelial dysfunction, vascular stiffness, inflammation, and disruption of vascular integrity [79]. Hyperglycemia and Metabolic Memory: Chronic hyperglycemia remains a central driver of diabetic retinopathy and contributes to macrovascular disease through several mechanisms, including polyol pathway activation, protein kinase C signaling, oxidative stress, and AGE-RAGE interactions [80]. The concept of “metabolic memory” suggests that earlier periods of poor glycemic control may have lasting effects on retinal and arterial tissues, even after later improvement in glycemia. This is particularly relevant in type 1 diabetes and long-duration diabetes, where cumulative glycemic exposure plays a major role. Long-term follow-up data from DCCT/EDIC and UKPDS support the importance of early intensive glycemic control, showing durable reductions in microvascular and macrovascular complications, including diabetic retinopathy and cardiovascular events, even after later convergence of glycemic control between treatment groups [81,82]. Thus, hyperglycemia, diabetes duration, and metabolic memory should be considered core mechanisms linking retinal and carotid vascular injury, while insulin resistance should be viewed mainly as an additional contributor in type 2 diabetes. Neovascularization as a hypothesis of disease progression Pathologic neovascularization (NV) may represent a conceptual link between microvascular disease in the eye and plaque vulnerability in the carotid artery, as both processes can be driven by hypoxia, VEGF signaling, inflammation, and vascular remodeling. Fig. 3 summarizes the shared angiogenic mechanisms contributing to proliferative diabetic retinopathy and carotid plaque instability. In the diabetic retina, chronic capillary nonperfusion induces tissue hypoxia, activating hypoxia-inducible factors such as HIF-1α, which upregulate vascular endothelial growth factor (VEGF) and other pro-angiogenic mediators. This cascade culminates in proliferative diabetic retinopathy (PDR), characterized by fragile, leaky new vessels on the optic disc and elsewhere, which predispose to vitreous hemorrhage and tractional retinal detachment. The central role of VEGF in this process is clinically validated by the success of intraocular anti-VEGF agents in preventing or regressing retinal NV and its complications [[83], [84], [85]]. A parallel angiogenic phenomenon occurs within the atherosclerotic carotid artery. As plaques enlarge beyond the diffusion limit, hypoxia develops in the intima and media, stimulating expansion of the vasa vasorum and intraplaque neovascularization (IPN) from the adventitia inward [86]. These microvessels are immature and highly permeable, facilitating inflammatory cell infiltration and predisposing to intraplaque hemorrhage, which accelerates plaque growth and destabilization, increasing the risk of cerebrovascular events [87]. Imaging and histopathology studies—including contrast-enhanced ultrasound, MRI, and microvessel density analysis—consistently demonstrate that exuberant IPN correlates with high-risk plaque features and clinical vulnerability [88]. Diabetes amplifies plaque neovascularization through multiple mechanisms. Beyond promoting dyslipidemia and endothelial dysfunction, the pro-inflammatory and pro-angiogenic milieu characteristic of diabetes independently associates with IPN in the carotid bed [89]. Clinical cohorts using contrast-enhanced ultrasound confirm diabetes as an independent risk factor for IPN, even after adjusting for age, blood pressure, and lipid levels, supporting the concept that systemic drivers of retinal ischemia and VEGF upregulation also favor angiogenic remodeling within carotid plaques [90]. At the interface of macro- and microcirculation, severe carotid inflow limitation can further exacerbate ocular NV. In ocular ischemic syndrome (OIS), typically caused by high-grade ipsilateral carotid stenosis, chronic hypoperfusion triggers anterior-segment NV (rubeosis iridis) and posterior-segment NV, often accompanied by mid-peripheral hemorrhages and, in advanced cases, neovascular glaucoma [91]. This represents a clinically striking example of carotid pathology driving ocular NV on top of any pre-existing DR. Recognition is critical because OIS carries a high risk of stroke and cardiovascular mortality, necessitating urgent carotid evaluation alongside ocular therapy. These shared angiogenic pathways have practical implications for risk stratification and management. Retinal NV or disproportionate nonperfusion on OCTA may serve as a marker of systemic vascular burden, while sonographic evidence of carotid IPN may identify vulnerable plaque biology, particularly in patients with diabetes. Anti-VEGF therapy effectively treats retinal NV locally, whereas IPN currently serves as an imaging biomarker rather than a therapeutic target, refining cerebrovascular risk assessment beyond the degree of stenosis [92]. Furthermore, studies using OCTA have reported improvements in retinal microcirculatory parameters, including increased macular and peripapillary flow density and reduced FAZ size, after carotid endarterectomy or stenting. These findings suggest that carotid perfusion may influence retinal vascular health, although further prospective studies are needed to confirm the strength and clinical relevance of this relationship [93,94]. In summary, neovascularization may represent a biologically plausible conceptual link between diabetic retinal microangiopathy and carotid plaque vulnerability. Both retinal NV and intraplaque NV may involve hypoxia-driven angiogenic signaling, which can be amplified by the metabolic and inflammatory environment of diabetes [[95], [96], [97], [98]]. This conceptual link may improve understanding of parallel ocular and vascular changes in diabetes and supports further investigation of coordinated ocular and systemic vascular assessment. Clinical implications and preventive strategies The observed association between carotid artery changes and retinal microvascular alterations in diabetes supports the concept that vascular injury may occur across multiple vascular beds. Because much of the available evidence is cross-sectional, these findings should be interpreted as evidence of coexistence and shared vascular risk rather than proof that retinal abnormalities can reliably identify carotid disease. Assessment of one vascular bed may provide useful clinical context regarding the other, but current evidence does not establish retinal abnormalities as indications for carotid imaging. The presence of diabetic retinopathy, especially when moderate to severe or progressive, should prompt comprehensive cardiovascular risk assessment and optimization of established preventive measures, including blood pressure control, lipid-lowering therapy, glycemic management, smoking cessation, and lifestyle modification, in accordance with contemporary diabetes guidelines [99,100]. At present, however, diabetic retinopathy alone should not be considered sufficient evidence to recommend routine carotid or coronary screening in asymptomatic patients. Similarly, detection of carotid atherosclerosis, such as carotid plaque or stenosis, may indicate a higher systemic vascular burden and should encourage clinicians to ensure that ophthalmologic follow-up for diabetic retinopathy is up to date. Emerging evidence raises the hypothesis that combined retinal and carotid assessment could add information to vascular risk stratification in diabetes, but this requires prospective validation and current data are insufficient to support universal integrated screening. In current practice, carotid imaging should be guided by established clinical indications independently of the presence or severity of diabetic retinopathy. Noninvasive carotid imaging is appropriate in patients with recent focal neurological or retinal ischemic symptoms, retinal embolic events, or suspected ocular ischemic syndrome [101,102]. By contrast, diabetic retinopathy severity, marked asymmetry of retinopathy, severe retinal ischemia, or a high burden of conventional cardiovascular risk factors alone should not currently be regarded as validated indications for carotid imaging. Whether these features can identify selected patients who derive incremental benefit from additional vascular assessment remains a hypothesis requiring prospective evaluation. Therapeutic and coordinated management implications The shared mechanisms underlying diabetic retinopathy and carotid atherosclerosis suggest that certain therapies may confer cross-benefits. For example, blockade of the renin–angiotensin system (RAS) with ACE inhibitors or ARBs not only reduces cardiovascular events but has also been associated with slowing of retinopathy progression, possibly through blood pressure–independent effects on microvascular integrity. Meta-analyses and large randomized trials demonstrate that RAS inhibitors significantly reduce the risk of diabetic retinopathy progression and increase the likelihood of disease regression [103]. These benefits are not solely attributable to blood pressure lowering; mechanistic studies and clinical trials indicate that RAS blockade improves retinal microvascular integrity through inhibition of VEGF expression, reduction of oxidative stress, and enhancement of retinal perfusion [104]. Similarly, glucagon-like peptide-1 receptor agonists (GLP-1 RAs) and sodium-glucose cotransporter-2 inhibitors (SGLT2is) both reduce major adverse cardiovascular events and all-cause mortality in type 2 diabetes, with SGLT2is providing additional benefit for heart failure and kidney outcomes [99]. SGLT2is may confer retinal protection, especially with longer-term use, while short-term use may transiently increase risk in some populations [105]. Preliminary data from GLP-1 RA studies hint at potential benefits in slowing DR progression, although one retrospective clinical cohort study reported a transient risk of worsening, underscoring the need for careful monitoring [106]. Despite ongoing mechanistic uncertainty, the transient worsening is not thought to result from direct retinal toxicity but rather from rapid glycemic improvement, supporting a pragmatic focus on therapies that strengthen systemic vascular health in diabetic retinopathy [107]. Some studies have reported that coexistence of proliferative diabetic retinopathy and significant carotid artery stenosis confers a high risk of stroke or transient ischemic attack [108,109]. Randomized evidence indicates that fenofibrate can slow diabetic retinopathy progression in selected patients. ACCORD-Eye reported benefit from intensive glycemic control and fenofibrate-based lipid therapy, but not from intensive blood-pressure control [110,111]. A large, population-based cohort study of 219,359 patients found that statin use was associated with a lower incidence of diabetic retinopathy and fewer vision-threatening complications, including proliferative diabetic retinopathy, vitreous hemorrhage, tractional retinal detachment, and diabetic macular edema. Statin therapy was also linked to reduced need for retinal interventions such as laser photocoagulation, intravitreal injections, and vitrectomy [112]. At present, coexistence of OCTA abnormalities and CIMT or plaque does not by itself establish a need to intensify therapy beyond guideline-directed risk management. Future Directions for integrated imaging and risk stratification Future research should move beyond correlation. Priority areas include prospective cohorts using standardized retinal and carotid imaging protocols, repeated assessments to clarify temporal relationships, external validation of OCT/OCTA thresholds, and prediction models evaluating their incremental value beyond established cardiovascular risk factors. Decision-curve analysis, impact studies, randomized screening trials, cost-effectiveness analyses, and patient-centered outcomes will be required before combined retinal-carotid imaging can be implemented as an evidence-based clinical pathway. Future studies should also evaluate whether artificial intelligence applied to retinal imaging can improve assessment of diabetes-related systemic vascular complications [113,114]. Retinal hemodynamic relationships in carotid disease [115], imaging responses to PCSK9 inhibition [116], and shared inflammatory pathways and targeted anti-inflammatory strategies [117,118] warrant further investigation. Conclusions Carotid macrovascular abnormalities and retinal microvascular injury are associated in diabetes, but the current evidence base is heterogeneous and largely observational. Diabetic retinopathy, particularly in more advanced stages, has been linked with higher CIMT, greater plaque burden, arterial stiffness, and in selected cohorts carotid stenosis. Shared mechanisms include chronic hyperglycemia, endothelial dysfunction, inflammation, oxidative stress, dyslipidemia, and hypoxia-driven angiogenic signaling, while insulin resistance is most relevant to type 2 diabetes and metabolic syndrome. The association between retinal and carotid findings has clinical value as a reminder to optimize vascular risk factors and coordinate ophthalmologic and cardiovascular care, but retinal abnormalities and OCTA metrics are not yet validated triggers for asymptomatic carotid screening. Retinal and carotid vascular findings may offer complementary perspectives on systemic vascular involvement in diabetes and may help inform comprehensive cardiovascular risk assessment; however, further prospective data are needed to clarify their combined clinical utility. Author contributions Gentian Bajraktari: Conceptualization and Writing – original draft. Fisnik Jashari: Conceptualization and Writing – original draft. Pranvera Ibrahimi: Conceptualization, writing – review and editing. Dren Boshnjaku: Writing – original draft. Ardian Sermaxhaj: Validation. Edita Pllana: Writing – review and editing. Per Wester: Supervision. Ethics approval and consent to participate Not applicable. Sources of funding The authors declare that there are no sources of funding. Declaration of interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.