Associations between Retinal and Choroidal Vascularization Parameters and Brachial Artery Flow-mediated Dilation in Type 2 Diabetes and Healthy Controls

1. INTRODUCTION

Type 2 Diabetes Mellitus (DM) has a significant impact on the eye, as diabetic retinopathy is one of the leading causes of preventable vision loss [1]. One of the primary tissues affected by diabetes is the vascular endothelium, a layer of specialized cells lining the interior surface of blood vessels. This tissue plays a crucial role in various functions, including hemostasis, inflammation modulation, vaso- constriction, and vasodilation [2]. It is suggested that damage to the vascular endothelium is critical in the development of diabetic retinopathy and that such damage also affects the choroid [3, 4].

Optical Coherence Tomography Angiography (OCTA) is a non-invasive examination that enables the visualization of the superficial and deep retinal capillary plexus vascular densities (SCP and DCP, respectively). These vessels consist of a single layer of endothelial cells, along with pericytes and a basement membrane [5]. Visualizing the choroidal vasculature is more challenging, although it may be assessed using the Choroidal Vascularity Index (CVI), as proposed by Agrawal et al. [6].

Vascular endothelial function is also impaired in cardiovascular disease. Flow-mediated vasodilation (FMD) is widely used to assess endothelial function in peripheral arteries [7]. This phenomenon describes the vasodilation of an artery in response to increased blood flow, such as after a period of ischemia. Studies have reported that FMD of the brachial artery is an indicator of cardio- vascular risk [8].

One of the main factors contributing to FMD is the release of endothelium-derived nitric oxide (NO). It has been shown that NO modulates basal flow through the retinal and choroidal vasculature and influences the response of other vasodilatory factors, such as bradykinin, acetylcholine, and histamine [9]. Ocular hemodynamics is impaired by endothelial dysfunction because it increases the synthesis of reactive oxygen species and decreases the absorption of NO [10]. FMD has previously been studied in regard to the retinal and choroidal diseases. It has been found to be decreased in advanced diabetic retinopathy and wet age-related macular degeneration [11, 12]. These findings suggest that peripheral endothelial function measured by FMD may influence the ocular vascular system.

Identifying ophthalmic equivalents of FMD could facilitate the detection of patients with potentially increased cardiovascular risk and enhance our under- standing of the pathophysiology of diabetic eye disease. This is particularly important because the diabetic population is already at an elevated risk for cardiovascular disease, which is further exacerbated by the presence of diabetic retinopathy [13]. To the best of our knowledge, it has not yet been determined whether vascular endothelial function, as expressed by FMD, translates into retinal and choroidal vascular status in patients with type 2 diabetes mellitus.

The aim of our study is to identify which ocular vascularisation parameters (SCP, DCP, or CVI) may be related to vascular endothelial function in the peripheral circulation, expressed by FMD, in type 2 diabetic patients and healthy population.

2. METHODS

This observational study included patients with type 2 diabetes without diabetic retinopathy and healthy controls. Exclusion criteria included type 1 diabetes, age under 18 years, intraocular pressure over 21 mmHg, previous ocular treatments, retinal diseases, ocular diseases that may occlude the visual axis (corneal opacities, cataract), systemic diseases other than diabetes and hypertension (HT). Data on HbA1C levels and hypertension history were collected.

Comprehensive ophthalmological examinations were conducted for each patient, which involved slit-lamp examination of the anterior segment and fundus, best-corrected visual acuity assessment using a Snellen chart (maximum limit of 1.0), axial length measurement (ARGOS biometer, Alcon, Switzerland), and air-puff intraocular pressure measurement (TRK-2p, Topcon, Japan). Spectral-domain optical coherence tomography (SD-OCT) and angio-OCT scans of the macula (3 × 3 mm) were taken using the RTVue XR Avanti (Optovue Inc., Fremont, CA, USA). SCP and DCP were calculated using built-in software. Only the eye with the better image quality in the angio-OCT scans was selected for analysis in each participant. All measurements were performed under consistent room lighting conditions in the evening.

The study was conducted in accordance with the guidelines of the Declaration of Helsinki and approved by the Institutional Ethics Committee of Wroclaw Medical University, Poland (KB-510/2022). Informed consent was obtained from all subjects involved in the study.

3. RESULTS

Thirty-two patients participated in the study; 15 were in the study group, and 17 were in the control group. The groups did not differ significantly in terms of sex and age (p=0.49 and p= 0.52, respectively) or other clinical characteristics (Table 1). In the diabetes group, the mean diabetes duration was 6.63 years, and the mean HbA1C level was 7.6%. There was no statistical difference in FMD between the DM and control group, despite the DM group having higher mean FMD (6.64 vs 5.67, p= 0.47, respectively). This lack of significance may be attributed to the higher standard deviation in the DM group (3.69 vs 2.67, respectively). Of the retinal and choroidal vascu- larization parameters, CVI-C and perifoveal CVI-P were significantly reduced in the study group compared to the control group (32.56 vs. 47.17 and 53.08 vs. 59.34, respectively) (Table 2 and Fig. 1).

FMD correlated with clinical characteristics, showing a negative correlation with age in the general population and the control group (r=-0.43, p= 0.01; r=-0.51, p=0.04, respectively). In the diabetic group, FMD was also negatively correlated to HT (r=-0.57, p=0.03). When examining the correlation between FMD and vascu- larisation parameters, a positive correlation was found with the SCP-P and CVI-P in the control group (r=0.57, p=0.02; r=0.58, p=0.01, respectively). Additionally, a positive correlation between FMD and DCP-P was observed in the study group, the control group, and the general population (r=0.58, p=0.04; r=0.86, p<0.001; r=0.68, p<0.001, respectively) (Table 3).

In the univariate regression analysis, age, SCP-P, and DCP-P were found to significantly influence FMD (b=-0.38, b=0.36, and b=0.57, respectively). Parameters with a p-value lower than 0.1, plus age and sex, were included in the multivariate analysis. Due to the collinearity of DCP-P and SCP-P, it was decided to omit the latter from the multivariate analysis. In the final model, FMD was found to be dependent on the presence of HT (b=-0.4) and DCP-P (b=0.47) (Table 4).

4. DISCUSSION

In this study, in diabetic patients and healthy population, FMD was found to be significantly positively correlated to DCP. Our findings also indicated that FMD was influenced negatively by age and hypertension. Additionally, after adjusting for age and sex, FMD decreased with the presence of hypertension and lower perifoveal DCP. In our studied population, only CVI was significantly lower in diabetic patients than in healthy controls.

In 2023, Heiss et al. published FMD reference values for healthy individuals. Their meta-analysis reported a mean FMD of 6.4%, which was similar to the 6.13% obtained in our study population, containing DM patients and healthy controls. Age significantly influenced the reduction in FMD, which was consistent with our findings. A high heterogeneity was observed between studies despite similar methodologies being used [14]. The association of FMD with hypertension has been previously reported, as HT impairs vascular endothelial function, which influences the degree of arterial vasodilation [15, 16]. Importantly, we did not observe a significant difference in FMD values between the diabetic and healthy groups. It may be explained by the fact that patients in the DM group did not have diabetic complications, and their mean HbA1C level was not significantly elevated (7.6%), so the impairment of endothelial function may not have been significant. It is also possible that this relationship would change in a larger study population. FMD is increasingly recognized for its value not only as a measure of endothelial function but also as a potential predictor of cardiovascular events. Numerous studies have demons- trated that impaired FMD is associated with an elevated risk of atherosclerosis and coronary artery disease, suggesting its utility as a non-invasive tool for early detection of cardiovascular dysfunction [17-19]. The ability to track improvements in FMD following lifestyle or pharmacological interventions makes it a useful metric in both clinical and research settings. However, while FMD has predictive value, it is important to acknowledge that variability in measurement and patient-specific factors can affect its reliability as a universal predictor of cardiovascular outcomes [20-22].

CVI was significantly reduced in the DM group compared to the control group, both in the central and perifoveal regions. This finding aligned with previous studies reporting reduced CVI in patients with diabetes, even in the absence of visible changes in the eye fundus [23, 24]. We did not observe a difference in SCP and DCP, likely due to the small sample sizes, as previous studies have reported reductions in these parameters in diabetic patients [25, 26].

FMD was positively correlated with DCP-P in both study groups and with SCP-P and CVI-P, but only in the control group. This is the first study to observe these correlations, as the relationship among FMD and SCP, DCP, and CVI, to the best of our knowledge, has not been studied in a similar population. Additionally, in the regression analysis, DCP-P emerged as a sex and age-independent factor positively affecting FMD. It is worth mentioning that SCP-P also positively influenced FMD in the univariate analysis and exhibited collinearity with DCP-P. The higher vessel density in the retinal plexuses of patients with higher FMD indicated that endothelial status in the large circulation may be translated to micro- circulation. The relationship between FMD and densities of retinal plexuses has been previously studied in post-COVID-19 patients, but no correlation among macular SCP, DCP, and FMD has been found [27].

Diabetes results in impaired vascular endothelial function, especially in patients with long-term disease. It may be caused by a reduction in nitric oxide (NO) production [28]. NO is a crucial molecule for vasodilation, and its deficiency leads to a decrease in FMD. Lower NO concentration can cause the constriction of arterioles and a subsequent reduction in capillary density. The observed correlation among FMD, SCP, and CVI exclusively in the healthy cohort indicates that endothelial function is dysregulated in individuals with diabetes. The association of FMD with DCP may persist in the diabetic population as this vascular bed is less susceptible to ischemic changes than SCP [29].

Vascular endothelial status, as expressed by FMD, has been studied in various ophthalmic conditions. Bojic et al. demonstrated lower brachial FMD in patients with glaucoma compared to a healthy control group [30]. Yao et al. observed reduced FMD in patients with non-arteritic anterior ischemic optic neuropathy. However, this finding was not confirmed by Ma et al. when comparing with a hypertensive group [31, 32]. FMD was also found to be reduced in patients with branch retinal vein occlusion compared to both healthy and hypertensive groups [33]. Additionally, some authors have suggested a link between endothelial dysfunction and the pathophysiology of wet age-related macular degeneration [12].

The findings of our study suggest that eye micro- circulatory status is associated with vascular endothelial function in systemic circulation. However, further investigations involving larger cohorts are warranted to validate this relationship and identify parameters that can aid in predicting cardiovascular risk based on retinal and choroidal parameters.

Our study has several noteworthy limitations. Firstly, edge tracking software was not used in the calculation of FMD, which may have posed a risk of less accurate capture of maximum arterial diastole. Additionally, the study did not encompass a large cohort, potentially resulting in undetected relationships. Finally, despite statistical insignificance, the control group contained more females.

CONCLUSION

This is the first study to report the association of retinal capillary plexus density parameters with peripheral vascular endothelial function expressed by FMD in a healthy population and type 2 diabetes. Higher retinal vascular density, especially in the deep capillary plexus, was found to be associated with higher FMD. This relationship may, in the future, be used to identify additional ocular indicators of cardiovascular function.

REFERENCES