Background
Cardiovascular disease is a common comorbidity of type 2 diabetes (T2D), affecting approximately one-third of patients with T2D worldwide [
1]. In addition to being a risk factor for the development of coronary heart disease, diabetes also affects the prognosis of acute coronary syndrome (ACS) [
2]. Zhou et al. showed that patients with ACS and diabetes had a two-fold increased risk of all-cause death and a 1.5-fold increased risk of major adverse cardiovascular and cerebrovascular events, compared to patients without diabetes [
3]. Macrovascular and microvascular complications are common in patients with diabetes. Studies have analyzed the effects of diabetes and ischemic heart disease on systemic blood flow, pathological mechanisms, and prognosis from the perspective of ocular blood flow [
4‐
6]. Most of these studies are based on the retinal vascular system, because of its clinical observability. In vitro studies have shown that diabetes reduces retinal Connexin 43 (Cx43) expression, and the development of pericyte loss and acellular capillaries is associated with decreased Cx43 expression in diabetic retinopathy [
7]. However, retinal vessels are affected by many confounding factors, such as age, sex, blood pressure, blood lipids, smoking, drinking, and drugs, which are difficult to completely exclude during research [
8].
The ophthalmic artery (OA), as the origin of the blood supply to the eye, can directly and accurately reflect relatively early changes in blood supply. At present, most studies have used color Doppler imaging (CDI) to observe hemodynamic changes in the OA. Most of these studies concluded that the resistive index was significantly higher in patients with T2D, while other hemodynamic changes were inconclusive [
9‐
11].This may be related to the fact that the results measured by CDI are significantly affected by human factors [
12]. Changes in patient position during the CDI examination, examiners’ operating pressure, and the probe position may influence hemodynamic results. It is difficult for CDI to locate the origin of the OA or display the full length due to the course variability and small diameter of the OA. In recent years, blood simulations based on computational fluid dynamics (CFD) have been widely used in the hemodynamic analysis of cardiovascular diseases [
13‐
15]. CFD numerical simulations based on computed tomography angiography (CTA), which can accurately obtain hemodynamic information on diseased vessels and reflect the real blood flow status in the body.
We previously reconstructed three-dimensional (3D) OA models and adopted CFD numerical simulations to examine the morphological and hemodynamic features of OA in patients with ACS and found a slower blood velocity [
16]. Based on the previous study, the current study aimed to explore the morphological and hemodynamic characteristics of the OA in patients with T2D, with or without ACS, and to assess the correlation between OA characteristics and clinical indicators.
Discussion
We reported several differences in OA morphology and hemodynamics in patients with ACS only, T2D only, and with both ACS and T2D. It is difficult to obtain hemodynamic data due to the small diameter and complex course of the OA. Most previous studies on OA diameter or blood velocity were based on autopsies [
19] or CDI. In comparison, the method in this study could show the full length of the OA and is less affected by human factors.
Nonetheless, in a study that also measured the OA diameter based on CTA [
20], the results were different, which may be due to the following reasons. First, the reconstruction methods were different. Second, there was a different measurement method compared to the diameter we measured at the OA’s origin, wherein they measured a diameter of 5 mm from the beginning of the OA. Finally, our control group comprised not completely healthy participants but also matching with the patients in the disease groups in terms of age, sex, smoking history, and hypertension. In our study, the OA diameter of patients with T2D only was smaller than that of patients with ACS, which may be a response of blood vessels to a high glucose status. Consistent with our results, previous studies have reported smaller retinal artery diameters in patients with diabetes. Moreover, arteriole diameter decreased with increasing duration of diabetes, and this difference could not be explained by age, mean arterial blood pressure, or smoking [
21,
22]. The main causes of abnormal vascular diameter may be the dysfunction of endothelial cells and abnormal tension of smooth muscle cells and pericytes. Local vasoactive substances play an important role in the regulation of the vascular wall. Additionally, endothelin-1 (ET-1) is a potent endogenous vasoconstrictor. Endogenous expression of ET-1 is increased in diabetes, and ET-1 mRNA levels in patients with diabetes are significantly higher than those in patients without diabetes [
23]. ET-1 tends to decrease the retinal artery diameter, but it has no effect on the retinal vein diameter [
24]. Notably, ET-1 is a vascular tension regulator of the OA, which can induce strong vasoconstriction and regulate local blood flow [
25]. Therefore, it is hypothesized that the increase in ET-1 level is possibly one of the main reasons for the decrease in OA diameter in patients with diabetes. Regarding macrovascularity, studies in recent years have reported its protective effect on aortic aneurysm disease [
26,
27], which is associated with the decrease in matrix metalloproteinase activity by high glucose level [
28], the non-enzymatic cross-linking of advanced glycation end products between the basement membrane of the extracellular matrix, and the stimulation of the tumor growth factor-β signaling pathway to maintain vascular smooth muscle cell homeostasis [
27]. However, studies have found that a smaller retinal artery diameter is associated with the incidence of ACS, especially in women [
29], which is inconsistent with the larger OA diameter in our study. This may be affected by the difference in observation time, because, in our study, patients with ACS underwent CTA examination during their hospitalization period. We hypothesize that this is related to the compensatory regulatory mechanism of the ocular blood vessels [
25].
In previous studies, CDI has been used to measure retrobulbar blood flow in patients with diabetes. It was found that there was a high resistance index in the OA of patients with diabetes. However, the results of OA blood velocity measurements are controversial. Such inconsistencies may be related to different measurement techniques, locations, and differences in patient population characteristics [
9‐
11]. After blood simulation, we found that the OA blood velocity in patients with ACS was relatively low, which is consistent with the results of our previous study [
16]. In this study, the blood velocity in patients with T2D only was lower than that in the control group, but it was higher than that in both ACS groups. This may be due to the increased resistance to outflow from the OA [
5]. It is now well established that Cxs plays a vital role in maintaining the activity of microcirculation homeostasis, as well as in pathologies that involve a tight regulation and coordination between cells in the blood vessel wall and circulating blood cells such as atherosclerosis and hypertension [
30,
31]. In diabetic retinopathy, breakdown of the blood-retinal barrier is typical and has been associated with pericyte loss and endothelial cellecell junction breakdown, consistent with reports that downregulation of Cx43 may promote vascular loss in the diabetic retina [
32]. Müller cell-pericyte communication is critical for maintenance of retinal homeostasis, and that high glucose-induced Cx43 downregulation plays a critical role in the demise of Müller cells and pericytes [
33]. Oxidative stress, involved in mitochondrial dysfunction and apoptosis in early damage of diabetic vascular complications [
34]. Therefore, maintaining mitochondrial function and attenuating reactive oxygen species production in retinal endothelial cells could be considered a significant approach for treating diabetic retinopathy. Although the potential use of mitochondria-targeted therapies in the regulation of diabetic retinopathy remains unclear, studies have shown that targeting of the mitochondrial superoxide scavenging by overexpression of mitochondrial superoxide dismutase prevents the development of retinopathy in diabetic mice [
35,
36]. Intracranial vessels are susceptible to oxidative stress, and destruction of the intracranial carotid blood–brain barrier may precede the formation of atherosclerotic lesions [
37]. As a branch of the ICA, OA velocity is easily affected by internal carotid atherosclerotic lesions. In addition, hemodynamic characteristics determine vascular integrity and blood cell transport [
38]. Red blood cell deformability was lower in patients with T2D and formed chain-like structures under low shear conditions. Excessive accumulation of red blood cells leads to arteriole blockage, which prevents normal oxygen transport in the surrounding tissues [
39]. Notably, the effect of flow perturbation on endothelial cell dysfunction is associated with atherosclerosis [
40]. The flow perturbation caused by ACS may also be related to the slowing of the blood flow velocity in the OA. Frank et al. [
41]. used CFD to redefine the central role of hemodynamically related microscopic lesions in atherogenesis. They suggested that hemodynamic stress-induced endothelial rupture triggered an inflammatory process that promoted atherosclerosis. This indicates that hemodynamic changes precede pathophysiological changes.
Diabetes causes ischemia in peripheral arterioles, whereas ACS, the acute manifestation of ischemic heart disease, may also affect the blood supply [
42]. The incidence of ischemic stroke in patients with diabetes is two to three times higher, due to the combined effect of multiple risk factors for atherosclerosis [
37]. Our calculation of the OA mass flow rate showed that there was a decrease in all disease groups, and patients with ACS had a lower flow rate than patients with diabetes. These results further confirm the influence of diabetes and ACS on peripheral blood flow and provide evidence on the pathogenesis of ocular ischemic lesions. Chronic hyperglycemia leading to diabetic retinopathy remains the most common vascular complication in patients with diabetes [
43]. Retinal microaneurysm is one of the earliest clinical symptoms of diabetes and one of the key lesions in diabetic retinopathy severity classification. Microaneurysms are local dilations of capillaries. The mechanisms of microaneurysms are unclear, including increased luminal pressure, retinal microenvironment changes, endothelial cell damage due to leukocyte deposition, and pericyte loss [
44]. Clinically, patients with ACS are at a higher risk not only for ocular ischemia but also for pressure changes than those with coronary artery disease. Patients with ACS were significantly more likely to develop retinal microaneurysms and dot bleeding than those with stable coronary artery disease [
45]. In this study, the OA pressure was higher in patients with both ACS and T2D. We hypothesized that this may increase the risk of dilation, which may further cause lesions, such as microaneurysms and OA aneurysm. It is blood flow that determines aneurysm formation and growth by guiding the collagen remodeling of the aneurysm wall [
46]. Pressure-induced fluid changes predate pathophysiological changes [
47], and OA pressure changes occur earlier than microaneurysm formation. Therefore, the early observation of hemodynamic changes is important for early clinical detection and treatment.
In the correlation analysis, we found that the OA blood velocity and mass flow ratio were associated with several clinical indicators. Increased fasting plasma glucose and HbA1c levels are important predictors of diabetic complications and are significantly associated with the risk of diabetes, cardiovascular disease, cancer, and all-cause mortality [
48]. Although commonly used to assess heart failure, NT-proBNP is independently associated with poor outcomes in patients with myocardial infarction [
49]. Meanwhile, the combination of troponin T and NT-proBNP contributes to the identification of patients with diabetes, at an extremely high absolute risk [
50]. The hemodynamic parameters of the OA are correlated with important clinical indicators of patient prognosis, which suggests a new study direction for the prediction of disease prognosis.
This study has some limitations. The slice thickness of the CTA scan limits the accuracy of 3D reconstruction. Additionally, the same boundary conditions were applied to all groups because of the lack of relevant study data. Further studies are required to improve the completeness of the results.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.