In view of the rising incidence of diabetes mellitus the search is on for new diagnostic methods to serve as markers of early diabetic changes. Diabetic retinopathy is the primary microvascular and neurodegenerative complication of diabetes. It is also the main ocular complication of diabetes and the most common reason behind vision loss among individuals between 20 and 74 years of age(1). Concurrently with the rise in the incidence of diabetes, there is a noticeable increase in the occurrence of chronic microvascular complications. After 20 years, diabetic retinopathy develops in nearly all patients with type 1 diabetes mellitus and in 60% of individuals with type 2 diabetes(2,3). A key role is attributed to chronic hyperglycemia leading to microcirculation disorders. Other contributory factors include duration of the disease, comorbid hypertension, hyperlipidemia, smoking, and pregnancy. While the risk factors are well-documented, the exact pathogenesis of diabetic retinopathy is not fully understood, and the mechanisms causing vascular damage are multifaceted. The onset of the disease is influenced by a variety of aspects, including genetics, and metabolic and vascular conditions.
Evaluation of vascular flow in the retina and its supplying blood vessels by color Doppler ultrasound could serve as a valuable tool in expanding our understanding of the progression of diabetic retinopathy.
Doppler ultrasound imaging is a non-invasive diagnostic procedure used in ophthalmology to evaluate changes in blood flow in the orbital blood vessels.
The most commonly assessed vessels include the ophthalmic artery (OA), the central retinal artery (CRA), and the temporal posterior ciliary arteries (TPCA), while the main studied parameters comprise peak systolic velocity (PSV), end-diastolic velocity (EDV), and resistance index (RI), which is a measure of peripheral vascular resistance. While it is possible to determine blood velocity in these vessels, direct measurement of their diameter poses a challenge because of their small size. Despite this limitation, velocity is regarded as a good indicator of blood flow(4,5).
There are only scarce reports in the literature that specifically address blood flow in the orbital vessels in diabetic patients. Most measurements are carried out in small, heterogeneous patient groups, which diminishes the diagnostic value of results. In addition, discrepancies in the findings are frequently noted. Some sources have reported a decrease in PSV, EDV and RI in retrobulbar blood vessels in patients with diabetes, while other authors have presented opposing observations(6,7,8,9,10,11,12). The discrepancies in findings might be due to differences in patient selection criteria and heterogeneity of study groups in terms of diabetes type, patient age, duration of diabetes, control of metabolic parameters, stage of retinopathy, and treatment method.
The aims of the study were to assess the blood flow parameters in the orbital vessels in the OA and CRA in patients with type 1 diabetes mellitus without changes in the fundus and with mild nonproliferative retinopathy, to compare the parameters with a control group of healthy individuals, and to examine the correlations between the blood flow values in the orbital vessels and the duration and metabolic control of diabetes measured by the level of glycated hemoglobin (HbA1c).
The study included a total of 161 participants, with 80 individuals aged 18–45 diagnosed with type 1 diabetes mellitus and 81 healthy individuals of a similar age range, serving as the control group. The study exclusion criteria included comorbidities, systemic diabetic complications, obesity, smoking, medications other than insulin, vision defects (up to ±3.0 diopters (D)), current or past eye conditions that may affect ocular vascular flow, ocular trauma, and history of ophthalmic surgery.
The participants gave their informed consent to take part in the study, which was approved by the Ethics Committee (decision No. KB/217/2016).
The study encompassed individuals aged 18–45 years who had been diagnosed with type 1 diabetes mellitus and treated for a period ranging from one to 25 years. The patients were divided into groups based on the presence or absence of diabetic fundus changes. Another classification was determined by the degree of diabetes control, as assessed by the level of HbA1c: group HbA1c 5.0–7.0%, group HbA1c 7.1–13.0%. The third criterion was the duration of diabetes (evaluated in increments of five years).
All participants in the study underwent a general medical assessment and laboratory tests. The results, including renal function and lipid profile, fell within the normal range for all subjects. A full ophthalmic examination with pupil dilation and indirect ophthalmoscopy using a Volk +90D focusing lens in white and red-free light were performed. Color and red-free fundus images were taken using a Topcon TRC-NW7SF Mark II fundus camera at 30° and 45°. Optical coherence tomography (OCT) was performed along with fluorescein angiography (FA) in the participants who consented to the examination with a contrast agent. Diabetic alterations were assessed according to the ETDRS (Early Treatment Diabetic Retinopathy Study) scale. Depending on the findings of the fundus examination, diabetic patients were allocated to a group without retinopathy or with mild non-proliferative retinopathy.
Doppler ultrasound examinations of blood flow in the orbital vessels were carried out using a Siemens Acuson X300 ultrasound system equipped with a VF10-5 linear transducer. Blood flow was assessed in the OA and CRA. During the examination, the patients were positioned lying on their back with closed eyelids. A linear probe with ultrasound gel was used, with no pressure applied to the eyeball. By capturing an image of the eye in B-scan presentation, the optic nerve was identified, serving as a reference point for the retrobulbar vessels. The visualization of the ophthalmic and central retinal arteries was routinely achieved with the eye positioned in a straight-ahead gaze. The OA was typically identified nasal to the optic nerve, following its path towards the medial part of the optic nerve. The CRA was visualized along with the homonymous vein in the lumen of the optic nerve, with measurements conducted prior to its passage through the lamina cribrosa. Adopting an appropriate scan depth, a narrow sampling gate (1.5 mm) was set in the vessel lumen, with an angle correction of 0–30°. In this set-up, a flow velocity spectrum was recorded for the OA (Fig. 1) and CRA (Fig. 2).
The following blood flow parameters were analyzed:
peak systolic velocity (PSV, m/s), i.e. maximum velocity of blood flow during the systolic phase of the cardiac cycle (heart contraction); end-diastolic velocity (EDV, m/s), i.e. minimum velocity of blood flow during the diastolic phase before the onset of heart contraction; vascular resistance index (RI), or Pourcelot index, i.e. ratio of the difference between systolic and diastolic velocities to the systolic velocity of blood flow.
Upon preliminary examination, the variables were found to deviate significantly from a normal distribution, prompting the adoption of non-parametric methods for the analysis. The basic descriptive statistics used in the study included the median and mean values, standard deviation (SD), confidence interval for the mean (95% CI), interquartile range (25–75P), and the minimum and maximum values. The primary measure compared was the median with interquartile range.
The level of statistical significance was set at α = 0.05, and compared with the test probability ‘
The study involved a cohort of 161 individuals. The type 1 diabetes mellitus group comprised 80 patients, including 34 women and 46 men between the ages of 18 and 45 years (mean age: 28.85 years). The control group consisted of 81 healthy volunteers aged 20–39 years (mean: 26.53 years), including 53 women and 28 men.
In the study group, the duration of diabetes ranged between one and 33 years (mean: 10.4 years). Diabetic control was assessed on the basis of the level of HbA1c and varied from 5.4 to 13.0% (mean: 7.58%). In the group of patients receiving insulin treatment, insulin injection pens and personal insulin pumps were used by 60 and 20 individuals, respectively. All the patients had normal blood pressure readings. The biochemical parameters (lipid, renal, and hormonal profiles) were within the normal ranges. Furthermore, no systemic complications of diabetes were found among the study participants. Aside from insulin, the patients did not use any medications and were non-smokers.
Examinations were carried out for the right and left eyes. The right eye was selected for the assessment and analysis of results. However, where the right eye met the study exclusion criteria, the left eye was selected.
Both the healthy subjects and patients had full visual acuity for both distance and near vision, and normal intraocular pressure. Appropriate correction was used to accommodate individuals with visual impairment. In the control group, no abnormalities were noted in the anterior segment and fundus of the eyes.
Patients diagnosed with type 1 diabetes mellitus were classified based on the fundus examination and digital photography, using the ETDRS (Early Treatment Diabetic Retinopathy Study) scale. Two groups were set up: one consisting of patients without clinical fundus lesions, and another with patients presenting with mild non-proliferative retinopathy, with lesions including sparse micro-aneurysms, petechiae, and isolated hard exudates, without concurrent macular edema.
The study determined the values of blood flow parameters (PSV, EDV and RI) in the ophthalmic and central retinal arteries in a group of healthy volunteers and in patients with type 1 diabetes mellitus.
Detailed descriptive statistics for blood flow parameters (PSV, EDV, RI in the CRA and OA) in the control group are shown in Tab. 1.
Detailed descriptive statistics for blood flow parameters (PSV, EDV, RI in the OA and CRA) in the control group
80 | 20.1 | 54.3 | 34.4 | from 29.05 to 40.10 | |
79 | 4.3 | 15.1 | 8.6 | from 6.65 to 10.10 | |
80 | 0.6 | 0.83 | 0.75 | from 0.73 to 0.77 | |
80 | 8.7 | 14.5 | 11.1 | from 10.30 to 12.00 | |
80 | 1.9 | 5.7 | 3.8 | from 3.30 to 4.35 | |
80 | 0.52 | 0.75 | 0.65 | from 0.62 to 0.68 |
Detailed descriptive statistics for blood flow parameters (PSV, EDV, RI in the CRA) in the study group are shown in Tab. 2.
Detailed descriptive statistics for blood flow parameters (PSV, EDV, RI in the CRA) in the group with type 1 diabetes mellitus
80 | 6 | 14.7 | 10.5 | from 8.70 to 11.90 | |
80 | 1.5 | 5.3 | 3 | from 2.40 to 3.70 | |
80 | 0.55 | 0.82 | 0.7 | from 0.65 to 0.74 |
Detailed descriptive statistics for blood flow parameters (PSV, EDV, RI in the OA) in the control and study groups are shown in Tab. 3.
Descriptive statistics for blood flow parameters in the OA in the group with type 1 diabetes mellitus
80 | 17.9 | 73.9 | 33.9 | from 28.95 to 39.95 | |
79 | 3.2 | 27.7 | 7.3 | from 5.90 to 10.67 | |
79 | 0.55 | 0.9 | 0.78 | from 0.71 to 0.81 |
Statistical relationships of blood flow parameters in the OA and CRA between the control and study groups are shown in Tab. 4.
Statistical relationships of blood flow parameters in the OA and CRA between the control and type 1 diabetes groups
80 | 33.9 | from 28.950 to 39.950 | 80 | 34.4 | from 29.05 to 40.10 | 0.985 | |
79 | 7.3 | from 5.900 to 10.675 | 79 | 8.6 | from 6.65 to 10.10 | 0.1376 | |
79 | 0.78 | from 0.712 to 0.810 | 80 | 0.75 | from 0.73 to 0.77 | 0.0805 | |
80 | 10.5 | from 8.700 to 11.900 | 80 | 11.1 | from 10.30 to 12.00 | 0.0064 | |
80 | 3 | from 2.400 to 3.700 | 80 | 3.8 | from 3.30 to 4.35 | <0.0001 | |
80 | 0.7 | from 0.650 to 0.740 | 80 | 0.65 | from 0.62 to 0.68 | <0.0001 |
Among individuals with type 1 diabetes mellitus, there were statistically significant differences in blood flow parameters in the central retinal artery (a decrease in PSV and EDV, and an increase in RI) compared to the control group (
Differences in blood flow parameters were found between the group of patients without fundus changes and the group of subjects with mild non-proliferative retinopathy. The patients with retinopathy showed a significant decrease in PSV and EDV. In patients with mild non-proliferative retinopathy, the median values of maximum and minimum flow in the central retinal artery were significantly lower: 8.0 (6.9–10.8) vs. 10.7 (9.2–12.0;
Statistical comparison of blood flow parameters in the OA and CRA in the group with type 1 diabetes mellitus depending on the presence of mild non-proliferative retinopathy
65 | 34.2 | from 29.42 to 40.00 | 15 | 33.9 | from 25.82 to 37.17 | 0.321 | |
64 | 7.4 | from 5.85 to 10.50 | 15 | 7.2 | from 6.30 to 11.20 | 0.822 | |
64 | 0.78 | from 0.72 to 0.82 | 15 | 0.75 | from 0.69 to 0.79 | 0.220 | |
65 | 10.7 | from 9.22 to 12.00 | 15 | 8 | from 6.92 to 10.87 | 0.016 | |
65 | 3.2 | from 2.50 to 3.72 | 15 | 2.7 | from 2.20 to 2.97 | 0.024 | |
65 | 0.7 | from 0.64 to 0.74 | 15 | 0.7 | from 0.65 to 0.72 | 0.917 |
Analyzing the blood flow parameters depending on the degree of diabetes control measured by the level of HbA1c, a statistically significant decrease in PSV was found in the central retinal artery in the group with poorly controlled diabetes, with PSV (
Statistical comparison of blood flow parameters in the OA and CRA in the group with type 1 diabetes mellitus depending on the degree of diabetes control (HbA1c)
46 | 31.3 | from 26.50 to 36.30 | 34 | 37.8 | from 32.90 to 43.40 | 0.0024 | |
46 | 6.8 | from 5.80 to 9.20 | 33 | 9 | from 6.25 to 11.82 | 0.0514 | |
46 | 0.78 | from 0.70 to 0.82 | 33 | 0.76 | from 0.72 to 0.81 | 0.7959 | |
46 | 9.7 | from 8.10 to 11.10 | 34 | 11.45 | from 9.50 to 12.00 | 0.0382 | |
46 | 2.8 | from 2.30 to 3.60 | 34 | 3.25 | from 2.70 to 3.70 | 0.0679 | |
46 | 0.695 | from 0.64 to 0.74 | 34 | 0.7 | from 0.65 to 0.73 | 0.9418 |
Furthermore, in the group with poorly controlled diabetes, an additional impact of retinopathy on blood flow parameters in the central retinal artery was observed, manifesting as a decrease in both systolic and diastolic velocities. In the subset of patients with long-term diabetes (over 20 years) and unstable glycemic control, a significant decline in PSV was found in the central retinal artery. There was no link between the methods of insulin delivery (insulin injection pen, personal insulin pump) and blood flow parameters in the examined arteries.
The ophthalmic artery was characterized by variability of blood flow parameters depending on the level of diabetes control, manifested as a significant reduction in PSV in the group with poorly controlled glycemia, as determined by the level of HbA1c. A statistically significant decrease in PSV was found in the OA, with the PSV values (
The assessment of blood flow parameters in the orbital vessels was carried out in a cohort of young patients with type 1 diabetes mellitus, without comorbidities, who had not yet developed diabetic changes in the fundus or exhibited early signs of mild non-proliferative diabetic retinopathy. By selecting this group of patients, it was possible to eliminate potential factors contributing to changes in the characteristics of blood flow in the orbital vessels in patients with type 1 diabetes mellitus, and to determine whether alterations in blood flow parameters in the CRA and OA could be seen in this group of patients despite the absence of changes or early changes in the fundus of the eye.
The study involved patients with a diagnosis of type 1 diabetes mellitus who exhibited changes in blood flow parameters compared to the control group, specifically a statistically significant reduction in PSV and EDV in the central retinal artery along with an increase in RI in relation to the control group.
A correlation was found between the blood flow parameters and the progression of changes in the eye fundus in this group. Alterations in blood flow parameters were evident already in individuals without diabetic changes in the fundus. However, the presence of retinopathy further decreased the blood flow velocity (specifically PSV and EDV) in the central retinal artery compared to the group of patients where no retinal changes were observed. The findings obtained in the study point to the early onset of hemodynamic alterations in the CRA, preceding the emergence of clinical abnormalities in the fundus.
Depending on the level of glycated hemoglobin (HbA1c) which serves as an indicator of diabetes control, changes in blood flow values in the OA and CRA were noted. In the group of patients with uncontrolled diabetes, significant reductions in PSV were observed both in the OA and CRA.
The duration of diabetes was found to have an impact on changes in blood flow parameters in the CRA, manifesting as a decrease in PSV, in patients who had diabetes for more than 20 years. In this group, a significant difference was observed between the individuals with stable glycemia compared to those with elevated levels of HbA1c.
The ophthalmic artery was characterized by variability of blood flow parameters depending on the level of diabetes control, manifested as a significant reduction in PSV in the group with poorly controlled glycemia.
Chronic hyperglycemia disturbs the body’s autoregulatory mechanisms, contributing to the onset of diabetic retinopathy by altering the diameter of blood vessels, reducing the perfusion pressure, and inducing changes in peripheral resistance(13,14). The authors’ research suggests that changes in blood flow within the orbital vessels occur in patients without and with retinopathy. Major abnormalities in blood flow parameters, specifically a significant decrease in the measured values of PSV and EDV, are noted in the central retinal artery. The deviations correlate with the severity of fundus lesions, metabolic control of diabetes, and duration of the condition. In the OA, variability in blood flow parameters was shown, depending on diabetes control, with a decrease in PSV observed in patients with elevated levels of HbA1c. Hemodynamic changes observed in the central retinal artery among the study participants with type 1 diabetes mellitus without evident fundus lesions and with mild non-proliferative retinopathy might, therefore, reflect early vascular changes. The central retinal artery seems to hold particular significance as a vessel for evaluating these initial alterations. As a branch of the ophthalmic artery, the central retinal artery is anatomically terminal, and studies have found significant changes in blood flow parameters in this vessel.
Blood flow in the vessels is contingent upon both perfusion pressure and vascular resistance. These two parameters, in turn, are influenced by a range of hyperglycemia-sensitive local factors encompassing blood vessel structure and diameter, obstacles in the flow path, neurogenic and angiogenic factors, and blood rheology(15,16).
The outcomes derived from this study, including diminished PSV and EDV levels in the CRA compared to the healthy group, might suggest that ischemia and inadequate perfusion precede the onset of clinical signs of retinopathy, and increased RI in the CRA may be linked to elevated peripheral vascular resistance.
A review of publications addressing the application of Doppler ultrasound for evaluating blood flow in the orbital vessels in diabetic patients reveals discrepancies in study findings. They may be due to the absence of standardized eligibility criteria across the study groups, including variations in age ranges(7,8,9,17) and type of diabetes (most publications focus on correlations with type 2 diabetes or combined diabetes types)(18,19). The groups under comparison also exhibited variations in the duration of the disease and the advancement of fundus lesions. The most commonly described hemodynamic parameters of the orbital vessels include PSV, EDV, and RI. A number of authors have reported a decrease in orbital blood flow velocity in diabetic patients, while others have observed an increase in blood flow. Vascular resistance index demonstrates variability across authors as well.(6,12,13,20).
There are only scarce reports in the literature describing hemodynamic changes in the orbital vessels in young patients with type 1 diabetes mellitus who have not yet developed diabetic changes in the fundus or the abnormalities are still in the early stages. The presence of statistically significant blood flow disruptions in the CRA in patients with type 1 diabetes mellitus without fundus changes and with early signs of retinopathy, as demonstrated in our study, appears to be an important observation, aligned with a limited number of existing reports in the literature and findings from other studies employing alternative measurement techniques for the CRA(21) or based on animal models(22), suggesting that reductions in CRA flow precede ocular fundus changes.
Our study findings align with the results of the metaanalysis of 13 studies conducted by Meng
In summary, the findings of the reported study may indicate that alterations in blood flow parameters in the central retinal artery reflect vascular dysfunction preceding the onset of visible signs of retinopathy. Thus, they might serve as sensitive markers of hemodynamic disorders in patients with type 1 diabetes mellitus without fundus changes, and they are more prominent in patients presenting with signs of retinopathy. Uncontrolled diabetes, measured by the level of HbA1c, further decreases vascular flow velocity both in the CRA and OA. The duration of diabetes and the level of glycemic control, which are widely recognized as the most critical risk factors for the development of microangiopathy, were also mirrored in the hemodynamic changes in the orbital vessels found in the study. However, what appears to be particularly important is the fact that changes in blood flow parameters occurred in the early stages of diabetes, even before the onset of retinopathy, which highlights the value of the adopted research methodology and substantiates the need for further studies in this area.