Association between resistivity index of central retinal artery and severity of diabetic retinopathy

Abstract

Purpose:

Diabetic retinopathy (DR) is a leading cause of ocular morbidity. Its progression depends mainly on retinal vasculature and ocular blood flow. Color Doppler imaging (CDI) is a noninvasive imaging technique that measures blood flow velocity. The resistivity index (RI), calculated by the CDI, reflects the vascular resistance distal to the measuring location. RI is independent of the doppler angle and position of the patient, making it a reliable and reproducible parameter. To the best of our knowledge, there is only one study in literature studying the association between resistivity index (RI) of the central retinal artery (CRA) and severity of DR.

Aim:

To determine the association between RI of CRA and severity of DR. To determine the association between RI of CRA and spectral-domain optical coherence tomography (SD-OCT) biomarkers for DR.

Methods:

Type II diabetics visiting our OPD underwent DR screening and were graded into three categories according to ETDRS classification which include Group A—No diabetic retinopathy (No DR), Group B—Nonproliferative diabetic retinopathy (Moderate–Severe–Very Severe NPDR), and Group C—Proliferative diabetic retinopathy (PDR). SD-OCT was performed. Ultrasonic color doppler imaging was done. RI of the CRA was noted. It was compared between the three groups and its association with severity of DR and OCT biomarkers (central subfield thickness, cube average thickness and ellipsoid zone disruption) was studied.

Results:

56 eyes of 28 patients were included in our study with 20 in Group A,14 in Group B, and 22 in Group C. RI of CRA compared within groups showed statistically significant association with severity of DR (P < 0.001). The presenting BCVA (LogMar) showed positive correlation with RI in all groups. OCT biomarker central subfield thickness showed a positive correlation with RI in Groups A (P < 0.001) and B. Ellipsoid zone (EZ) disruption showed a statistically significant association with RI in Group C (P < 0.001).

Conclusion:

The RI of CRA is a reliable biomarker for the assessment of the severity of DR. Patients with high RI of CRA had higher chances of EZ disruption and presented with poor visual acuity.

Diabetes mellitus (DM) is a leading cause of ocular morbidity. It manifests as an array of microvascular complications like retinopathy, neuropathy, and nephropathy. The progression of diabetic retinopathy (DR) depends mainly on retinal vasculature and ocular blood flow. The four main risk factors affecting the progression of DR are the age of the patient, duration of diabetes, mean glycated hemoglobin (HBA1c), and diastolic blood pressure.[1] The central retinal artery (CRA) supplies the inner six layers of retina.[2] Changes in retinal blood flow and choroidal blood flow have been reported in patients with DR.[3] SD-OCT is a noninvasive diagnostic imaging technique to examine cross sections of the retina in high resolution for qualitative and quantitative analysis. It produces reliable, reproducible, and objective retinal images. The second band on OCT is called as ellipsoid zone (EZ) of photoreceptors.[4] Color Doppler imaging [CDI] is a noninvasive imaging technique that measures blood flow velocity. CDI has successfully demonstrated changes in orbital hemodynamics associated with a variety of pathological conditions, including central retinal artery and vein occlusions, cranial arteritis, nonarteritic ischemic optic neuropathy, and carotid disease.[5] In one cardiac cycle, the peak systolic velocity (PSV), end-diastolic velocity (EDV), and mean blood flow velocities are measured. The resistivity index (RI), calculated by the machine using PSV and EDV reflects the vascular resistance distal to the measuring location. RI is independent of the doppler angle, making it a reliable and reproducible parameter. It is independent of the position of the patient.[6] Previous study[7] has shown no significant difference in RI, when CDI was done by different qualified radiologists.

Sandeep Saxena et al.[7] proved that RI of CRA is a reliable bio imaging biomarker for the severity of DR. Yoshitaka Ueno et al.[8] suggested that diabetes mellitus (DM) causes pathological changes in blood vessels that leads to a reduction in the blood flow and increase in resistance, which leads to a hypoxic state causing sight-threatening complications in DR. These observations gave us an idea to study the RI of CRA and its association with severity of DR.

Methods

A hospital-based, cross-sectional observational study was conducted in the department of ophthalmology in collaboration with the department of radiology for a period of 1.5 years. Patients with type 2 DM without any preference for sex were included and irrespective of severity, and both eyes of patients were included. Patients with preexisting glaucoma, with other coexisting retinal pathologies, with significant cataracts (in whom OCT had poor signal strength), on antioxidants, antiplatelet drugs, anticoagulants, and calcium channel blockers were excluded from the study.

We aimed to determine the association between the RI of the CRA and the severity of DR and also to determine the association between the RI of CRA and the SD-OCT biomarkers for DR.

Patients who met the inclusion criteria were subjected to detailed ophthalmic evaluation. Visual acuity was assessed by Snellen's chart and was converted into a LogMar chart for statistical analysis. Thorough anterior segment evaluation by slit-lamp, fundus evaluation by 90 D lens, and direct and indirect ophthalmoscopy with 20D lens was performed. For the purpose of this study, patients were classified into three groups based on the ETDRS classification for DR, which include Group A—No diabetic retinopathy (No DR), Group B—Nonproliferative diabetic retinopathy (Moderate–Severe–Very Severe NPDR), and Group C—Proliferative diabetic retinopathy (PDR).

History of the patient's duration of diabetes and associated comorbidities were noted. Duration of diabetes was defined as the known years of disease at the time of recruitment. OCT macula (CirrusTM HD-OCT-spectral domain technology) was performed for all the patients except patients with dense vitreous hemorrhage (VH) and tractional retinal detachment (TRD). We studied central subfield thickness (CST), cube average thickness (CAT), and ellipsoid zone (EZ), in all the patients as they are proven biomarkers correlating with visual acuity. CST is defined as the thickness of a central circle of 1 mm diameter in the circular ETDRS grid map. CAT is the overall average thickness of the tissue layers between the internal limiting membrane and retinal pigmented epithelium (ILM-RPE) over the entire scanned area, the mean of thicknesses in nine sections. The SD-OCT reports were scanned by a single skilled doctor who was unaware of the stage of DR and the status of EZ was documented.

CDI was carried out by a single skilled and trained radiologist using (Philips EPIQ 5G ultrasound system). PSV, EDV, and mean blood flow velocities were noted. RI was calculated by the machine using the formula: RI = (PSV-EDV)/PSV. All the values (age, sex, duration of diabetes, associated systemic comorbidities, LogMar BCVA, RI of CRA, and OCT Biomarkers) were tabulated and statistical analysis was performed.

Statistical analysis

The mean value of RI and the standard deviation (SD) was studied in all the groups. ROC curve analysis was performed to study the efficacy of RI as a biomarker of severity of DR. RI is compared between groups and within groups using ANOVA technique. As the association was statistically significant post hoc analysis was performed. RI was taken as a dependent variable and multiple comparison analysis (Bonferroni post hoc test) was done within groups. LogMar visual acuity was correlated with RI by Pearson correlation coefficient in all the groups. SD-OCT biomarkers (CST, CAT, EZ disruption) were correlated with RI by Pearson correlation. The ANOVA technique is used to compare CST, CAT, and EZ status between groups and within groups.

Result

56 eyes of 28 patients were included in our study, out of which 20 eyes were included in Group A,14 eyes in Group B, and 22 eyes in Group C. The mean RI was calculated in all the groups individually. It was compared between all three groups. It showed a statistically significant (P value < 0.001) [Table 1] association with severity of DR, which states that RI increased with an increase in severity of DR. It was also compared between the groups individually, that is, Group A (No DR) was compared with Group B (NPDR), similarly group B with Group C, and Group A with Group C. We found a statistical significance when RI was compared between Group A and Group B and between Group A and Group C.

Table 1.

Association between RI and severity of DR

Groups	n	Mean RI of CRA	SD	P
No DR	20	0.684	0.061	<0.001
NPDR	14	0.766	0.054
PDR	22	0.787	0.051

The ROC curve analysis showed the diagnostic accuracy of RI in discriminating patients with No DR from patients with PDR (AUC-89.3%, P < 0.001) with a cutoff value of 0.735 [Fig. 1]. Out of 20 samples in the No DR group, only three samples had RI > 0.735 and 17 samples had RI < 0.735. Out of 22 samples in the PDR group, 18 samples had RI > 0.735 and only four samples had RI < 0.735. RI had a sensitivity of 80%, and specificity of 77.78%. The positive predictive value is 66.67%, and the negative predictive value is 87.50% with an accuracy of 78.57%.

Figure 1.

ROC curve—RI of CRA. With area under curve -89.3% and cutoff value of RI of CRA -0.735, P value = 0.001

On correlating LogMar visual acuity with RI, a positive correlation was observed in all the groups [Fig. 2], which shows that patients with high RI had higher LogMar visual acuity, that is, poor vision.

Figure 2.

Association between LogMar Visual acuity and RI. (a) Scatter diagram showing a positive correlation between RI and LogMar Visual acuity in Group A. (b) Scatter diagram showing a positive correlation between RI and LogMar Visual acuity in Group B. (c) Scatter diagram showing a positive correlation between RI and LogMar Visual acuity in Group C

On correlating CST with RI, in Group A, a mild positive correlation (r = 0.465) was found with statistical significance (P value of 0.039) which represents that with an increase in RI there is an increase in CST. In Group B, very low positive correlation was found with a P value of 0.810. In Group C, a low negative correlation (r = -0.198) was found with a P value of 0.480. On correlating the status of EZ disruption with RI, Group C showed a good positive correlation (r = 0.512) with statistical significance (P value of 0.043). This showed that eyes with PDR had high RI and had a significant percentage of EZ disruption on SD-OCT [Fig. 3]. Group B showed a low positive correlation (r = 0.175) (P value of 0.551). On correlating CAT with RI, Group A showed a mild negative correlation (r = -0.234) with a P value of 0.321. Group B showed a mild negative correlation (r = -0.320) with a P value of 0.265. Group C showed a mild negative correlation (r = -0.134) with a P value of 0.635. There was no statistical significance in all three groups.

Figure 3.

Association between EZ disruption and RI. (a) Scatter diagram showing a positive correlation between RI and EZ disruption in Group B. (b) Scatter diagram showing a positive correlation between RI and EZ disruption in Group C

Discussion

Our primary objective was to study the association between RI of CRA and severity of DR. RI was found to correlate significantly with severity of DR. RI was found to increase with an increase in vascular resistance in a blood vessel, which reduces the blood flow. It also represents the vascular compliance, that is, ability of vessel walls to dilate in response to autoregulation. Kohner et al.[9] reported that there is an increase in ocular blood flow in the early stages of DR and a decrease in ocular blood flow in proliferative stages based on fluorophotometry. Initially, there is significant vasodilation in response to autoregulation with an increase in the vessel wall diameter. An increase in the vessel diameter and volume of blood flow induces shear stress in the vessel wall. Along with this, there are impaired rheological properties in patients with DM[9]; that is, there is increased plasma viscosity and increased erythrocyte rigidity, which contribute to the altered blood flow. As the shear stress increases, the vascular compliance of the vessel wall reduces causing mechanical injury to the basement membrane. As the disease progresses, the vessel wall loses its ability to auto regulate due to the damage to the basement membrane and increased capillary pressure, thus affecting vascular compliance. According to Laplace law, the ability of a vessel walls to counteract the distortion pressure of a dilating vessel in response to autoregulation is inversely proportional to the radius of the vessel. In macrovessels like CRA, as the counteracting force is less, the vessel wall remains dilated in response to autoregulation. As a result, the pressure within the vessel increases, and shear stress is easily transmitted to the vessel wall damaging the basement membrane, reducing the vascular compliance. In our study, we also observed that RI increased with an increase in severity of DR, which can be explained by the reduced vascular compliance of CRA in response to increase in vessel wall damage.

L. Schmetterer et al.[10] studied the pathophysiology of ocular blood flow in patients with DR and stated that endothelin-1-mediated vasoconstriction is more prominent in ocular arteries. This vasoconstriction results in reduced ocular blood flow leading to tissue hypoxia and capillary nonperfusion, which induces the development of neovascular proliferation thus producing sight-threatening complications. Histopathological studies[11,12,13,14] have also shown that hyperglycemia is associated with an increase in the retinal vascular basement membrane thickening associated with pericyte loss, oxidative stress, and decreased capillary perfusion all of which leads to capillary endothelium damage and increased vascular resistance. All these mechanisms suggest that multiple factors contribute to the increased vascular resistance in patients with DR and the resistance increases with an increase in the severity of DR, which supports our observation in this study.

Our study indicated that there is a significant difference in RI between the No DR group and the NPDR group as well as between the No DR group and the PDR group. Arend O et al.[15] stated that, in the early stages of DM, there is increased retinal blood flow which over a long time causes damage to the vessel due to shear stress inducing the production of endothelin-1 and other mediators of vasoconstriction causing reduced ocular blood flow and increased vascular resistance which could possibly explain our observation.

Birog et al.[16] stated that the deep capillary plexus (DCP) provides 10-15% of oxygen to the inner photoreceptor segment. It derives its branches from CRA. It is located on the outer side of the inner nuclear layer. Choroidal vasculature has no auto regulatory mechanism.[17] In conditions of hypoxia, CRA through DCP supplies the inner retina and outer retinal as well. As DR progresses ischemia increases and reduces the blood supply to the outer layers of the retina resulting in thinning of the outer nuclear layer and photoreceptor disruption.[18] EZ on SD-OCT corresponds inner segment of photoreceptors and it shows a strong correlation with visual acuity. Our study also confirmed that patients with PDR and high RI had a higher frequency of ellipsoid zone disruption and poor presenting visual acuity.

In our study, SD-OCT biomarker CST showed a positive correlation with RI in all the three groups. This could be explained by a study performed by Bhagat N et al.,[19] which states that hypoxia, ischemia, and accumulation of free radicals lead to increased production of VEGF, which causes a breakdown in the blood–retinal barrier (BRB). Breakdown of the BRB leads to the accumulation of plasma proteins which leads to intraretinal edema and retinal thickening. Damage to the blood vessels results in the production of vasoconstrictors which increase the resistance to blood flow, thus increasing the RI. Saxena et al.[7] stated that CST and CAT increased with an increase in severity of DR. They also stated that an increase in CST and CAT was associated with increased odds of EZ disruption. A positive correlation between CAT and RI of CRA could not be established in our study. Table 2 summarizes the results of the previous studies done in the same context and compares it with our study.

Table 2.

Comparison of our study with previous studies in literature

Study	Sample	RI and Severity of DR	RI and EZ disruption	RI and CST	RI and CAT	RI and LogMar BCVA
Arai T, et al.,[23] 2000 Cross-sectional	52 eyes	Significant	Not Done	Not Done	Not Done	Not Done
Manila Khatri, Sandeep Saxena, et al.,[7] 2020 Cross-sectional Lucknow - KGH	81 eyes	Significant	Significant	Significant	Significant	Significant
My study 2020-2021 Cross-sectional Kochi - AIMS	56 eyes	Significant	Significant in Group C Positive correlation in Group B	Positive correlation in Group A, B Negative correlation in Group C	Negative correlation in all Groups (A, B, C)	Positive correlation in Group A, B, C

Hyperglycemia induces microangiopathy (retinopathy, nephropathy, and neuropathy) and macroangiopathy. Macro vascular disease is more diffuse, severe, and is an accelerated form. Atherosclerosis is also known to progress aggressively in diabetics due to impaired metabolism. Diabetic macroangiopathy is characterized by intraplaque new vessel formation due to abnormal neo vasculogenesis and angiogenesis leading to an atherosclerotic plaque rupture and plaque hemorrhage, causing cardiovascular and cerebrovascular dysfunction.[20,21,22] As CRA is a macro vessel, the RI also represents the status of macroangiopathy in diabetics. High RI can also be used as an indication for complete systemic workup and early intervention to prevent life-threatening complications like stroke and cardiac arrest.

Limitations

Limited sample size due to COVID-19. Confounding factors for vision like macular edema could not be eliminated completely. Ours was a cross-sectional study, a prospective study in this direction can help us to study the efficacy of RI as a risk factor for progression of DR and DME. Once the relationship is established, it can also be used to predict the response to treatment in patients with DME.

Conclusion

RI of CRA, a vascular parameter, has diagnostic accuracy in discriminating patients with DR and PDR. This makes it a sensitive and specific biomarker for the severity of DR. It has a significant positive correlation with the severity of DR. It also showed a positive correlation with LogMar visual acuity, confirming that patients with high RI presented with poor vision. With an increase in RI, there is a high chance of ellipsoid zone disruption on SD-OCT. EZ disruption is a documented OCT biomarker that is associated with poor prognosis even post-treatment. Considering the significant association between EZ disruption and RI, patients with high RI are likely to have poor responses to treatment.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.