Article Text

Download PDFPDF
Correlation of retinal arterial wall thickness with atherosclerosis predictors in type 2 diabetes without clinical retinopathy
  1. Shigeta Arichika,
  2. Akihito Uji,
  3. Tomoaki Murakami,
  4. Kiyoshi Suzuma,
  5. Norimoto Gotoh,
  6. Nagahisa Yoshimura
  1. Department of Ophthalmology and Visual Sciences, Kyoto University Graduate School of Medicine, Kyoto, Japan
  1. Correspondence to Dr Akihito Uji, Department of Ophthalmology and Visual Sciences, Kyoto University Graduate School of Medicine, 54 Shogoin Kawahara-cho, Sakyo-ku, Kyoto 606-8507, Japan; akihito1{at}kuhp.kyoto-u.ac.jp

Abstract

Aims To evaluate retinal arterial wall thickness (WT) using high-resolution retinal imaging in patients with type 2 diabetes and assess its correlation with risk factors for arteriosclerosis.

Methods Outer diameter, inner diameter and WT of the retinal artery were measured using adaptive optics scanning laser ophthalmoscopy in 28 patients with type 2 diabetes without clinically apparent diabetic retinopathy and normal volunteers. Laboratory values and intima-media thickness (IMT) in the common carotid artery were measured.

Results Retinal arterial WT was significantly greater in patients with type 2 diabetes than in controls (p=0.02). There was a significant correlation of retinal artery WT with IMT in patients with diabetes (r=0.40, p=0.04). WT in patients with diabetes was positively correlated with haemoglobin A1c (HbA1c) (r=0.49, p=0.001), total cholesterol (r=0.47, p=0.002) and low-density lipoprotein cholesterol (r=0.47, p=0.003).

Conclusions Microvasuclar thickness was greater in patients with diabetes than in controls. Furthermore, WT was positively correlated with HbA1c, total cholesterol and low-density lipoprotein cholesterol levels and IMT in the diabetic group. These results suggest that retinal artery wall measurements can be potential surrogate markers of early diabetic microangiopathy.

  • Imaging
  • Retina

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Diabetic retinopathy (DR) is one of the leading causes of blindness, and the number of patients with DR and vision-threatening DR is expected to increase.1 To prevent microvascular complications, including DR, nephropathy and neuropathy, the UK Prospective Diabetes Study group recommended intensive blood glucose control and blood pressure (BP) control in patients with type 2 diabetes mellitus (DM).2–4 The Kumamoto Study showed that the maintenance of haemoglobin A1c (HbA1c) levels below 7% prevented the development of microvascular complications in patients with non-insulin-dependent DM (NIDDM).5 Moreover, a follow-up study showed that strict glycaemic control delayed macrovascular complications, also known as the legacy effect or metabolic memory.6 Generally, DR is a metabolic disorder caused by hyperglycaemia; these increase the synthesis and accumulation of basement membrane components to result in a thickened basement membrane.7–10 However, these early changes are undetectable by standard fundus examinations. If very early microvascular changes can be visualised directly and non-invasively before the appearance of clinically apparent DR, then patients can be encouraged towards better glycaemic control and vision-threatening DR as well as prevention of life-threatening macrovascular complications.

Traditionally, evaluation of the retinal microvasculature using direct ophthalmoscopic examinations is considered to aid in the assessment of systemic diseases.11 ,12 Several retinal calibre studies to predict early microvascular changes using fundus photography before the appearance of clinically apparent DR have been conducted.13–15 In patients with diabetes, examination of retinal vascular changes revealed arteriolar dilation as a preclinical marker of early DR,16 arteriolar narrowing as a predictor of DM,13 venous narrowing as a predictor of DR17 and DR as a marker of subclinical atherosclerosis.18 However, the resolution of fundus photography is not adequate for the identification of retinal vessel walls and the measurement of their thickness; therefore, vascular calibre changes in patients with diabetes require further investigation using improved methods.

Adaptive optics technology can correct the optical wavefront aberrations and was recently found to provide high-resolution images of photoreceptors,19 ,20 blood flow21–25 and retinal vessel walls, which cannot be observed on colour fundus photographs, in a non-invasive manner.26–29

In this study, we directly visualised and evaluated the retinal arterial wall thickness (WT) using a high-resolution retinal imaging modality known as adaptive optics scanning laser ophthalmoscopy (AOSLO) in patients with type 2 diabetes without clinically apparent DR and compared these values with those in normal volunteers. In addition, we assessed its correlation with the risk factors for arteriosclerosis.

Methods

This study was approved by the Institutional Review Board and Ethics Committee at Kyoto University Graduate School of Medicine. All study conduct adhered to the tenets of the Declaration of Helsinki. After the study design and risks and benefits of participation were thoroughly explained, written informed consent was obtained from each participant.

Subjects

This prospective observational cross-sectional study included 28 patients with type 2 diabetes without clinically apparent DR (53.1±9.6 years; range, 23–72 years), who visited the Kyoto University Hospital, Kyoto, Japan, between July 2013 and February 2015. Data were collected from 31 eyes of 31 healthy volunteers (54.5±11.0 years; range, 24–71 years) retrospectively via our database of normal volunteers. We obtained AOSLO videos for both eyes of 28 patients, after dilating their pupils using one drop of tropicamide (0.5%) and phenylephrine hydrochloride (0.5%). We determined the diagnosis of DM and diabetes type from the clinical medical records of patients. Retinal specialists determined the international clinical classification via indirect ophthalmoscopic and slit-lamp biomicroscopic fundus examinations.30 We also obtained fundus photographs for all patients. Patients who had any of the following were excluded from the study: best-corrected visual acuity poorer than 20/25, high myopia (more than −6 dioptres and an axial length >26.5 mm), average systolic BP (SBP) ≥140 mm Hg and average diastolic BP (DBP) ≥90 mm Hg, intraocular pressure >21 mm Hg, discrepant DR stage between the right and left eyes, pregnancy, secondary DM and any other ocular diseases. The BP of all patients was measured after they maintained a seated posture for at least 5 min. Measurements were obtained three times on the day of examination, and the values were averaged for analysis. The 31 normal volunteers were included as the control group. All volunteers satisfied the above-mentioned criteria and did not suffer from any disease including DM.

AOSLO imaging

We used the prototype AOSLO system (Canon, Tokyo, Japan), which comprises an AO system, a high-resolution confocal SLO imaging system and a wide-field imaging subsystem, for this study. The light source wavelength was 840 nm. The imaging area was a 0.34×0.34 mm square. Details of this system are described in a previous study.31 AOSLO videos, with the optical focus on the layer where the arterial wall could be appropriately visualised, were recorded for a total of 10 s per scan area at the rate of 32 frames/s. We selected images that were clear enough to enable vessel calibre measurements. We examined each eye for approximately 10 min, with the patient in a seated posture, and used the axial length obtained using an optical biometer (IOL Master; Carl Zeiss Meditec, Dublin, California, USA) to convert the degree to the actual distance to the retina.32 A pulse oximeter (Oxypal Neo; Nihon Kohden, Japan) was attached to the earlobes of patients for synchronising cardiac pulsations with the AOSLO video frames.27

Retinal vascular measurements

We selected zone B, which was 0.5–1 disc diameter from the optic disc margin, as the area for vascular measurements,15 which were obtained using AOSLO Retinal Image Analyzer software (Canon), dedicated to our prototype AOSLO system. We have described the details of this system in a previous report.21 ,27 ,31 Raw videos were corrected for scanning distortions and were stabilised to correct for eye motion using a scan line warping method described previously.21 After image stabilisation, we generated an averaged image using stabilised frames with a specific range of relative cardiac cycles. This enabled division of the AOSLO images into five segments according to cardiac pulsations. We applied an edge-preserving filter to the frame-averaged image, following which we manually set control points for a length of at least 150 µm on the retinal artery axis and walls to detect the borders of the retinal artery walls along its course. We continuously determined the wall borders using the natural spline interpolation method on the control points. As vascular measurements, we obtained the outer diameter (OD), inner diameter (ID) and WT values for the retinal artery. A 6 μm interval was maintained between measurements on the lined vessels. Once all vascular measurements for each of the five segments were obtained, the mean of the five measurements was taken as the average value. Intraclass correlation coefficients (ICC) were used to evaluate interobserver and intraobserver agreement. All arterial measurements were independently performed by two retinal specialists (SA and AU) in a masked manner. In the analyses, the measurement values obtained by the former were adopted.

Macrovascular measurements

We reviewed the patients' medical records and obtained their laboratory values within 2 weeks of AOSLO imaging. Furthermore, we measured the intima-media thickness (IMT) in the common carotid artery using carotid ultrasonography (Xario XG; Toshiba, Tokyo, Japan or ProSoundα10; Hitachi-Aloka, Tokyo, Japan) and the cardioankle vascular index (CAVI) using a BP pulse wave inspection apparatus (VaSera VS-1500A; Fukuda Denshi, Tokyo, Japan).

Statistical analysis

All values are presented as mean±SD. Bivariate correlations were analysed using Pearson's correlation coefficient. Multivariate regression analysis (backward stepwise method) was performed to identify the predictors of WT. All analyses except ICC and multivariate regression analysis, which was performed using SPSS (V.19; IBM, Armonk, New York, USA), were performed using StatView (V.5.0; SAS Institute, Cary, North Carolina, USA). A p value of <0.05 was considered to be statistically significant.

Results

AOSLO facilitated clear arterial wall visualisation in patients with diabetes (figure 1). Clear vascular images were successfully acquired for 40 of the 56 eyes (71.4%). Clear images could not be obtained for the remaining 16 eyes because of poor eye fixation. Patient characteristics and average measurements are summarised in table 1. Compared with the control group, patients with diabetes had greater body mass index (BMI) and WT. The mean laboratory values were as follows: fasting glucose, 131.6±27.2 mg/dL; HbA1c, 9.5±2.3%; total cholesterol, 197.2±48.4 mg/dL; high-density lipoprotein (HDL) cholesterol, 51.9±14.3 mg/dL; low-density lipoprotein (LDL) cholesterol, 122.1±44.0 mg/dL and triglyceride, 163.4±109.1 mg/dL. The IMT was 0.74±0.25 mm, common carotid artery calibre was 5.5±0.62 mm, internal carotid artery calibre was 4.4±1.24 mm and CAVI was 7.6±1.00. AOSLO vascular measurements had high interobserver and intraobserver agreement with all ICCs ≥0.89 (interobserver: OD=0.89, ID=0.94, WT=0.95; intraobserver: OD=0.94, ID=0.92, WT=0.94).

Table 1

Clinical and ophthalmological characteristics of patients with type 2 diabetes without clinical retinopathy and normal controls

Figure 1

Direct retinal arterial wall visualisation in a patient with type 2 diabetes without clinical retinopathy. (A) A fundus photograph of the right eye of a man aged 62 years. His blood pressure was 109/67 mm Hg and haemoglobin A1c level was 7.1%. (B) Magnified images of the area outlined in white in panel A. (C) Adaptive optics images of the area outlined in white in panel A. Scale bar, 100 µm. (D) Adaptive optics images of the area outlined in white in panel A, with semiautomatic segmentation of the vessel borders. Scale bar, 100 µm. The outer diameter, inner diameter and wall thickness of the retinal artery are 139.7, 108.7 and 30.9 µm, respectively. (E) A fundus photograph of the left eye of a volunteer aged 53 years. His blood pressure was 108/67 mm Hg. (F) Magnified images of the area outlined in white in panel D. (G) Adaptive optics images of the area outlined in white in panel D. Scale bar, 100 µm. (H) Adaptive optics images of the area outlined in white in panel D, with semiautomatic segmentation of the vessel borders. Scale bar, 100 µm. The outer diameter, inner diameter and wall thickness of the retinal artery are 122.5, 102.3 and 20.2 µm, respectively.

Correlations between vascular measurements and predictors of atherosclerosis

IMT was significantly correlated with WT (r=0.40, p=0.04), but not with OD (r=0.01, p=0.95) and ID (r=−0.15, p=0.45; table 2). The common carotid artery calibre, internal carotid artery calibre, CAVI, age, diabetes duration and BMI did not show a correlation with any vascular measurements (tables 2 and 3). SBP was negatively correlated with OD (r=−0.33, p=0.04) and ID (r=−0.35, p=0.03), but not with WT (r=−0.009, p=0.96), while DBP was not correlated with any of these parameters (r=−0.06, p=0.73; r=−0.09, p=0.60; and r=0.07, p=0.67, respectively).

Table 2

Correlation between retinal artery parameters and systemic vascular parameters in patients with type 2 diabetes without clinical retinopathy

Table 3

Correlation between artery parameters and medical examination values in patients with type 2 diabetes without clinical retinopathy

Correlations between vascular measurements and laboratory values

OD and ID showed no correlation with any laboratory value, including fasting glucose, HbA1c, total cholesterol, HDL cholesterol, LDL cholesterol and triglyceride levels. Meanwhile, WT showed a positive correlation with HbA1c (r=0.49, p=0.001), total cholesterol (r=0.47, p=0.002) and LDL cholesterol levels (r=0.47, p=0.003) and no significant correlation with fasting glucose, HDL cholesterol and triglyceride levels (table 3). Multivariate regression analysis with a backward stepwise method to analyse the factors associated with WT among age, diabetes duration, BMI, SBP, DBP, fasting glucose, HbA1c and total cholesterol levels revealed only HbA1c levels to be an independent factor (table 3).

Discussion

We directly visualised and evaluated the retinal arterial WT using a high-resolution retinal imaging modality known as AOSLO in patients with type 2 diabetes without clinically apparent DR and assessed its correlation with the risk factors for arteriosclerosis in this study. AOSLO enabled direct and non-invasive visualisation of the retinal arterial wall. Compared with the control group, WT in patients with diabetes was significantly larger. The microvascular retinal artery WT showed a significant positive correlation with IMT, which is the macrovascular WT. In addition, WT showed a positive correlation with HbA1c, total cholesterol and LDL cholesterol levels.

IMT33 and CAVI34 are well known for their associations with the clinical risk factors for atherosclerosis. In particular, the association between IMT and the clinical manifestations of atherosclerotic vascular diseases in patients with NIDDM is widely known.35 IMT assessments in a previous study showed that DR is a risk factor for subclinical atherosclerosis in patients with NIDDM.18 Moreover, Pannacciulli et al evaluated the carotid IMT in individuals with a genetic predisposition to type 2 diabetes, and reported a significant difference between those with a family history and those without. They suggested this might accelerate development of atherosclerosis and increase the risk of coronary heart diseases.36 A review by Ciccone et al suggested that, in addition to the increase in cardiovascular risk in individuals with a family history of diabetes, having a prediabetic condition further increases the risk of atherosclerosis development.18 Although the association between DR and IMT has been evaluated in the search for a reliable marker of early atherosclerosis, the mechanism underlying this association remains unclear. Malecki et al showed an association between DR and IMT and suggested that endothelial dysfunction is a common denominator in the pathogenesis underlying microvascular complications and atherosclerosis in patients with type 2 DM.37 Even though our patients did not have clinically apparent DR, we observed a correlation between the retinal artery WT and IMT. Although a few previous studies on the association between retinal artery WT and IMT have been reported, thickened arterial walls on AOSLO images are considered to reflect the synthesis and accumulation of basement membrane components, which result in a thickened basement membrane.7–10 To the best of our knowledge, this is the first study involving direct visualisation of the retinal artery WT and determination of the association between macrovascular and microvascular WTs.

Although the number of patients with diabetes is increasing worldwide, few studies have directly and non-invasively evaluated the retinal artery WT in these patients. Using fundus photograph-based calibre measurements, Nguyen et al17 found that an increased venous calibre, but not arteriolar calibre, was associated with increased fasting glucose and HbA1c levels in individuals with and without diabetes. Our study showed that the retinal artery ID did not correlate with HbA1c and fasting glucose levels in patients with diabetes, which is consistent with the finding of Nguyen et al.17 Meanwhile, HbA1c was positively correlated with the retinal artery WT measured by AOSLO. This positive association implied that arterial WT measurement using AOSLO might be a potentially effective method for the evaluation of microvascular damage even in the absence of clinically apparent DR, although DR stage has been a widely accepted marker of retinal vascular changes until date.

The first sign of hypertension is small artery remodelling, which is characterised by inward eutrophic and outward hypertrophic remodelling.36 In the Beaver Dam Eye study38 using fundus photograph-based calibre measurements, including the vascular inner calibre, the retinal arteriolar diameter was found to decrease with an increase in BP, while OD remained unmeasurable. We previously showed that SBP and DBP were negatively correlated with ID and positively correlated with WT using AOSLO imaging, while no significant correlation was found between SBP or DBP and OD.27 In this study, however, we found that SBP contributes to arterial narrowing against both OD and ID, while WT showed no significant change in patients with diabetes without clinically apparent DR. One possible reason for this discrepancy in results may be the inclusion of patients with a normal BP in this study, resulting in a linear decrease in OD and ID with an increase in BP and the absence of pathological wall thickening.

In general, atherosclerosis involves the accumulation of intimal smooth muscle cells together with accumulated macrophages and T lymphocytes, the formation of large amounts of connective tissue matrix by the proliferated smooth muscle cells and the accumulation of lipids.39 In this process, the oxidative modification of LDL is necessary for macrophage uptake and cellular accumulation of cholesterol. With regard to the retinal microvascular changes induced by lipid accumulation, previous studies have shown an association between increased lipid levels and diabetic complications.40 For example, one study found that patients with diabetes and elevated total cholesterol and LDL cholesterol levels were twice as likely to exhibit hard retinal exudates compared with patients with normal levels,41 while another study found a significant association between increased LDL cholesterol levels and diabetic macular oedema.42 However, the association between lipid levels and retinal vascular wall changes has not been reported. Interestingly, total cholesterol and LDL cholesterol levels showed a positive correlation with the retinal artery WT in the present study, while HDL cholesterol and triglyceride levels showed no such correlations. Despite the absence of clinically apparent DR, total cholesterol and LDL cholesterol levels were found to be associated with the retinal artery WT, which suggests that direct evaluation of the retinal vascular wall can facilitate the monitoring of initial microvascular changes associated with hyperlipidaemia.

This study has several limitations. First, we included few patients aged >70 years, because cataract degrades AOSLO image quality. Second, there could have been systematic exclusion of eyes with thicker vessel walls or patients with poor blood sugar control that may have had more media opacity; however, we only selected eyes with clear images. Third, patients in this study did not have diabetes for a long period and the effects might have been greater if individuals with a longer duration of diabetes were studied. Last, DM often coexists with other risk factors for atherosclerosis, such as hypertension or dyslipidemia; therefore, our results should be carefully interpreted. Indeed, in this study, multiple associations were found between vascular measurements and some laboratory values, although multivariate regression analysis revealed only HbA1c levels to be an independent factor associated with the retinal artery WT.

In conclusion, we used AOSLO imaging for direct arterial wall visualisation in patients with diabetes without clinically apparent DR. We found that the retinal artery WT (microvascular thickness) was larger than the WT in the control group and was positively correlated with HbA1c, total cholesterol, LDL cholesterol levels and IMT (macrovascular WT). These results suggest that retinal artery wall measurements can be used as potential surrogate markers of early diabetic microangiopathy.

References

Footnotes

  • Contributors Conception and design: SA and AU. Data collection: SA, AU, TM and NG. Analysis and interpretation: SA and AU. Writing the article: SA and UA. Preparation, review and approval of the manuscript: SA, AU, TM, KS, NG and NY.

  • Funding This work was partly supported by the Innovative Techno-Hub for Integrated Medical Bio-Imaging as part of the Project for Developing Innovation Systems, by the Ministry of Education, Culture, Sports, Science and Technology in Japan.

  • Competing interests NY: Topcon (financial support, Tokyo, Japan), Nidek (financial support, consultant, Aichi, Japan), Canon (financial support, Tokyo, Japan).

  • Patient consent Obtained.

  • Ethics approval The Institutional Review Board and Ethics Committee at Kyoto University Graduate School of Medicine.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost Jonas