Printer Friendly

Retinal vessel evaluation.

This article considers the importance of retinal vessel assessment with an emphasis on arterio-venous ratio measurement and its relationship to systemic vascular disease. The author explores issues surrounding variable subjective interpretation of vessel ratios and details an alternative approach to assist with these measures using retinal photography.

Course code: C-36462 | Deadline: June 20, 2014

Acne. Learning objectives

To be able to recognise arterio-venous changes by comparing measures against existing records and manage the patient accordingly (Group 2.2.5)

To be able to recognise the relationship between arterio-venous changes and systemic vascular disease (Group 6.1.13)

Learning objectives

To be able to understand the relationship between arterio-venous changes and systemic vascular disease (Group 8.1.4) ocular


As part of any retinal assessment, whether undertaken by direct ophthalmoscopy, slit lamp binocular indirect ophthalmoscopy or using a fundus camera, optometrists always evaluate the state of the retinal vasculature. Besides looking for signs of haemorrhages, cotton wool spots and pigmentary changes, the retinal arteries and veins are checked for tortuosity, nipping, focal narrowing, neovascularisation, obstruction and relative size. There are numerous case reports, clinical guidelines and reports on fundus evaluation. However, limited information is available for evaluating retinal vessel measurements, and their application for risk stratification and monitoring.

Arterio-venous assessment

A wide range of patients are encountered in routine clinical practice, many of whom suffer from underlying vascular diseases such as hypertension, heart disease and diabetes, although not every patient presents with visible adverse retinal manifestations.


Despite the widespread use of digital imaging technology, practitioners still rely upon subjective methods for assessing and recording retinal vessel characteristics, including the vessel reflex and the arteriovenous ratio (AVR). Subjective visual grading requires experience acquired from examining large numbers of patients at regular intervals, paired with excellent differential diagnostic skills. If the same examiner is assigned to the same patient for all follow-up measurements inter-observer variability is limited, but this is not always possible in everyday clinical practice. A further limitation of subjective evaluation of retinal vessels is that subtle signs can be overlooked as visual grading is often too coarse and not sensitive enough to detect small changes over time, particularly the assessment of arteriolar narrowing. (1,2) Semi-automated measurements of retinal vessel diameters offer a much more reliable method to achieve this goal. Current literature on the relationship between retinal vessel diameter changes and its potential for risk assessment and links to underlying vascular disease strongly supports the objective measurement of retinal vessel diameters. (3-5)

Subjective grading of AVR is a poor clinical tool, with variability in respect to the best approach for establishing clinically useful measures. (6-8) While most literature states that a 'normal' AVR is reflected by 2:3 or 3:4, others state that the AVR can lie between 1:3 to 1:1, with little or no guidance on incremental steps. This leads to most examiners applying an almost binary approach to grading AVR by stating 2:3 in most cases, whereas in cases of vessels of the same size as 1:1 and retinal arterioles classed as generalised narrow, referred to as 1:2.

Clinical experience also plays a part in reliable subjective AVR grading. While an inexperienced grader will grade with much larger increments, a more experienced grader will not. But how much is the difference in reality? An interesting evaluation of this was undertaken at Aston University by asking a final year optometry student and two qualified optometrists, with four and 20 years experience respectively, to grade the same set of retinal images of 70 healthy individuals and comparing them to semi-automated AVR results. While there was no difference on the average AVR grading between the student and the more recently qualified optometrist 0.60 (SD 0.08) and 0.61 (SD 0.10), there was a significant difference to the optometrist with 20 years of experience 0.71 (SD 0.06). In addition, all subjective values, regardless of grader experience, were significantly lower than the values obtained using a semiautomated methodology (see Figure 1).

AVR calculation

Historically, the AVR was thought to be a good measure to describe retinal vessel geometry being proposed by Stokoe and Turner in 1966. (9) In 1974, Parr and Spears devised a formula to calculate the central retinal arterial equivalent (CRAE). (10, 11) This method takes into account all arteriolar trunk and branch vessels around the optic nerve head (ONH) in a pre-defined concentric measurement zone.

Each vessel diameter is measured and paired vessels are combined to estimate the trunk vessel before the paired trunk vessels are combined into a summary measure to provide a value for CRAE using the following formula:

[W.sub.c] = v(0.87[W.sub.a.sup.2] + 1.01[W.sub.b.sup.2]-0.22[W.sub.a][W.sub.b]-10.76)

where [W.sub.c] is the calibre of the trunk arteriole, [W.sub.a] is the calibre of the smaller branch arteriole and [W.sub.b] is the calibre of the larger branch arteriole.

However, this approach is time consuming and highly dependent on the tracing of individual paired vessels. Later, Hubbard and colleagues (1992) devised a similar measure to derive the central retinal venous equivalent (CRVE):

[W.sub.c] = v(0.72[W.sub.a.sup.2] +0.91[W.sub.b.sup.2] + 450.05)

where [W.sub.c] is the calibre of the trunk venule, [W.sub.a] is the calibre of the smaller branch venule and [W.sub.b] is the calibre of the larger branch venule.

Following further developments in 1999 it was possible to calculate CRAE and CRVE by measuring all vessels crossing through a concentric annulus around the ONH (see Figure 2). (12) The measurement annulus is typically 0.5 disc diameters (DD) in width and 0.5DD distant from the ONH rim. This technique is less time consuming and offers a quick and reliable calculation of AVR. This current method has undergone further changes; those vessels with a calibre of less than 25 microns are excluded and if they are larger than 80 microns the vessel branches are taken into account rather than the trunk. The Atherosclerosis and Risks in Communities (ARIC) study was the first to utilise this objective approach to measure AVR semiautomatically in a large patient cohort. As the measurements were dependent upon the number of vessels selected, Knudtson and colleagues (13) developed a revised formula based on the six largest arterioles and venules passing through the measurement zone (see Figure 2). This revised formula is in agreement with the previous calculation devised by Parr-Hubbard, but independent of the number of vessels measured.


Value and interpretation of AVR

Originally, it was thought appropriate to quantify generalised arteriolar attenuation due to evidence that arterioles would be more affected by narrowing in response to cardiovascular disease processes than venules. (12,14) Subsequently, this approach was used in a number of large community-based epidemiology studies such as ARIC, Blue Mountain Eye Study (BMES), Wisconsin Epidemiologic Study of Diabetic Retinopathy, Cardiovascular Healthy Study, Beaver Dam Eye Study (BDES) and the Rotterdam Study. Results from these studies show agreement between right and left eyes and AVR's potential to reflect generalised arteriolar narrowing. However, owing to its use in these large cohorts with patients comprising a wide range of risks and underlying pathology (such as hypertension, heart disease and diabetes mellitus) it became evident that AVR was not an ideal marker for risk prediction of most cardiovascular pathologies, including coronary heart disease, stroke, myocardial infarction and the presence of carotid disease. (15,16) The major limitation of AVR is that arterioles and venules have a variable response to different pathologies, that is the increase/decrease of AVR can occur due to arteriolar and/or venular changes or simultaneous changes. For example, arterioles can be attenuated with hypertension, whereas venules get dilated due to inflammatory processes. Therefore, the independent use of CRAE and CRVE provides superior information into certain pathological states compared with using AVR alone.

Measuring CRAE, CRVE and AVR from fundus images

One basic requirement to measure retinal vessel diameters objectively is to obtain a fundus image, ideally with the ONH centred. The preferred way to measure vessel diameters is from monochromatic (red-free) retinal images as these provide best contrast, but colour images or those converted to monochrome by selecting one colour channel post image acquisition can also be used. Depending on the camera and software used, some manufacturers already provide software to calculate CRAE, CRVE and AVR using semiautomated methods. Alternatively, there is a range of software available which can be used independent of the camera system used to obtain the image, for example VesselMap (IMEDO Systems, Germany), IVAN (IVAN, Department of Ophthamology and Visual Science, University of Wisconsin-Madison, US) and SIVA (Singapore I Vessel Assessment, Singapore Eye Research Centre, Singapore) which can be licensed individually, but is also integrated within Topcon's ImageNet software. Besides these commercially available programmes ImageJ is a free software package with a wide range of measurement tools suitable for the analyses of retinal images.

Once the image is obtained, retinal vessel diameters can be measured in the concentric annulus around the ONH and all parameters (CRAE, CRVE and AVR) calculated according to the formulae outlined earlier. Depending on the software used, the examiner is only required to select the relevant vessels and all calculations are completed automatically; some programmes provide an output of individual vessel diameters in addition to tools that enable monitoring of progression.


The benefit of objective measurements

Figure 3 shows a series of different fundi along with their semi-automatically derived values for CRAE, CRVE and AVR to illustrate the subtle differences that can be overlooked when using a subjective approach.

The main benefit of objectively measuring AVR is that it provides highly reproducible results along with the summarised arteriolar and venular diameters. (17) Objective AVR calculation provides values of CRAE and CRVE, which cannot be obtained by subjective grading, and are superior to AVR itself in regards to risk prediction and monitoring microvascular change over time.

Value of AVR, CRAE and CRVE measures

Publications on the utility, clinical validity and applications for retinal vessel diameters are numerous but to date not widely integrated into clinical practice. Liew and colleagues reported from the ARIC study that the major systemic determinant for smaller CRAE is higher blood pressure whereas wider CRVE is mainly due to current cigarette smoking, higher blood pressure, systemic inflammation and obesity. Those with higher blood pressure (75th percentile) had on average 4.8 microns smaller CRAE and 2.6 microns wider CRVE than those with lower blood pressure (25th percentile). (18) A more recent study found a strong negative correlation between renal function and retinal parameters (CRAE and CRVE) in a cohort of eighty healthy individuals which suggests a common determinant in pre-clinical target organ damage. (19) This is in support of earlier studies, (20,21) examining the association between retinal vascular signs and incident hypertension providing evidence that a decrease in CRAE is indeed an antecedent to clinical onset of hypertension and occurs prior to other signs of target organ damage.

Besides the value of CRAE in predicting hypertension, it also shows great potential in other pathologies including stroke and diabetes. Generalised arteriolar narrowing as reflected by a decrease in CRAE is associated with an increased risk of stroke with odds ratios reported between 1.1 and 3.0. (15,22,23) While in diabetes, an increase of CRVE is associated with increased incidence of diabetic retinopathy (DR), progression of DR, progression to proliferative DR and macular oedema, but is unrelated to CRAE. (24)

Apart from its potential for risk prediction, screening and monitoring systemic pathologies, retinal vessel parameters have been shown to be useful in ocular-vascular abnormalities such as glaucoma and AMD too. Results of the Handan Eye Study showed the association of increased CRAE with early AMD (average odds ratio 1.34). (25) While an increased risk of open angle glaucoma was associated with a decrease in CRAE (average odds ratio 1.77). (26)


The retina is an ideal location to observe vascular changes non-invasively. Although optometrists thoroughly assess its structure, vasculature and overall appearance, this is mostly done by subjective visual grading despite the widespread use of fundus cameras. A steep increase of patients at risk, or suffering from cardiovascular pathology such as hypertension, diabetes and heart disease paired with an ageing population brings about an increased necessity for screening and monitoring. Optometrists already play an integral part in the screening of DR and with the strong evidence of retinal vessel parameters associated with systemic and ocular pathology, these markers might provide enhanced diagnostic insight for existing pathology and for those at risk of developing future ocular-vascular pathology. While retinal vascular changes themselves do not always lead to imminent loss in visual function, they are useful markers for future risk. Any assessment a patient has to undergo should be of clinical use; AVR as determined by subjective grading is an outdated measure, superseded by objectively determined diameter measurements. Routine retinal photography as part of a standard eye examination is becoming commonplace with imaging technology now an affordable tool for everyday clinical practice.

Static retinal vessel analyses are useful clinical measures to assess future risk of systemic and ocular vascular complications. This methodology can be carried out on most retinal photographs, including 'older' digitised images and, unlike with subjective AVR, we can now determine whether the arterial, the venular or both vessel types have changed over time which in turn provides more accurate prognostics on the underlying pathology.

Liew and colleagues also highlight that more objective methods of static retinal vessel assessment need to be implemented in clinical practice to build upon the existing clinical evidence base. Optometrists are in a prime position to contribute to this by implementing retinal vessel analysis as part of routine fundus photography.

Reflective learning

Having completed this CET exam, consider whether you feel more confident in your clinical skills--how will you change the way you practice? How will you use this information to improve your work for patient benefit?

Exam questions

Under the enhanced CET rules of the GOC, MCQs for this exam appear online at Please complete online by midnight on June 20, 2014. You will be unable to submit exams after this date. Answers will be published on and CET points will be uploaded to the GOC every two weeks. You will then need to log into your CET portfolio by clicking on 'MyGOC' on the GOC website ( to confirm your points.


Visit clinical, click on the article title and then on 'references' to download.

Dr Rebekka Heitmar is a lecturer at Aston University. Her research interests include ocular haemodynamics and the impact of systemic vascular disease upon ocular circulation and visual function.
COPYRIGHT 2014 Ten Alps Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2014 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:1 CET POINT
Author:Heitmar, Rebekka
Publication:Optometry Today
Geographic Code:4EUUK
Date:May 23, 2014
Previous Article:Sports vision--prescribing and dispensing.
Next Article:Diary dates.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters