Computer-assisted retinal vessel measurement in an older population: correlation between right and left eyes

BMI and retinal vascular caliber in children

OBJECTIVE

In adult populations, changes in retinal vascular caliber have been linked with obesity and metabolic syndrome. We examined the association of BMI and weight with retinal vascular caliber in children.

RESEARCH METHODS AND PROCEDURES

This was a school-based, cross-sectional study of 768 children, 7 to 9 years old, randomly sampled from the Singapore Cohort Study of the Risk Factors for Myopia. Participants had digital retinal photographs. Retinal vascular caliber was measured using a computer-based program and combined to provide average calibers of arterioles and venules in that eye. Weight and height were measured using standardized protocol. These data were used to calculate BMI.

RESULTS

In this population, the mean retinal arteriolar and venular calibers were 156.40 microm [95% confidence interval (CI), 155.44 to 157.36] and 225.43 microm (95% CI, 224.10 to 226.74) respectively. After controlling for age, gender, race, parental monthly income, axial length, birth weight, and birth length, each 3.1 kg/m2 (standard deviation) increase in BMI was associated with a 2.55-microm (95% CI, 1.21 to 3.89; p < 0.001) larger retinal venular caliber. In multivariable analysis, greater weight was also significantly associated with larger retinal venular caliber. BMI and weight were not associated with retinal arteriolar caliber. Height was not significantly associated with retinal arteriolar or venular caliber.

DISCUSSION

Greater BMI and weight are associated with larger retinal venular caliber in healthy children.

Rationale and design of the AdRem study: Evaluating the effects of blood pressure lowering and intensive glucose control on vascular retinal disorders in patients with type 2 diabetes mellitus

The ADVANCE Retinal Measurements (AdRem) Study is a large intervention study evaluating the effects of target driven intensive glucose control and placebo controlled blood pressure lowering on retinal vascular changes. AdRem is a sub-study of the ADVANCE Study (Action in Diabetes and Vascular disease), a 2x2 factorial randomized controlled trial with an ACE inhibitor-diuretic combination (perindopril-indapamide) and a gliclazide MR-based regimen in patients with type 2 diabetes mellitus. The AdRem study is based on seven-field stereoscopic retinal photographs of both eyes. These are taken within 3 months after randomization in ADVANCE (baseline), at the biennial and at the final visit. The primary outcome is progression of two or more steps in ETDRS classification. Secondary outcomes include progression of retinal vascular lesions and distortion of retinal vascular geometry. Retinal photographs are made on film and digitized at a central laboratory. The AdRem study uses fully digitized quality control and grading. Between August 2002 and January 2004 1978 patients were included in the AdRem study, from 39 centers in 14 countries. Approximately 85% comply with the strict AdRem quality requirements. Publication of the results is expected in early 2008. The AdRem study is designed to provide reliable evidence on the effects of intensive glucose control and blood pressure lowering on both diabetic retinopathy and abnormalities of retinal vasculature in patients with type 2 diabetes mellitus.

Cardiovascular risk factors and retinal microvascular signs in an adult Japanese population: the Funagata Study

OBJECTIVE

To describe the prevalence of retinal vascular signs and their association with cardiovascular risk factors in a Japanese population.

DESIGN

Population-based cross-sectional study.

PARTICIPANTS

Adult persons aged 35 years or older from Funagata, Yamagata Prefecture, Japan (n = 1481).

METHODS

The Funagata Study is a Japanese population-based study of persons aged 35 years or older, and included 1961 nondiabetic participants (53.3% of 3676 eligible subjects). A nonmydriatic retinal photograph was taken of 1 eye to assess retinal microvascular signs. Retinal arteriolar wall signs (focal arteriolar narrowing, arteriovenous nicking, enhanced arteriolar wall reflex) and retinopathy were assessed in 1481 participants without diabetes (40.3% of eligible persons) using a standardized protocol. Using a computer-assisted method, retinal vessel diameters were measured in 921 participants with gradable retinal image (25.1% of eligible persons).

MAIN OUTCOME MEASURES

Prevalence of retinal microvascular signs and their association with cardiovascular risk factors.

RESULTS

Moderate or severe focal arteriolar narrowing, arteriovenous nicking, enhanced arteriolar wall reflex, and retinopathy were found in 8.3%, 15.2%, 18.7%, and 9.0%, respectively, of the study population. Mean (+/-standard error) values for retinal arteriolar diameter were 178.6+/-21.0 mum, and mean values (+/-standard error) for venular diameter were 214.9+/-20.6 mum. Older persons were more likely to have retinal arteriolar wall signs, retinopathy, and narrower retinal vessel diameters. After adjusting for multiple factors, each 10-mmHg increase in mean arterial blood pressure was associated with a 20% to 40% increased likelihood of retinal arteriolar signs and a 2.8-mum reduction in arteriolar diameter. Retinopathy was associated with higher body mass index and both impaired glucose tolerance and impaired fasting glucose.

CONCLUSIONS

In nondiabetic Japanese adults, retinal arteriolar wall signs were associated with older age and increased blood pressure, whereas retinopathy was associated with older age, higher body mass index, impaired glucose tolerance, and impaired fasting glucose. These findings are comparable with data from white populations.

Dyslipidaemia and microvascular disease in the retina

PURPOSE

There are few data on the effect of serum lipids on microvascular disease. This study assessed the relationships between serum lipid levels and microvascular disease, as seen in the retina, among participants who attended a population-based study in Australia (n=3654, aged 49+years).

METHODS

Diameters of retinal arterioles and venules were measured from digitised photographs of each participant to obtain an estimate of generalised arteriolar narrowing. Focal arteriolar narrowing, arteriovenous nicking, and retinopathy lesions (microaneurysms, haemorrhages, hard/soft exudates) were graded using a standard protocol. Fasting blood tests were performed in 89% of subjects. Adjusted means were calculated using general linear models. Logistic regression models were used to determine the odds ratios for retinal microvascular signs.

RESULTS

After controlling for age, sex, body mass index, smoking, and mean arterial blood pressure, elevated high-density lipoprotein cholesterol was associated with narrower retinal arterioles (Ptrend=0.002) and venules (Ptrend=0.03) and with increased odds of generalised arteriolar narrowing (odds ratio 1.6, 95% confidence interval 1.1-2.2 for the highest vs the lowest quintile of high-density lipoprotein cholesterol). Serum triglyceride had a U-shaped relationship with venular diameter (Ptrend=0.003). We found no consistent pattern of association between serum total cholesterol or low-density lipoprotein cholesterol and any retinal microvascular signs.

CONCLUSIONS

These findings suggest that microvascular disease in the retina may result from processes distinct from dyslipidaemia.

Five-year incidence and progression of vascular retinopathy in persons without diabetes: the Blue Mountains Eye Study

PURPOSE

To assess the 5-year incidence of vascular retinopathy and its associations in an older nondiabetic population.

METHODS

The Blue Mountains Eye Study examined 3654 residents aged 49+ years (82.4% response rate) during 1992-1994, and re-examined 2335 (75.1% of survivors) during 1997-1999. Retinopathy lesions (microaneurysms, haemorrhages, hard or soft exudates) were assessed from 6-field retinal photographs in persons without diabetes. Incident retinopathy was assessed in those at risk. Hypertensive status was defined following the WHO/International Society of Hypertension guidelines.

RESULTS

Of the 2335 re-examined, 195 had retinopathy lesions at baseline and 1725 were at risk of retinopathy after excluding subjects with diabetes (n=261), retinal vein occlusion (n=52) or missing/un-gradable photographs (n=102). The cumulative 5-year incidence was 9.7% (95% confidence intervals (CI) 8.3-11.1%). Age was the only factor significantly associated with incident retinopathy (Pfor trend=0.012). Neither fasting blood glucose (age-sex-adjusted P=0.147) nor hypertension (adjusted Pfor trend=0.43) was associated with incident retinopathy. Of the 195 with retinopathy lesions at baseline, 3.5% developed diabetes, 13.3% progressed, and 72.3% regressed/disappeared over 5 years. Progression was positively associated with elevated blood pressure (BP) (adjusted odds ratio (OR) 1.3, 95% CI 1.1-1.6 per 10 mmHg systolic BP) and inversely associated with fasting glucose level (OR 0.36, CI 0.14-0.92 per mmol/l increase). Aspirin use was weakly associated with regression (OR 2.4, CI 1.0-6.0).

CONCLUSIONS

Over 5 years, retinopathy developed in 10% of older people without diabetes, while 72% of baseline lesions regressed. Age was significantly associated with the development of these lesions.

Retinal image analysis: concepts, applications and potential

As digital imaging and computing power increasingly develop, so too does the potential to use these technologies in ophthalmology. Image processing, analysis and computer vision techniques are increasing in prominence in all fields of medical science, and are especially pertinent to modern ophthalmology, as it is heavily dependent on visually oriented signs. The retinal microvasculature is unique in that it is the only part of the human circulation that can be directly visualised non-invasively in vivo, readily photographed and subject to digital image analysis. Exciting developments in image processing relevant to ophthalmology over the past 15 years includes the progress being made towards developing automated diagnostic systems for conditions, such as diabetic retinopathy, age-related macular degeneration and retinopathy of prematurity. These diagnostic systems offer the potential to be used in large-scale screening programs, with the potential for significant resource savings, as well as being free from observer bias and fatigue. In addition, quantitative measurements of retinal vascular topography using digital image analysis from retinal photography have been used as research tools to better understand the relationship between the retinal microvasculature and cardiovascular disease. Furthermore, advances in electronic media transmission increase the relevance of using image processing in 'teleophthalmology' as an aid in clinical decision-making, with particular relevance to large rural-based communities. In this review, we outline the principles upon which retinal digital image analysis is based. We discuss current techniques used to automatically detect landmark features of the fundus, such as the optic disc, fovea and blood vessels. We review the use of image analysis in the automated diagnosis of pathology (with particular reference to diabetic retinopathy). We also review its role in defining and performing quantitative measurements of vascular topography, how these entities are based on 'optimisation' principles and how they have helped to describe the relationship between systemic cardiovascular disease and retinal vascular changes. We also review the potential future use of fundal image analysis in telemedicine.

Systemic associations of retinal microvascular signs: a review of recent population-based studies

Retinal microvascular signs, such as generalized retinal arteriolar narrowing, focal arteriolar narrowing, arteriovenous nicking and retinal haemorrhages, microaneurysms and cotton wool spots, are common fundus findings in the general population, even in individuals without hypertension or diabetes. Recent population-based studies have provided new insights into the systemic associations and clinical significance of these retinal signs. Studies show that these retinal microvascular signs are strongly associated with elevated blood pressure (BP). Generalized retinal arteriolar narrowing may be associated with markers of inflammation and risk of diabetes and hypertension. Retinal haemorrhages, microaneurysms and cotton wool spots are associated with risk of subclinical and clinical stroke, cognitive impairment, renal dysfunction and cardiovascular mortality, independent of BP and cardiovascular risk factors. A consistent pattern of association between retinal microvascular signs and ischaemic heart disease has not been demonstrated. This suggests that patients with some retinopathy signs (retinal haemorrhages, microaneurysms and cotton wool spots) may benefit from a careful systemic evaluation and, if supported by further research, appropriate risk reduction therapy.

Retinal vessel diameter and open-angle glaucoma: the Blue Mountains Eye Study

PURPOSE

To examine the relationship between open-angle glaucoma (OAG) and retinal vessel diameter among baseline participants in the Blue Mountains Eye Study.

DESIGN

Population-based cross-sectional study.

PARTICIPANTS

The study included 3654 persons older than 49 years, representing 82.4% of permanent residents living in an area west of Sydney.

METHODS

Participants had a detailed eye examination, including automated perimetry and stereo optic disc photography. A computer-assisted program measured retinal vessel diameters from digitized photographs of right eyes.

MAIN OUTCOME MEASURES

Open-angle glaucoma was diagnosed from matching visual field defects and optic disc cupping, without reference to intraocular pressure (IOP) level. Ocular hypertension was defined as IOP of >21 mmHg in either eye, without matching glaucomatous optic disc and field changes. Average retinal vessel diameters, measured from right eyes, were summarized as arteriolar and venular equivalents. The lowest quintile of the arteriolar equivalent or arteriole-to-venule ratio was used to define generalized retinal arteriolar narrowing.

RESULTS

The study included 3314 participants, after excluding those with incomplete data or nonglaucomatous optic nerve disease. Of persons included, 59 (1.8%) had evidence of glaucomatous damage affecting the right eye, 3065 (92.5%) had no damage to either eye, and 163 (4.9%) had ocular hypertension. Right eyes with glaucomatous damage had significantly narrower retinal arteriolar diameters (183+/-2.6 microm) than eyes without glaucoma (194+/-0.4 microm, P = 0.0001) or eyes with ocular hypertension (195+/-1.6 microm, P = 0.0002), after adjusting for age, mean arterial blood pressure, and other confounding variables, including refraction. Right eyes with glaucomatous damage were at least 2 times more likely to have generalized retinal arteriolar narrowing than eyes without glaucoma (odds ratio, 2.7; 95% confidence interval, 1.5-4.8).

CONCLUSIONS

These population-based data suggest that generalized retinal arteriolar narrowing, an indicator of localized vascular change, is significantly associated with optic nerve damage caused by OAG. It is not clear whether such a retinal arteriolar change reflects an ischemic process leading to optic nerve damage or results from loss of retinal neurons secondary to glaucoma.

Does hormone replacement therapy influence retinal microvascular caliber?

Previous population-based data suggest that retinal arteriolar diameter is wider in women than in men. Estrogen exposures could account for this difference. To evaluate the effects of HRT on small blood vessels, we assessed the relationship between use of hormone replacement therapy (HRT) and retinal microvascular diameter in older women in the Blue Mountains Eye Study baseline population (n = 1993, age >/= 49 years). Information on HRT use was recorded by trained interviewers. A computer-assisted program measured retinal vessel diameters from digitized photographs. Average arteriolar diameter was calculated as the central retinal arteriolar equivalent (CRAE); the lowest quintile of CRAE was considered generalized arteriolar narrowing. Of the 1897 women with complete data, 79 (4%) were premenopausal and 315 (17%), 224 (12%), and 1279 (67%) were current, past, and never HRT users, respectively. Among women aged < 65 years, multivariate-adjusted mean CRAE for the four groups was 196, 198, 201, and 200 microm (P < 0.0001), respectively. Among postmenopausal women >/= 65 years, multivariate-adjusted mean CRAE was 187, 188, 191 microm (P < 0.0001), respectively, for current, past, and never users. Current HRT users were 40% more likely than never users to have generalized arteriolar narrowing (95% confidence interval 1.0-2.1). It appears that the use of HRT in older women may not have any long-term vasodilatory effects on retinal arterioles. These data do not support our hypothesis that exogenous estrogen exposures account for observed wider retinal arteriolar diameters in women.

Impact of current and past blood pressure on retinal arteriolar diameter in an older population

OBJECTIVE

To examine the effects of current and past blood pressure on retinal arteriolar diameter in a general older population.

DESIGN

Cross-sectional and longitudinal studies.

SETTING

Population-based cohort study of older residents from an area west of Sydney, Australia.

PARTICIPANTS

Two thousand three hundred and thirty-five individuals (n = 2335) (aged > or = 54 years) who attended the 5-year follow-up Blue Mountains Eye Study during 1997-99.

MAIN OUTCOME MEASURES

A computer-assisted method measured vessel diameters from digitized right eye retinal photographs. The narrowest quintile of central retinal arteriolar equivalent or arteriole-to-venule ratio (AVR) defined generalized arteriolar narrowing. Blood pressure was measured using a mercury sphygmomanometer.

RESULTS

After simultaneous adjustment for age, sex, body mass index, smoking and current or past blood pressure, elevated levels of both current and past blood pressure were associated with narrower retinal arterioles [Ptrend = 0.009 and 0.007 for current and past systolic blood pressure (SBP), respectively] and lower AVR [Ptrend = 0.001 and 0.0009 for current and past diastolic blood pressure (DBP), respectively]. Generalized arteriolar narrowing was associated with both current blood pressure [adjusted odds ratio (OR) 2.4, 95% confidence interval (CI) 1.5-3.8 for the highest versus lowest quintile of current DBP] and past blood pressure (adjusted OR 1.7, 95% CI 1.1-2.6, for the highest versus lowest quintile of past DBP). Hypertension duration or control status at baseline had no additional effect on arteriolar diameter after adjusting for current blood pressure.

CONCLUSIONS

Our data document the independent effects of both current and past blood pressure on small vessel calibre in the retina, suggesting that retinal arteriolar narrowing may result from the cumulative effects of long-standing hypertension.

Retinal vessel wall signs and the 5 year incidence of age related maculopathy: the Blue Mountains Eye Study

AIMS

To assess whether retinal arteriolar wall changes (focal narrowing and arteriovenous nicking) are associated with an increased 5 year risk of age related maculopathy (ARM).

METHODS

The Blue Mountains Eye Study examined 3654 residents aged 49+ years living in a defined area, during 1992-4 (82.4% participation). After 5 years, 2335 surviving participants (75.1%) were re-examined during 1997-9. Retinal photographs were graded using the Wisconsin ARM grading system. Incident late (neovascular or atrophic) or early stage ARM was defined using a side by side grading method. Focal arteriolar narrowing (localised constricted arteriolar segments causing a sausage-like appearance), and arteriovenous (AV) nicking (constriction on both sides of the venule where crossed by an arteriole), were graded by comparison with standard photographs. All retinal vessels passing through a circumferential zone 0.5-1.0 disc diameters from the optic disc margin were measured from digitised images. Summarised estimates for central retinal arteriolar equivalent (CRAE) represent an average diameter of arterioles for that eye. Associations were assessed after adjusting for age, sex, smoking, mean arterial blood pressure, and other vascular risk factors.

RESULTS

Of 2314 baseline participants at risk of late stage ARM, either late stage lesion developed in 34 participants (1.5%). Of 2203 at risk of early stage ARM, this sign developed in 197 participants (8.9%). Focal arteriolar narrowing was present at baseline in at least one eye of 162 survivors (6.9%) and severe AV nicking was present in 187 people (8.1%). Over 5 years, 4.9% of subjects with and 1.2% of those without focal narrowing developed either late stage ARM lesion, age adjusted relative risk (RR) 2.3, 95% confidence interval (CI) 1.0 to 5.1, multivariate adjusted odds ratios (OR) 2.1 (95% CI 0.9 to 4.9). Similarly, 3.7% of subjects with and 1.3% of those without severe AV nicking developed late ARM lesions, age adjusted RR 2.1 (95% CI 0.9 to 5.1), multivariate adjusted OR 2.2 (95% CI 0.9 to 5.6). Corresponding age adjusted RR and multivariate adjusted OR for development of early stage ARM were 1.4 (95% CI 0.9 to 2.0) and 1.3 (95% CI 0.8 to 2.1) for focal arteriolar narrowing, and 1.6 (95% CI 1.0 to 2.3) and 1.8 (95% CI 1.1 to 2.9) for severe AV nicking, respectively. No associations between baseline CRAE and 5 year incident late or early stage ARM were found.

CONCLUSIONS

Although of borderline statistical significance, the consistent associations found in this study suggest that structural retinal arteriolar changes may either contribute to ARM progression or may share common pathological pathways with ARM.

Hypertensive retinal vessel wall signs in a general older population: the Blue Mountains Eye Study

To describe cross-sectional relations between hypertension and retinal vessel wall signs in an older white population. These signs were defined from fundus photographs in 3654 Blue Mountains Eye Study participants > or =49 years of age. Focal arteriolar narrowing and arteriovenous nicking were graded through the use of standard protocol. Photographs were digitized to measure retinal vessel diameters. Average arteriolar diameter, summarized as central retinal arteriolar equivalent and arteriole-to-venule ratio, were used as indexes of generalized arteriolar narrowing. Blood pressure was measured with the use of a mercury sphygmomanometer. Hypertension was defined through the use of antihypertensive medications, systolic blood pressure > or =160 mm Hg, or diastolic blood pressure > or =95 mm Hg. Hypertension was categorized as controlled (using medication, normal blood pressure), uncontrolled (using medication, high blood pressure), or untreated (not using medication). Hypertensive subjects had higher prevalence of all retinal microvascular signs. After adjusting for age, gender, body mass index, and smoking, persons with controlled (18.2%), uncontrolled (13.8%), or untreated hypertension (13.8%) were significantly more likely than normotensive subjects (54.2%) to have (a) lower central retinal arteriolar equivalent: adjusted odds ratios 1.5, (95% CI, 1.1 to 1.9), 2.1 (1.6 to 2.7), and 2.1 (1.6 to 2.7), respectively, and lower arteriole-to-venule ratio: 1.3 (1.0 to 1.6), 1.4 (1.1 to 1.8), and 1.7 (1.3 to 2.2), respectively; (b) focal arteriolar narrowing: 1.3 (0.9 to 1.9), 2.2 (1.5 to 3.2), and 2.5 (1.8 to 3.6), respectively; and (c) arteriovenous nicking: 1.3 (0.9 to 1.8), 2.3 (1.6 to 3.2), and 1.9 (1.3 to 2.7), respectively. Our findings demonstrate a strong relation between presence and severity of hypertension and retinal microvascular structural changes.

Correlation between generalized and focal retinal arteriolar narrowing in an older population

This study aimed to assess the associations between generalized and focal retinal arteriolar narrowing in the Blue Mountains Eye Study population (n = 3654 persons aged 49+ years). Summarized estimates of the average retinal arteriolar diameter, the central retinal arteriolar equivalent (CRAE) and arteriole-to-venule ratio were obtained from right eye photographs of each participant using a computer-assisted method. Focal retinal arteriolar narrowing was graded from retinal photographs of both eyes using a 'light-box' method. The proportion of subjects with focal arteriolar narrowing increased as CRAE decreased from the widest (2.2%) to the narrowest (7.8%) quintile (P for trend = 0.001). A receiver operating characteristic curve was constructed using CRAE to detect focal arteriolar narrowing. With the diagnostic cut point set at the 4th decile of CRAE, the sensitivity and specificity in detecting focal arteriolar narrowing were 64% and 61%,respectively. These data suggest a moderate correlation between generalized (CRAE) and focal arteriolar narrowing. However, no significant association was found between arteriole-to-venule ratio and focal arteriolar narrowing.