Skip main navigation

Retinal Microvascular Changes and Risk of Stroke

The Singapore Malay Eye Study
Originally publishedhttps://doi.org/10.1161/STROKEAHA.113.001738Stroke. 2013;44:2402–2408

Abstract

Background and Purpose—

To examine the relationship between retinal microvascular measures and incident stroke in an Asian Malay population.

Methods—

We conducted a prospective, population-based cohort study of Asian Malay persons 40 to 80 years at baseline. Retinal microvascular signs were assessed from baseline retinal photographs including quantitative retinal microvascular parameters (caliber, branching angle, tortuosity, and fractal dimension) and qualitative retinopathy signs. Incident stroke cases were identified during the follow-up period. Cox proportional-hazards regression and incremental usefulness analysis (calibration, discrimination, and reclassification) were performed.

Results—

A total of 3189 participants were free of prevalent stroke at baseline. During the follow-up (median, 4.41 years), 51 (1.93%) participants had an incident stroke event. In Cox proportional-hazards models adjusting for established stroke predictors (age, sex, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, smoking, glycosylated hemoglobin, and antihypertensive medication), retinopathy (hazard ratio, 1.94; 95% confidence interval, 1.01–3.72) and larger retinal venular caliber (hazard ratio, 3.28; 95% confidence interval, 1.30–8.26, comparing fourth versus first quartiles) were associated with risk of stroke. Compared with the model with only established risk factors, the addition of retinal measures improved the prediction of stroke (C-Statistic 0.826 versus 0.792; P=0.017) and correctly reclassified 5.9% of participants with incident stroke and 3.4% of participants with no incident stroke.

Conclusions—

Retinal microvascular changes are related to an increased risk of stroke in Asian Malay, consistent with data from white populations. Retinal imaging improves the discrimination and stratification of stroke risk beyond that of established risk factors by a significant but small margin.

Introduction

There is increasing evidence that the clinical and pathogenetic pattern, and risk factor profile, of stroke varies among different ethnic groups.1 Asians, for example, have a high prevalence of intracranial large artery disease and cerebral small vessel disease and a stronger relationship between blood pressure and stroke risk.2,3 Investigating any race-ethnic disparity may improve understanding of stroke pathogenesis, and thus, the design of specific stroke prevention programs in different populations.

The retinal blood vessels, measuring 100 to 300 µm in size, offer a unique and easily accessible window to study correlates and consequences of cerebral microvascular disease, because the retina and brain share similar anatomic features and physiological properties.4 In the last decade, there have been several population-based and clinical studies in white populations which have consistently shown that retinal microvascular signs (eg, retinopathy and retinal venular widening), assessed from retinal photographs, are predictors of clinical stroke events,59 stroke death1013 and are associated with subclinical measures of cerebral disease.1417 However, there exist 2 major gaps in the current literature. First, there have been no prior studies examining these relationships in Asian populations. Second, although most studies have examined the role of the caliber (diameter) of the retinal vessels in stroke, a range of newer retinal microvascular features, such as retinal fractal dimension and tortuosity, which reflect the branching network and complexity of retinal vasculature, have recently been linked to stroke.1820 However, few studies have prospectively examined the relationship of these new retinal associations with stroke, and importantly, whether there is incremental benefit of a retinal photographic assessment, in terms of discrimination, calibration, and reclassification in stroke risk beyond established risk factors, is unknown.21

In this study, we examined prospectively the association of a comprehensive range of retinal microvascular abnormalities, including new microvascular parameters, and risk of stroke in an Asian population. We also assessed the incremental usefulness of adding retinal photographic measures to established risk factors for stroke prediction.

Methods

Study Population

Data for this analysis were derived from the Singapore Malay Eye study (SiMES), a population-based study of eye diseases in urban Malay adults ranging in age between 40 and 80 years residing in Southwestern Singapore. In brief, subjects were selected using an age-stratified (by 10-year age group) random sampling method, from a computer-generated list provided by the Singapore Ministry of Home Affairs. Of 4168 eligible participants, 3280 (78.7%) participated in the study, performed from August 2004 through June 2006. The methodology and objectives of the study population are reported in detail elsewhere.22 Written informed consent was obtained from each participant; the study performed adhered to the Declaration of Helsinki. Ethical approval was obtained from the Singapore Eye Research Institute Institutional Review Board. Participants underwent a standardized interview, systemic and ocular examination, and laboratory investigations at baseline.

Retinal Microvascular Assessment

Retinal Photography

Digital fundus photography was taken using a 45° digital retinal camera (Canon CR-DGi with a 10D SLR digital camera back, Canon, Japan) after pupil dilation using tropicamide 1% and phenylephrine hydrochloride 2.5%. Two retinal images of each eye were obtained, one centered at the optic disc and the other centered at the fovea. The spatial resolution of each image was 3072 by 2048 pixels and the images were stored without compression before analysis.

Quantitative Measurements of Retinal Microvasculature

We used a semiautomated computer-assisted program (Singapore I Vessel Assessment [SIVA], version 1.0) to measure the retinal microvascular parameters quantitatively from digital retinal photographs. Figure 1 shows a screenshot of retinal microvasculature analyses using the SIVA program. Trained graders, masked to participant characteristics, executed the SIVA program to measure the retinal microvasculature. SIVA automatically identifies the optic disc, places a grid with reference to the center of optic disc, identifies vessel type, and calculates the retinal microvascular parameters. Graders are responsible for the visual evaluation of SIVA-automated measurement and manual intervention if necessary, following a standardized grading protocol. All visible vessels coursing through the specified zone were measured. The intra- and intergrader reliability was assessed and reported previously.23 A series of retinal microvascular parameters were measured and outputted automatically to quantify the retinal vasculature comprehensively using the SIVA program including caliber, tortuosity, bifurcation, and fractal (Appendix in the online-only Data Supplement). The details of the retinal microvascular parameters are reported elsewhere.2325

Figure 1.

Figure 1. Retinal fundus photograph assessed quantitatively by the Singapore I Vessel Assessment (SIVA) software with superimposed modified Atherosclerosis Risk in Communities study (ARIC) grid and vessel tracing (arterioles are in red and venules are in blue). The modified ARIC grid is composed of 4 concentric circles which demarcate an average optic disc, zone B defined as the region from half-disc diameter to 1-disc diameter from the disc, and zone C defined as the region from half-disc diameter to 2-disc diameter from the disc.

Qualitative Assessment of Retinopathy Lesions

The digital retinal photographs were analyzed for qualitative changes by trained graders based on a standardized protocol and compared with a standard set of images. Retinopathy signs (microaneuryms, hemorrhages, or exudates) were assessed.23 Retinopathy was defined as being present if the retinopathy score (a scale modified from the Airlie House classification system) was at level ≥15.26 Typical examples of these signs are shown in the Figure in the online-only Data Supplement.

Assessment of Stroke Outcome

Previous stroke events were ascertained through self-reporting. Incident stroke from all causes that occurred during the period between baseline examination and December 31, 2009, was obtained by linking with the stroke cases registered by National Registry of Diseases Office, Singapore, by record linkage. Because the Register is electronically captured and is compulsory by law, and because of a unique identifier number for all Singapore citizens and residents (National Registration Identity Card), the Registry is expected to capture virtually all stroke cases. New stroke cases are identified by the Registry through various sources including notifications by healthcare professionals, hospital data for cases admitted to hospitals, and pathology and laboratory reports. The Registry manages the database with quality assurance to ensure that all data collected for the registries have been validated and are properly anonymized before analysis.27,28

Assessment of Established Risk Factors and C-reactive protein

The established risk factors (age, sex, systolic blood pressure, total cholesterol, high-density lipoprotein cholesterol, glycosylated hemoglobin, body mass index, smoking, and use of antihypertensive medication) and C-reactive protein were assessed under standardized conditions.

Systolic and diastolic blood pressures were measured using a digital automatic blood pressure monitor (Dinamap model Pro Series DP110X-RW, 100V2, GE Medical Systems Information Technologies Inc, Milwaukee, WI), after the subject was seated for at least 5 minutes. Blood pressure was measured twice, 5 minutes later. A third measurement was made if the systolic blood pressure differed by >10 mm Hg or the diastolic by >5 mm Hg. The mean between the 2 closest readings was then taken as the blood pressure of that individual. Hypertension was defined as systolic blood pressure of ≥140 mm Hg, diastolic blood pressure of ≥90 mm Hg, a self-reported history of physician-diagnosed hypertension or use of antihypertensive medication.

Nonfasting venous blood samples were analyzed at the National University Hospital Reference Laboratory for biochemical testing of serum total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglycerides, glycosylated hemoglobin, creatinine, and glucose. Diabetes mellitus was identified if the random plasma glucose was ≥11.1 mmol/L, participants’ self-reported use of diabetic medication, or had physician-diagnosed diabetes mellitus. Current smokers were defined as those currently smoking any number of cigarettes (ie, current versus past/never). Alcohol consumption was defined as those currently drinking alcoholic beverages daily or on some days (ie, current versus past/never). Body mass index was calculated as body weight (in kilograms) divided by body height (in meters squared).

Serum high-sensitivity C-reactive protein (hsCRP) was measured in frozen plasma that had been stored at −80°C at the National University Hospital Reference laboratory using an immuno-turbidimetric assay (Intra-assay precision 0.6–1.3%; Interassay precision 2.3–3.1%) implemented on a Roche Integra 400 (Roche Diagnostics, Rotkreuz, Switzerland). The detection limit of this assay is 0.07 mg/L and the coefficient variation is 2.9% at 6.3 mg/L and 3.9% at 108 mg/L mean value.

Statistical Analysis

Independent t test or χ2 tests were used to examine the differences in baseline characteristics between included and excluded subjects in SiMES. Cox proportional-hazards regression was used to examine the relation of selected individual retinal microvascular parameters measured by SIVA and retinopathy signs to the risk of stroke, adjusting for age-sex and established risk factors (age, sex, systolic blood pressure, total cholesterol, high-density lipoprotein cholesterol, glycosylated hemoglobin, body mass index, smoking, and use of antihypertensive medication) at baseline.

Next, we examined the incremental usefulness of adding retinal measures for stroke prediction. First, we only added the significant predictors from Cox proportional-hazards regression models. Second, we added all the outputted SIVA retinal vascular parameters and retinopathy signs in this analysis because these will all be included in the retinal assessment when extending the screening of retinal abnormalities in routine clinical practice in the next phase. We used principal component analysis to summarize the SIVA data to minimize loss of information. We retained the first to 12th unrotated components which explained 85% of the total variance in the principle component analysis. We calculated Harrell C-Statistic for the Cox regression models to quantify the predictive discrimination. We performed 5 models to examine the predictive discrimination. Model 1 included the established risk factors, model 2 included established risk factors+hsCRP, model 3 included retinal venular caliber and presence of retinopathy (the significant predictors from Cox proportional-hazards regression analysis) and established risk factors, model 4 included all retinal measures and established risk factors, and model 5 includes all retinal measures, established risk factors and hsCRP. Differences in C-Statistic after the addition of retinal measures and hsCRP to a model with established risk factors were estimated and compared.

We graphically assessed the calibration of risk prediction by comparing the predicted frequency of stroke events with the observed frequency. We calculated Hosmer–Lemeshow χ2 statistic for the models to test goodness-of-fit.29P<0.05 represents a significant difference between the expected and observed event rates, indicating that the model is not well-calibrated.

Finally, we examined the net reclassification improvement in stroke risk prediction when adding retinal measures. We first generated predicted probabilities for the incident stroke outcome using established risk factors only, and classified the subjects into low- (0% to 2%), intermediate (2% to 5%), and high-risk (>5%) categories.8 We then generated predicted probability for the incident stroke outcome again using both retinal measures and established risk factors. We used the same cut-off points and the subjects were reassigned to low-, intermediate, or high-risk categories.

In supplementary analyses, we used the factors from Framingham Risk score (age, sex, total cholesterol, high-density lipoprotein cholesterol, smoking, diabetes mellitus, systolic blood pressure, and use of antihypertensive medication) to further examine the incremental benefit of retinal measures.

We regarded P values of <0.05 from 2-sided tests to indicate statistical significance. All statistical analyses were performed using STATA version 12.

Results

The baseline characteristics of the SiMES population are shown in the Table in the online-only Data Supplement. We excluded the subjects with prevalent stroke (n=82), none or ungradable retinal images (n=348), and individuals with missing data (n=206). A total of 2644 participants (80.6% of SiMES) were included in the final analysis. Excluded participants were older, and had higher levels of blood pressure, serum creatinine, glucose, and glycosylated hemoglobin, comparing with the included subjects. During the follow-up period (median, 4.41 years; range, 0.15–5.35 years), 51 participants developed stroke in SiMES. Of these incident stroke cases, 82% were ischemic.

Cox Regression

The hazard ratios of selected retinal vascular parameters and retinopathy signs associated with stroke risk are shown in Table 1. In Cox proportional-hazards models adjusting for established risk factors, wider central retinal vein equivalent (hazard ratio [HR], 3.28; 95% confidence interval, 1.30–8.26, comparing fourth versus first quartiles) and presence of retinopathy (HR, 1.94; 95% confidence interval, 1.01–3.72) were associated with risk of stroke. Straighter retinal arteriole (HR, 0.38; 95% CI, 0.15–0.94, comparing second versus first quartiles) was also associated with risk of stroke, but the association was not significant when retinal arteriolar tortuosity was analyzed as continuous variable. Other retinal vascular parameters were not associated with risk of stroke. With the given sample size, we had a statistical power of ≥80% with a P value of 0.05 significance level to detect a HR of ≥1.5 with each SD increase in retinal vascular parameters and presence of retinopathy.

Table 1. Relation of Selected Major Retinal Vascular Parameters and Retinopathy Signs to Risk of Stroke

Age-Sex Adjusted Hazard Ratio (95% CI)P ValueMultivariable* Adjusted Hazard Ratio (95% CI)P Value
Central retinal artery equivalent
 First quartile1.25 (0.57–2.72)0.5751.23 (0.55–2.76)0.610
 Second quartile1.11 (0.50–2.44)0.8001.13 (0.51–2.52)0.767
 Third quartile1.15 (0.53–2.49)0.7181.15 (0.53–2.50)0.730
 Fourth quartileReferentReferent
 Per SD decrease1.20 (0.92–1.57)0.1861.19 (0.91–1.57)0.208
Central retinal vein equivalent
 First quartileReferentReferent
 Second quartile2.84 (1.09–7.42)0.0332.70 (1.03–7.08)0.043
 Third quartile2.24 (0.81–6.18)0.1212.03 (0.73–5.65)0.177
 Fourth quartile3.96 (1.60–9.82)0.0033.28 (1.30–8.26)0.012
 Per SD increase1.36 (1.07–1.73)0.0111.27 (1.00–1.63)0.052
Retinal arteriolar tortuosity
 First quartileReferentReferent
 Second quartile0.37 (0.15–0.91)0.0300.38 (0.15–0.94)0.036
 Third quartile0.95 (0.48–1.88)0.8890.97 (0.49–1.92)0.933
 Fourth quartile0.63 (0.29–1.38)0.2480.76 (0.34–1.67)0.493
 Per SD increase0.86 (0.62–1.20)0.3720.92 (0.66–1.27)0.616
Retinal venular tortuosity
 First quartileReferentReferent
 Second quartile0.92 (0.40–2.11)0.8470.86 (0.38–1.97)0.726
 Third quartile1.52 (0.72–3.21)0.2721.45 (0.69–3.07)0.330
 Fourth quartile1.26 (0.58–2.73)0.5581.18 (0.54–2.56)0.680
 Per SD increase1.01 (0.77–1.32)0.9600.99 (0.74–1.32)0.945
Retinal arteriolar branching angle
 First quartileReferentReferent
 Second quartile1.15 (0.56–2.37)0.7051.21 (0.58–2.50)0.611
 Third quartile0.86 (0.38–1.97)0.7210.89 (0.39–2.05)0.790
 Fourth quartile1.10 (0.51–2.34)0.8091.19 (0.55–2.56)0.653
 Per SD increase1.00 (0.77–1.30)0.9761.02 (0.78–1.33)0.865
Retinal venular branching angle
 First quartileReferentReferent
 Second quartile0.53 (0.24–1.18)0.1200.53 (0.24–1.18)0.120
 Third quartile0.57 (0.26–1.27)0.1690.54 (0.24–1.20)0.131
 Fourth quartile0.92 (0.46–1.84)0.8060.91 (0.45–1.83)0.786
 Per SD increase0.96 (0.74–1.25)0.7750.96 (0.74–1.24)0.743
Retinal vascular fractal dimension
 First quartileReferentReferent
 Second quartile0.77 (0.35–1.68)0.5040.76 (0.35–1.67)0.501
 Third quartile1.41 (0.68–2.92)0.3531.61 (0.77–3.35)0.206
 Fourth quartile1.17 (0.52–2.63)0.7041.36 (0.60–3.07)0.455
 Per SD increase1.02 (0.78–1.34)0.8851.08 (0.81–1.43)0.602
Presence of retinopathy2.90 (1.61–5.24)<0.0011.94 (1.01–3.72)0.048

CI indicates confidence interval; HbA1c, glycosylated hemoglobin; and HDL, high-density lipoprotein.

*Adjusted for age, sex, systolic blood pressure, total cholesterol, HDL cholesterol, smoking, HbA1c, and antihypertensive medication

Discrimination

Table 2 shows C-Statistic for Cox regression models predicting stroke in the whole sample. The C-Statistic for discrimination of stroke prediction increased from 0.792 to 0.826 after all retinal measures are added in the models with established risk factors (Model 4). The difference in the C-Statistic was statistically significant (P=0.017). Adding retinal venular caliber and presence of retinopathy (Model 3) and hsCRP (Models 2 and 5) did not improve the discrimination of stroke prediction significantly.

Table 2. C-Statistic for Cox Regression Models Predicting Stroke

Whole SampleC-Statistic for StrokeP Value*
Model 1: established risk factors0.792Referent
Model 2: established risk factors+hsCRP0.7920.812
Model 3: established risk factors+retinal venular caliber and presence of retinopathy0.8030.272
Model 4: established risk factors+all retinal measures0.8260.017
Model 5: established risk factors+all retinal measures and hsCRP0.8260.018

Established risk factors include age, sex, systolic blood pressure, total cholesterol, HDL cholesterol, HbA1c, body mass index, smoking, and use of antihypertensive medication at baseline. HDL indicates high-density lipoprotein; HbA1c, glycosylated hemoglobin; and hsCRP, high-sensitivity C-reactive protein.

*P values are for the comparison with the model with established risk factors.

Calibration

Figure 2 shows the calibration assessment by dividing the population at risk into deciles of predicted risk versus the observed event rate. The Hosmer-Lemeshow χ2 statistics was 5.31 (P=0.72) for the model with established risk factors and was 3.43 (P=0.90) with the addition of retinal measures (Model 4), indicating that neither model had a significant lack of fit.

Figure 2.

Figure 2. Calibration plots for the predictive models with (A) established risk factors and (B) established risk factors+retinal measures, by dividing the population at risk into deciles of predicted stroke risk vs the observed stroke event rate. The Hosmer-Lemeshow χ2 is used to test the goodness-of-fit.

Reclassification

Table 3 shows the reclassification of stroke risk after addition of retinal measures in the prediction models. Among the subjects with incident stroke event, 6 subjects (11.8%) were reclassified to a higher risk category and 3 subjects (5.9%) were reclassified to a lower risk category. Among the subjects without incident stroke event, 285 subjects (11.0%) were reclassified to a lower risk category and 198 subjects (7.6%) were reclassified to a higher risk category. The overall net reclassification improvement was 9.3% (P<0.001), indicating a significantly greater number were being reclassified appropriately than were being reclassified inappropriately.

Table 3. Five-Year Risk of Stroke Predicted by Models With and Without Retinal Measures, Singapore Malay Eye Study, 2004 to 2011

Retinal+Established Risk CategoryReclassificationNet Correct
LowIntermediateHighTotalHigher RiskLower RiskReclassification, %
Subjects with incident stroke event
 Established risk category
  Low101011635.9
  Intermediate210517
  High012223
  Total12122751
Subjects with no incident stroke event
 Established risk category
  Low1714971018211982853.4
  Intermediate19826491553
  High879132219
  Total19204402332593
Net reclassification improvement, %9.3(P<0.001)

Established risk factors include age, sex, systolic blood pressure, total cholesterol, high-density lipoprotein cholesterol, glycosylated hemoglobin, body mass index, smoking, and use of antihypertensive medication at baseline.

We repeated the analyses of discrimination and reclassification by using the factors from Framingham Risk score. The addition of retinal measures consistently improved discrimination and risk stratification for stroke in the models with Framingham Risk score.

Discussion

In this study, we demonstrated that retinal microvascular signs (retinopathy signs and larger retinal venular caliber), were associated with an increased risk of stroke in this Asian Malay population, consistent with studies in white populations.8,30 Furthermore, we showed that the addition of retinal measures significantly improves discrimination and risk stratification for stroke when compared with the models on established risk factors and hsCRP, although the absolute improvement in prediction was modest.

It has long been known that established or traditional risk factors do not fully predict stroke risk, accounting for only ≈60% of the attributable risk.31 Thus, there is value in examining additional biomarkers which may further improve stroke prediction. Microvascular processes which cause ischemia are important contributors to the brain diseases. Pathological studies have shown a close correlation between cerebral and retinal vascular findings.32 With advance in retinal photography, retinal vascular imaging has been proposed to be a noninvasive tool to assess the cerebral microvasculature objectively and reliably for better understanding of the pathophysiological processes involved in cerebral small vessel disease.4,33

To our best knowledge, this is the first study to examine the relationship between retinal microvascular abnormalities and the risk of stroke in an Asian population. Our finding adds to existing data, largely in white populations, which convincingly show links between retinal vascular changes (retinopathy signs and retinal venular widening) and clinical59,34 and subclinical stroke,1417 providing strong evidence that these 2 specific retinal microvascular changes are associated with stroke and the role of microvascular pathology in stroke development is also important in Malays.

In addition to retinopathy signs and retinal vascular caliber, we have also explored a range of new retinal vascular parameters, quantifying the geometric branching network (fractals, tortuosity, and bifurcation) of retinal vasculature, with risk of stroke in this study. These new retinal parameters may reflect the optimality and efficiency of blood distribution in the microvasculature,35,36 and have been recently linked to stroke.1820 However, there are limited prospective data on stroke and the results are inconsistent. Witt37 has reported that retinal vascular tortuosity and bifurcation are not associated with stroke mortality in a nested case-control study in a white population. In another prospective study, Kawasaki20 has showed that low spectrum fractal dimension is associated with 2-fold higher risk of incident stroke event or stroke-related mortality in a nested case-control study in a white population. We found that straighter retinal arteriole was marginally associated with risk of stroke; however, the association was not significant when retinal arteriolar tortuosity was analyzed as continuous variable, and we did not find any significant associations between retinal vascular fractal dimension and branching angle with stroke risk in this Asian cohort. The reason for the discrepancies may be related to smaller sample size that limited the power of our study to detect association with smaller effect sizes (HR<1.5). Alternatively, this may be related to ethnic differences (Asian Malays versus white), suggesting that microvascular pathology, as reflected by retinal vascular fractal dimension, may not play an equally important role in different ethnicities.

In terms of clinical prediction, there seems to be incremental value of adding all these new retinal vascular parameters, in addition to retinopathy signs and retinal venular caliber, for stroke prediction. We showed that addition of all retinal measures improves discrimination and reclassification for stroke, when compared with the models using only established risk factors. In addition, we did not observe any improvement for discrimination of stroke prediction when hsCRP was added in the models, consistent with reports that its discriminative value may be limited.3840 Our data suggest that the addition of retinal measures, which provides surrogate information on microvascular processes in cerebral vasculature, can improve risk stratification in identifying patients who are at high risk of stroke more accurately. However, further work is needed to optimize the risk-prediction algorithm, to determine how it may influence in clinical decision making, to identify the most appropriate subgroup of patients, and to estimate the cost-effectiveness and acceptability to patients in different settings.

Strengths of this study include a prospective design, a study population drawn from the community, standardized retinal measures from photographs, validated ascertainment of incident stroke cases, and detailed information on a variety of potential confounders. Potential limitations should also be mentioned. First, we do not have data on left ventricular hypertrophy and atrial fibrillation for SiMES which are factors for Framingham stroke risk score. Despite such limitation, the models still had generally good discrimination (via the C-Statistic) and calibration. Second, ischemic stroke are more common in Singapore.28 In this prospective cohort, 82% incident stroke cases were ischemic stroke. We did not have adequate power to perform stratified analysis by stroke subtypes. Third, only ethnic Malays were examined in this study and the findings may vary in other ethnic groups.

In summary, our population-based study in an Asian Malay cohort confirms previous findings in white populations that specific retinal microvascular signs (retinopathy and retinal venular widening) are associated with stroke risk. We further showed that the addition of retinal assessment significantly improves stroke risk discrimination and stratification beyond established risk factors and hsCRP, although the absolute improvement in risk prediction, and thus, the clinical significance of retinal vascular imaging for stroke risk stratification, requires further evaluation.

Disclosure

Dr Wong is advisory board member and consultant to Abbott, Novartis, Pfizer, Allergan, and Bayer. The other authors have no conflicts to report.

Footnotes

The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.113.001738/-/DC1.

Correspondence to Carol Y. Cheung, PhD, Singapore Eye Research Institute, Singapore National Eye Centre, 11 Third Hospital Ave, Singapore 168751. E-mail

References

  • 1. Sacco RL, Boden-Albala B, Abel G, Lin IF, Elkind M, Hauser WA, et al. Race-ethnic disparities in the impact of stroke risk factors: the Northern Manhattan Stroke study.Stroke. 2001; 32:1725–1731.LinkGoogle Scholar
  • 2. Eastern Stroke and Coronary Heart Disease Collaborative Research Group. Blood pressure, cholesterol, and stroke in eastern Asia.Lancet1998; 352:1801–1807.CrossrefMedlineGoogle Scholar
  • 3. Wong LK. Global burden of intracranial atherosclerosis.Int J Stroke. 2006; 1:158–159.CrossrefMedlineGoogle Scholar
  • 4. Wong TY. Is retinal photography useful in the measurement of stroke risk?Lancet Neurol. 2004; 3:179–183.CrossrefMedlineGoogle Scholar
  • 5. Mitchell P, Wang JJ, Wong TY, Smith W, Klein R, Leeder SR. Retinal microvascular signs and risk of stroke and stroke mortality.Neurology. 2005; 65:1005–1009.CrossrefMedlineGoogle Scholar
  • 6. Ikram MK, de Jong FJ, Bos MJ, Vingerling JR, Hofman A, Koudstaal PJ, et al. Retinal vessel diameters and risk of stroke: the Rotterdam Study.Neurology. 2006; 66:1339–1343.CrossrefMedlineGoogle Scholar
  • 7. Yatsuya H, Folsom AR, Wong TY, Klein R, Klein BE, Sharrett AR; ARIC Study Investigators. Retinal microvascular abnormalities and risk of lacunar stroke: Atherosclerosis Risk in Communities Study.Stroke. 2010; 41:1349–1355.LinkGoogle Scholar
  • 8. McGeechan K, Liew G, Macaskill P, Irwig L, Klein R, Klein BE, et al. Prediction of incident stroke events based on retinal vessel caliber: a systematic review and individual-participant meta-analysis.Am J Epidemiol. 2009; 170:1323–1332.CrossrefMedlineGoogle Scholar
  • 9. Wieberdink RG, Ikram MK, Koudstaal PJ, Hofman A, Vingerling JR, Breteler MM. Retinal vascular calibers and the risk of intracerebral hemorrhage and cerebral infarction: the Rotterdam Study.Stroke. 2010; 41:2757–2761.LinkGoogle Scholar
  • 10. Wong TY, Klein R, Klein BE, Tielsch JM, Hubbard L, Nieto FJ. Retinal microvascular abnormalities and their relationship with hypertension, cardiovascular disease, and mortality.Surv Ophthalmol. 2001; 46:59–80.CrossrefMedlineGoogle Scholar
  • 11. Wong TY, Klein R, Nieto FJ, Klein BE, Sharrett AR, Meuer SM, et al. Retinal microvascular abnormalities and 10-year cardiovascular mortality: a population-based case-control study.Ophthalmology. 2003; 110:933–940.CrossrefMedlineGoogle Scholar
  • 12. Wang JJ, Liew G, Klein R, Rochtchina E, Knudtson MD, Klein BE, et al. Retinal vessel diameter and cardiovascular mortality: pooled data analysis from two older populations.Eur Heart J. 2007; 28:1984–1992.CrossrefMedlineGoogle Scholar
  • 13. Sairenchi T, Iso H, Yamagishi K, Irie F, Okubo Y, Gunji J, et al. Mild retinopathy is a risk factor for cardiovascular mortality in Japanese with and without hypertension: the Ibaraki Prefectural Health Study.Circulation. 2011; 124:2502–2511.LinkGoogle Scholar
  • 14. Wong TY, Klein R, Sharrett AR, Couper DJ, Klein BE, Liao DP, et al; ARIC Investigators.Atheroslerosis Risk in Communities Study. Cerebral white matter lesions, retinopathy, and incident clinical stroke.JAMA. 2002; 288:67–74.CrossrefMedlineGoogle Scholar
  • 15. Cooper LS, Wong TY, Klein R, Sharrett AR, Bryan RN, Hubbard LD, et al. Retinal microvascular abnormalities and MRI-defined subclinical cerebral infarction: the Atherosclerosis Risk in Communities Study.Stroke. 2006; 37:82–86.LinkGoogle Scholar
  • 16. Kawasaki R, Cheung N, Mosley T, Islam AF, Sharrett AR, Klein R, et al. Retinal microvascular signs and 10-year risk of cerebral atrophy: the Atherosclerosis Risk in Communities (ARIC) study.Stroke. 2010; 41:1826–1828.LinkGoogle Scholar
  • 17. Cheung N, Mosley T, Islam A, Kawasaki R, Sharrett AR, Klein R, et al. Retinal microvascular abnormalities and subclinical magnetic resonance imaging brain infarct: a prospective study.Brain. 2010; 133(Pt 7):1987–1993.CrossrefMedlineGoogle Scholar
  • 18. Cheung N, Liew G, Lindley RI, Liu EY, Wang JJ, Hand P, et al; Multi-Center Retina & Stroke Study Collaborative Group. Retinal fractals and acute lacunar stroke.Ann Neurol. 2010; 68:107–111.CrossrefMedlineGoogle Scholar
  • 19. Doubal FN, MacGillivray TJ, Patton N, Dhillon B, Dennis MS, Wardlaw JM. Fractal analysis of retinal vessels suggests that a distinct vasculopathy causes lacunar stroke.Neurology. 2010; 74:1102–1107.CrossrefMedlineGoogle Scholar
  • 20. Kawasaki R, Che Azemin MZ, Kumar DK, Tan AG, Liew G, Wong TY, et al. Fractal dimension of the retinal vasculature and risk of stroke: a nested case-control study.Neurology. 2011; 76:1766–1767.CrossrefMedlineGoogle Scholar
  • 21. Hlatky MA, Greenland P, Arnett DK, Ballantyne CM, Criqui MH, Elkind MS, et al; American Heart Association Expert Panel on Subclinical Atherosclerotic Diseases and Emerging Risk Factors and the Stroke Council. Criteria for evaluation of novel markers of cardiovascular risk: a scientific statement from the American Heart Association.Circulation. 2009; 119:2408–2416.LinkGoogle Scholar
  • 22. Foong AW, Saw SM, Loo JL, Shen S, Loon SC, Rosman M, et al. Rationale and methodology for a population-based study of eye diseases in Malay people: the Singapore Malay Eye Study (SiMES).Ophthalmic Epidemiol. 2007; 14:25–35.CrossrefMedlineGoogle Scholar
  • 23. Cheung CY, Tay WT, Mitchell P, Wang JJ, Hsu W, Lee ML, et al. Quantitative and qualitative retinal microvascular characteristics and blood pressure.J Hypertens. 2011; 29:1380–1391.CrossrefMedlineGoogle Scholar
  • 24. Cheung CY, Hsu W, Lee ML, Wang JJ, Mitchell P, Lau QP, et al. A new method to measure peripheral retinal vascular caliber over an extended area.Microcirculation. 2010; 17:495–503.MedlineGoogle Scholar
  • 25. Hubbard LD, Brothers RJ, King WN, Clegg LX, Klein R, Cooper LS, et al. Methods for evaluation of retinal microvascular abnormalities associated with hypertension/sclerosis in the Atherosclerosis Risk in Communities Study.Ophthalmology. 1999; 106:2269–2280.CrossrefMedlineGoogle Scholar
  • 26. Wong TY, Cheung N, Tay WT, Wang JJ, Aung T, Saw SM, et al. Prevalence and risk factors for diabetic retinopathy: the Singapore Malay Eye Study.Ophthalmology. 2008; 115:1869–1875.CrossrefMedlineGoogle Scholar
  • 27. National Registry of Diseases Office Singapore. http://www.nrdo.gov.sg. Accessed November 2, 2012.Google Scholar
  • 28. Ministry of Health Singapore. Trends in stroke in Singapore 2005–2010.http://www.nrdo.gov.sg/uploadedFiles/NRDO/Health%20Factsheet_stroke%20(INP-12–4)_final.pdf. Accessed November 2, 2012.Google Scholar
  • 29. May S, Hosmer DW. A simplified method of calculating an overall goodness-of-fit test for the Cox proportional hazards model.Lifetime Data Anal. 1998; 4:109–120.CrossrefMedlineGoogle Scholar
  • 30. Doubal FN, Hokke PE, Wardlaw JM. Retinal microvascular abnormalities and stroke: a systematic review.J Neurol Neurosurg Psychiatry. 2009; 80:158–165.CrossrefMedlineGoogle Scholar
  • 31. Donnan GA, Fisher M, Macleod M, Davis SM. Stroke.Lancet. 2008; 371:1612–1623.CrossrefMedlineGoogle Scholar
  • 32. Goto I, Katsuki S, Ikui H, Kimoto K, Mimatsu T. Pathological studies on the intracerebral and retinal arteries in cerebrovascular and noncerebrovascular diseases.Stroke. 1975; 6:263–269.LinkGoogle Scholar
  • 33. Wong TY, Klein R, Klein BE, Tielsch JM, Hubbard L, Nieto FJ. Retinal microvascular abnormalities and their relationship with hypertension, cardiovascular disease, and mortality.Surv Ophthalmol. 2001; 46:59–80.CrossrefMedlineGoogle Scholar
  • 34. Wong TY, Klein R, Couper DJ, Cooper LS, Shahar E, Hubbard LD, et al. Retinal microvascular abnormalities and incident stroke: the Atherosclerosis Risk in Communities Study.Lancet. 2001; 358:1134–1140.CrossrefMedlineGoogle Scholar
  • 35. Murray CD. The physiological principle of minimum work: I. the vascular system and the cost of blood volume.Proc Natl Acad Sci USA. 1926; 12:207–214.CrossrefMedlineGoogle Scholar
  • 36. Sherman TF. On connecting large vessels to small. The meaning of Murray’s law.J Gen Physiol. 1981; 78:431–453.CrossrefMedlineGoogle Scholar
  • 37. Witt N, Wong TY, Hughes AD, Chaturvedi N, Klein BE, Evans R, et al. Abnormalities of retinal microvascular structure and risk of mortality from ischemic heart disease and stroke.Hypertension. 2006; 47:975–981.LinkGoogle Scholar
  • 38. Elkind MS, Leon V, Moon YP, Paik MC, Sacco RL. High-sensitivity C-reactive protein and lipoprotein-associated phospholipase A2 stability before and after stroke and myocardial infarction.Stroke. 2009; 40:3233–3237.LinkGoogle Scholar
  • 39. Cook NR, Buring JE, Ridker PM. The effect of including C-reactive protein in cardiovascular risk prediction models for women.Ann Intern Med. 2006; 145:21–29.CrossrefMedlineGoogle Scholar
  • 40. Lloyd-Jones DM, Liu K, Tian L, Greenland P. Narrative review: Assessment of C-reactive protein in risk prediction for cardiovascular disease.Ann Intern Med. 2006; 145:35–42.CrossrefMedlineGoogle Scholar

eLetters(0)

eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.