Abstract
Background and Purpose
Retinal microvasculature reflects cumulative small vessel damage from hypertension and other vascular processes. No study has prospectively examined retinal findings in relation to incidence of clinical lacunar stroke in comparison with other ischemic stroke subtypes.
Methods
In 10,496 adults initially free of stroke, we related retinal findings imaged during 1993-95 with incidence of hospitalized ischemic strokes through 2005.
Results
During a median of 11.2 years 338 incident ischemic strokes occurred (lacunar: 66, nonlacunar thrombotic: 192, cardioembolic: 80). Generalized arteriolar narrowing as measured by central retinal arteriole equivalent (CRAE) was associated with an increased incidence of lacunar stroke (multivariate-adjusted hazard ratio (HR) per 1-standard deviation (SD) decrement of CRAE: 1.67, 95% confidence interval (CI): 1.23-2.26), but was not associated with other ischemic stroke subtypes. Generalized venular widening as measured by central venule equivalent (CRVE) was also positively associated with only lacunar stroke (multivariate-adjusted HR per 1-SD increment: 1.44, 95% CI: 1.09-1.91). Retinal microvascular abnormalities were positively associated with lacunar stroke incidence (HR for focal arteriolar narrowing: 2.22, 95% CI: 1.11-4.48; for arteriovenous nicking: 2.38, 95% CI: 1.20-4.71), whereas retinopathy signs (microaneurysms, retinal hemorrhages, and others) were positively associated with nonlacunar thrombotic (HR: 2.41, 95% CI: 1.47-3.95) and cardioembolic stroke incidence (HR: 2.25, 95% CI: 1.09-4.65).
Conclusions
Narrower CRAE, wider CRVE, focal arteriolar narrowing and arteriovenous nicking were predictive of lacunar stroke. Retinal imaging is useful in understanding the pathophysiology and mechanisms of cerebral small vessel disease.
Introduction
Retinal imaging has been increasingly used as a noninvasive tool to directly evaluate subtle abnormalities or damage in the retinal microvasculature such as focal arteriolar narrowing, arteriovenous (AV) nicking, retinopathy and retinal vascular caliber.1 Because the retinal microvasculature shares embryologic and anatomic characteristics with that of the cerebral circulation,2 studying the associations of retinal abnormalities may provide clues to understanding the underlying pathophysiology of different cerebrovascular diseases.3 In the Atherosclerosis Risk in Communities (ARIC) Study, retinal microvascular abnormalities and generalized arteriolar narrowing were associated with an increased incidence of clinical stroke events,4 and in persons without clinical stroke, with a higher prevalence of MRI-defined subclinical cerebral infarction5 and white matter lesions.6 These findings have been confirmed in other populations.7-9 Furthermore, a clinic-based cross-sectional study among acute stroke patients found an association between retinal microvascular abnormalities and generalized arteriolar narrowing, and the presence of lacunar stroke, suggesting that microvascular disease pathways were more important in lacunar stroke.10 No previous studies, however, have prospectively examined whether the presence of retinal microvascular abnormalities is associated with an increased incidence of lacunar stroke in comparison with other ischemic stroke subtypes.11
We conducted a prospective analysis in the ARIC Study, where the standardized measurements of retinal images were performed for over 10,000 participants and the subsequent occurrence of stroke followed up for more than 10 years. We hypothesized that retinal microvascular abnormalities are associated prospectively with increased incidence of clinical lacunar stroke but not with other subtypes of ischemic stroke.
Methods
Study population
The ARIC Study included a cohort of 15,792 persons between 45 and 64 years of age at recruitment in 1987 through 1989.12 The present study sample is composed of individuals who participated in the third examination in 1993-95, when retinal photography was first performed. Of the 12,887 who participated in this examination, we excluded 38 whose race was neither black nor white, 245 with no retinal photographs, 1,566 with ungradeable photographs, 20 with retinal vein occlusions and 2 with retinal artery occlusions, and 303 with missing information on any of the covariates. We further excluded 120 with a history of stroke at visit 1, and 56 who developed stroke (43 ischemic and 13 hemorrhagic) between visit 1 and visit 3, leaving 10,496 subjects (1,381 black women, 797 black men, 4,485 white women, and 3,833 white men) free of stroke at visit 3 for the present analysis.
Assessment of retinal vascular caliber
The retinal photography procedure and grading of retinal microvascular signs are described elsewhere.13 Briefly, a 45-degree retinal photograph of 1 randomly selected eye of each participant was taken at visit 3 following 5 minutes of dark adaptation. This photograph was centered on the region of the optic disc and the macula and was taken using an autofocus camera.
Trained graders masked to participant characteristics used a computer-assisted approach to measure the calibers of all arterioles and venules coursing through a specified area surrounding the optic disc. Individual vessel measurements were combined into summary indices—the central retinal arteriole equivalent (CRAE) and the central retinal venule equivalent (CRVE). These indices represent the estimated central retinal arteriolar and venular caliber of the eye after taking into account the branching patterns. These measurements are reliable, with intragrader and intergrader reliability coefficients of 0.69 and 0.74 for CRAE and 0.89 and 0.77 for CRVE, respectively.13
Retinal microvascular abnormalities
Trained graders masked to the clinical status of the participants also assessed for presence of microvascular abnormalities. Definite focal arteriolar narrowing was defined if an arteriole estimated to be 50-μm diameter of greater had a constricted area of 2/3 or less the width of proximal and distal vessel segments. Definite AV nicking was defined if the venous blood column was tapered on both sides of its crossing under the arteriole. Retinopathy signs were evaluated without assumption of cause as described previously, and were considered present if any of the following findings (definite) were noted: retinal hemorrhages, microaneurysms, soft and hard exudates, macular edema, intraretinal microvascular abnormalities, venous beading, new vessels at the disc or elsewhere, fibrovascular proliferation, vitreous hemorrhage, disc swelling, and laser photocoagulation scars.4
Ascertainment of incident stroke
Hospitalized ischemic strokes that occurred by December 31, 2005 (median follow-up 11.2 years, from visit 3) were included in the present study. Details on quality assurance for ascertainment and classification of stroke are described elsewhere.14 All definite ischemic strokes were further classified as either lacunar or nonlacunar on the basis of the recorded neuroimaging results. A stroke was classified as “lacunar” when two criteria were met: (1) typical location of the infarct (basal ganglia, brain stem, thalamus, internal capsule, or cerebral white matter) and (2) infarct size of ≤ 2 cm or unstated size. Definite or probable “cardioembolic” stroke required the same criteria as ischemic infarction, plus either (1) autopsy evidence of an infarcted area in the brain and a source of possible cerebral emboli in a vessel or the presence of an embolus in the brain or (2) medical record evidence of a possible source of embolus such as moderate or greater valvular heart disease, atrial fibrillation, cardiac or arterial procedure, or intracardiac thrombus. Definite or probable ischemic strokes that were not deemed lacunar or embolic were labeled “nonlacunar.” CT or MRI was available for all the ischemic stroke cases except one case adjudicated as cardioembolic stroke with carotid ultrasound information. Strokes secondary to trauma, neoplasm, hematological abnormality, infection, or vasculitis were excluded.
Statistical analysis
Since no interactions of race or sex with retinal vascular caliber quintiles or retinal microvascular abnormality categories was observed (P>0.1, Wald test), analyses were done combining the race- and sex-groups. Age-, sex- and race- adjusted means and proportions among categories of retinal findings were calculated and tested by a general linear model. Since arteriolar and venular calibers are positively correlated, and the width of these calibers are determined not only by pathological processes but normal anatomical variation or measurement error, we simultaneously adjusted for the fellow-vessel caliber as a covariate in the analyses for retinal vascular calibers, as described previously.15 Cox proportional hazards regression was used to calculate minimally-adjusted (as described above) and multivariate-adjusted hazard ratios (HRs) and their 95% confidence intervals (CIs) for subtypes of ischemic stroke incidence in relation to quintiles of the retinal vascular calibers or the presence of microvascular abnormalities. Multivariate-adjusted models included variables in the minimal model plus educational level (more than high school graduate or not), smoking status (current, past, or never), usual ethanol intake (grams/week), leisure time sport index score (1.0-1.9, 2.0-2.4, 2.5-2.9, 3.0-5.0),16 systolic blood pressure (SBP), antihypertensive medication use or antiplatelet medication use, prevalent diabetes, waist circumference, and plasma high-density lipoprotein (HDL) cholesterol. All the covariates were assessed at visit 35 except for SBP for which we used the average value of visit 1 and visit 3 measurements. A sport score was calculated ranging from 1 (lowest) to 5 (highest). Diabetes was defined a history of, or treatment for, diabetes, a fasting glucose level of 126 mg/dl or greater, or a casual blood glucose level of 200 mg/dl or greater. Aspirin use encompassed use for seven or more days in the previous two weeks of the interview for the prevention of heart attack or stroke. A trend test for the retinal vascular caliber analysis was performed by assigning the median value of each quintile to corresponding individuals and treating it as a continuous variable in the model.
CRAE and CRVE were also examined as continuous variables as divided by one standard deviation (SD) of each variable. For both vessel quintiles and each microvascular abnormality, the hazard proportionality assumption was verified prior to the analyses. We tested whether proportional hazard regression coefficients for each retinal finding variable differed significantly by subtype, that is, between lacunar and nonlacunar and between lacunar and cardioembolic, using competing risks analysis.17 This was done by computing a test statistic: (b1-b2)2/{[se(b1)]2+[se(b2)]2} where b1 is the coefficient for a retinal finding variable for lacunar stroke, and b2 is the corresponding coefficient for the outcome to be compared, and se(b1) is the standard error of the variable for lacunar stroke and se(b2) is the corresponding standard error for the other outcome. The test statistic was compared to a 1-degree of freedom Wald chi-square distribution (two-sided, α=0.05). All statistical analyses were performed by SAS.
Results
Mean (SD) CRAE and CRVE were 162.4 (16.7) μm and 193.1 (16.7) μm, respectively. Men had a narrower CRAE than women (160.3 vs. 164.0 μm, p<0.0001) and blacks had a narrower CRAE than whites (160.9 vs. 162.7 μm, p<0.0001). Men had a wider CRVE than women (195.2 vs. 191.4 μm) and blacks had wider CRVE than whites (198.5 vs. 191.6 μm), both p<0.0001. CRAE was inversely associated with age (Table 1). Usual ethanol intake, antihypertensive medication use, SBP, and waist circumference were inversely associated with CRAE. In contrast, the prevalence of current smoking, ethanol intake, antihypertensive medication use, SBP, and diabetes were positively and HDL cholesterol was inversely associated with CRVE.
Table 1.
Age-, sex-, race, and CRVE- or CRAE-adjusted baseline characteristics according to generalized arteriolar width (CRAE quintile) and generalized venular width (CRVE quintile), ARIC, 1993-95
| Generalized arteriolar width (CRAE quintile)* |
Generalized venular width (CRVE quintile)† |
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | Q5 | Q1 | Q2 | Q3 | Q4 | Q5 | |||
| Range (Minimum- Maximum, μm) |
81.4 148.6 |
148.7 158.4 |
158.5 166.2 |
166.3 176.0 |
176.1 240.9 |
p‡ | 127.4 179.1 |
179.2 188.6 |
188.7 196.7 |
196.8 206.5 |
206.6 276.0 |
p‡ |
| Age | 60.0 | 60.0 | 59.7 | 59.3 | 58.9 | <.0001 | 60.2 | 59.7 | 59.4 | 59.3 | 59.2 | <.0001 |
| Men (%) | 54.7 | 47.3 | 43.9 | 40.3 | 34.1 | <.0001 | 32.1 | 40.7 | 44.6 | 48.3 | 54.9 | <.0001 |
| Black (%) | 23.2 | 20.7 | 22.4 | 21.2 | 16.2 | <.0001 | 9.2 | 15.3 | 19.8 | 25.9 | 33.6 | <.0001 |
| More than high school graduate (%) | 48.7 | 50.2 | 46.7 | 46.5 | 45.7 | 0.042 | 49.5 | 51.2 | 48.3 | 45.8 | 42.9 | <.0001 |
| Current smoker (%) | 17.4 | 15.1 | 15.6 | 16.2 | 22.5 | <.0001 | 9.2 | 12.0 | 15.2 | 19.4 | 31.1 | <.0001 |
| Usual ethanol intake (g/week) | 49.3 | 44.7 | 40.7 | 42.4 | 31.2 | <.0001 | 33.3 | 37.0 | 40.0 | 45.1 | 53.3 | <.0001 |
| Leisure time sports index (>=3) (%) | 30.8 | 32.8 | 33.4 | 36.0 | 36.5 | 0.0029 | 37.6 | 35.7 | 33.5 | 33.0 | 29.5 | <.0001 |
| Antihypertensive medication use (%) | 38.2 | 33.3 | 28.9 | 26.7 | 21.0 | <.0001 | 25.4 | 28.7 | 28.9 | 30.2 | 35.0 | <.0001 |
| Diabetes mellitus (%) | 15.0 | 13.3 | 14.5 | 13.8 | 13.0 | 0.38 | 11.8 | 13.0 | 12.8 | 13.1 | 18.8 | <.0001 |
| Aspirin use (%) | 10.3 | 10.9 | 9.9 | 10.3 | 11.0 | 0.78 | 11.2 | 10.7 | 10.6 | 9.7 | 10.2 | 0.64 |
| Waist circumference (cm) | 103.0 | 101.4 | 100.3 | 100.0 | 97.5 | <.0001 | 98.2 | 99.1 | 100.6 | 101.7 | 102.6 | <.0001 |
| Mean systolic blood pressure (mmHg) | 128.9 | 124.1 | 121.2 | 119.3 | 114.2 | <.0001 | 118.5 | 120.6 | 121.4 | 122.5 | 124.8 | <.0001 |
| High density lipoprotein cholesterol (IU/L) | 1.36 | 1.35 | 1.35 | 1.34 | 1.35 | 0.71 | 1.39 | 1.38 | 1.35 | 1.33 | 1.31 | <.0001 |
CRAE denotes central retinal arteriole equivalent; CRVE, central retinal vein equivalent.
Age, sex, race, and CRVE (continuous) adjusted.
Age, sex, race, and CRAE (continuous) adjusted.
Adjusted means and proportions and according to quintiles, and p-values for overall difference, were calculated by a general linear model.
The prevalence of definite focal arteriolar narrowing, AV nicking, and retinopathy signs were 7.0%, 5.8%, and 3.8%, respectively (Table 2). All the microvascular abnormalities were significantly associated with older age, greater antihypertensive medication use, higher SBP, and greater waist circumference.
Table 2.
Age-, sex-, and race-adjusted baseline characteristics according to presence of focal narrowing, arteriovenous nicking and retinop signs, ARIC, 1993-95
| Focal narrowing* |
Arteriovenous nicking* |
Retinopathy signs* |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Absent n=9,671 |
Present n=731 |
p† | Absent n=9,828 |
Present n=609 |
p‡ | Absent n=9,597 |
Present n=375 |
p‡ | |
| Age (y)‡ | 59.4 | 61.8 | <0.0001 | 59.5 | 60.8 | <0.0001 | 59.4 | 60.3 | 0.003 |
| Men (%)‡ | 44.2 | 42.9 | 0.5 | 44.1 | 44.1 | 1.0 | 43.6 | 50.1 | 0.01 |
| Black (%)‡ | 21.2 | 15.8 | <0.001 | 20.5 | 26.4 | <0.001 | 19.6 | 43.3 | <0.0001 |
| More than high school graduate (%) | 47.8 | 45.3 | 0.2 | 47.9 | 42.5 | 0.010 | 48.1 | 37.5 | <0.0001 |
| Current smoker (%) | 17.4 | 16.1 | 0.4 | 17.1 | 21.7 | 0.004 | 17.3 | 16.5 | 0.7 |
| Usual ethanol intake (g/week) | 41.5 | 42.6 | 0.7 | 41.2 | 46.0 | 0.2 | 41.7 | 23.0 | <0.0001 |
| Leisure time sports index (>=3) (%) | 34.1 | 30.3 | 0.04 | 34.0 | 31.2 | 0.2 | 34.2 | 26.2 | 0.001 |
| Antihypertensive medication use (%) | 29.1 | 37.3 | <0.0001 | 29.1 | 38.1 | <0.0001 | 29.0 | 46.9 | <0.0001 |
| Diabetes mellitus (%) | 13.9 | 13.5 | 0.8 | 13.8 | 14.2 | 0.8 | 12.1 | 56.0 | <0.0001 |
| Aspirin use (%) | 10.6 | 8.5 | 0.08 | 10.3 | 13.0 | 0.03 | 10.3 | 14.7 | 0.005 |
| Waist circumference (cm) | 100.3 | 102.5 | <0.0001 | 100.3 | 103.2 | <0.0001 | 100.3 | 105.1 | <0.0001 |
| Mean systolic blood pressure (mmHg) | 120.8 | 131.1 | <0.0001 | 121.3 | 125.7 | <0.0001 | 121.3 | 126.2 | <0.0001 |
| High density lipoprotein cholesterol (IU/L) | 1.35 | 1.36 | 0.8 | 1.35 | 1.32 | 0.05 | 1.36 | 1.24 | <0.0001 |
The numbers of subjects missing information were 94 for focal narrowing, 59 for arteriolar nicking, and 524 for retinopathy signs.
Age-, sex-, and race-adjusted means and proportions and according to presence of each microvascular abnormality, and p-values were calculated by a general linear model.
Sex- and race-adjusted mean age, Age- and race-adjusted proportion of men, and Age- and sex-adjusted proportion of blacks.
During a median of 11.2 years of follow-up (max=12.8 y), there were 338 incident ischemic strokes, 66 of which were lacunar, 192 nonlacunar thrombotic, and 80 cardioembolic. In the Cox regression analyses, CRAE was inversely associated only with lacunar stroke incidence (multivariate-adjusted HR for Q1 vs. Q5: 5.21, 95% CI: 1.92-14.1; HR for 1 SD decrement of CRAE: 1.67, 95% CI: 1.23-2.26, Table 3). A borderline association between CRAE and nonlacunar thrombotic stroke was totally attenuated by adjustment for SBP (66% reduction in the parameter estimate for CRAE after inclusion of SBP in the model). Competing risks analysis confirmed that the HR for 1 SD decrement of CRAE was significantly greater for lacunar stroke than nonlacunar thrombotic (P=0.03) or cardioembolic stroke (P=0.005).
Table 3.
Generalized arteriolar width (CRAE quintile) and incidence and hazard ratios (95% CIs) of ischemic stroke subtypes, ARIC, 1993-2005
| Generalized arteriolar width (CRAE quintile) | |||||||
|---|---|---|---|---|---|---|---|
| trend p | HR1 | ||||||
| Ischemic stroke subtype | Q5 | Q4 | Q3 | Q2 | Q1 | ||
| Lacunar | |||||||
| Number of cases | 7 | 12 | 17 | 10 | 20 | ||
| Person-years | 22,598 | 22,617 | 22,134 | 22,107 | 22,048 | ||
| Incidence rate* | 0.3 | 0.5 | 0.8 | 0.5 | 0.9 | ||
| Minimally-adjusted† | 1 (reference) | 2.15 (0.84-5.53) | 3.70 (1.49-9.21) | 2.68 (0.97-7.40) | 7.06 (2.69-18.49) | <0.001 | 1.85 (1.38-2.48)§ |
| Multivariate-adjusted‡ | 1 (reference) | 1.95 (0.75-5.03) | 3.31 (1.32-8.30) | 2.19 (0.78-6.12) | 5.21 (1.92-14.13) | 0.0014 | 1.67 (1.23-2.26)§ |
| Nonlacunar thrombotic | |||||||
| Number of cases | 37 | 28 | 38 | 46 | 43 | ||
| Person-years | 22,598 | 22,617 | 22,134 | 22,107 | 22,048 | ||
| Incidence rate* | 1.6 | 1.2 | 1.7 | 2.1 | 2.0 | ||
| Minimally-adjusted† | 1 (reference) | 0.76 (0.46-1.25) | 1.08 (0.67-1.73) | 1.35 (0.85-2.15) | 1.36 (0.82-2.25) | 0.065 | 1.18 (0.99-1.40) |
| Multivariate-adjusted‡ | 1 (reference) | 0.73 (0.44-1.19) | 1.00 (0.62-1.61) | 1.22 (0.76-1.96) | 1.05 (0.62-1.77) | 0.4 | 1.07 (0.90-1.28) |
| Cardioembolic | |||||||
| Number of cases | 18 | 20 | 15 | 12 | 15 | ||
| Person-years | 22,598 | 22,617 | 22,134 | 22,107 | 22,048 | ||
| Incidence rate* | 0.8 | 0.9 | 0.7 | 0.5 | 0.7 | ||
| Minimally-adjusted† | 1 (reference) | 1.19 (0.63-2.28) | 0.99 (0.49-2.02) | 0.87 (0.40-1.89) | 1.25 (0.57-2.74) | 0.8 | 1.05 (0.81-1.36) |
| Multivariate-adjusted‡ | 1 (reference) | 1.06 (0.55-2.03) | 0.85 (0.41-1.74) | 0.67 (0.31-1.45) | 0.78 (0.35-1.76) | 0.3 | 0.89 (0.68-1.17) |
HR denotes hazard ratio; CI, confidence interval; CRAE, central retinal artery equivalent.
Incidence rate is expressed as rate per 1,000 person-years.
Minimally-adjusted model includes age, sex, race and central retinal vein equivalent (CRVE).
Multivariate model includes age, sex, race, CRVE, smoking status, usual alcohol consumption, physical activity, education level, antihypertensive medication, prevalent diabetes, waist circumference, and HDL cholesterol all assessed at visit 3, and mean systolic blood pressure assessed at visit 1 and 3.
Trend test was performed by assigning the median value of each quintile to corresponding individuals and treating it as a continuous variable in the model. HR1: HR per 1 standard deviation (16.7 μm) decrement of CRAE.
HR1 for lacunar stroke significantly stronger than for nonlacunar (p=0.03) or cardioembolic (p=0.005) by competing risks analysis.
CRVE was positively associated also only with lacunar stroke incidence (multivariate-adjusted trend p=0.032; HR for 1 SD increment of CRVE: 1.44, 95% CI: 1.09-1.91, Table 4). Similarly to the CRAE analyses, SBP alone eliminated the significant associations between CRVE and nonlacunar thrombotic and cardioembolic stroke (25% and 20% reductions in the parameter estimates, respectively, after inclusion of SBP in the model) (data not shown). Despite the apparent stronger association of CRVE with lacunar stroke than other ischemic stroke subtypes, competing risks analysis showed that the HRs for 1 SD increment of CRVE were not significantly different across subtypes.
Table 4.
Generalized venular width (CRVE quintile) and incidence and hazard ratios (95% CIs) of ischemic stroke subtypes, ARIC, 1993-2005
| Generalized venular width (CRVE quintile) | |||||||
|---|---|---|---|---|---|---|---|
| trend p | HR1 | ||||||
| Ischemic stroke subtype | Q1 | Q2 | Q3 | Q4 | Q5 | ||
| Lacunar | |||||||
| Number of cases | 12 | 6 | 14 | 12 | 22 | ||
| Person-years | 22,467 | 22,322 | 22,469 | 22,398 | 21,848 | ||
| Incidence rate* | 0.5 | 0.3 | 0.6 | 0.5 | 1.0 | ||
| Minimally-adjusted† | 1 (reference) | 0.61 (0.23-1.66) | 1.52 (0.68-3.44) | 1.39 (0.58-3.32) | 2.93 (1.26-6.84) | 0.0028 | 1.71 (1.29-2.25)§ |
| Multivariate-adjusted‡ | 1 (reference) | 0.54 (0.20-1.47) | 1.32 (0.58-3.02) | 1.20 (0.50-2.91) | 2.03 (0.86-4.83) | 0.032 | 1.44 (1.09-1.91)§ |
| Nonlacunar thrombotic | |||||||
| Number of cases | 32 | 43 | 36 | 29 | 52 | ||
| Person-years | 22,467 | 22,322 | 22,469 | 22,398 | 21,848 | ||
| Incidence rate* | 1.4 | 1.9 | 1.6 | 1.3 | 2.4 | ||
| Minimally-adjusted† | 1 (reference) | 1.48 (0.93-2.37) | 1.30 (0.79-2.15) | 1.09 (0.64-1.87) | 2.15 (1.28-3.62) | 0.019 | 1.26 (1.06-1.49) |
| Multivariate-adjusted‡ | 1 (reference) | 1.44 (0.90-2.29) | 1.20 (0.73-1.97) | 0.95 (0.55-1.64) | 1.62 (0.95-2.75) | 0.2 | 1.12 (0.94-1.33) |
| Cardioembolic | |||||||
| Number of cases | 14 | 11 | 13 | 13 | 29 | ||
| Person-years | 22,467 | 22,322 | 22,469 | 22,398 | 21,848 | ||
| Incidence rate* | 0.6 | 0.5 | 0.6 | 0.6 | 1.3 | ||
| Minimally-adjusted† | 1 (reference) | 0.77 (0.35-1.72) | 0.89 (0.41-1.96) | 0.86 (0.38-1.93) | 1.90 (0.89-4.07) | 0.045 | 1.38 (1.06-1.78) |
| Multivariate-adjusted‡ | 1 (reference) | 0.75 (0.34-1.67) | 0.85 (0.39-1.86) | 0.77 (0.34-1.74) | 1.46 (0.67-3.16) | 0.2 | 1.26 (0.97-1.63) |
HR denotes hazard ratio; CI, confidence interval; CRVE, central retinal vein equivalent.
Incidence rate is expressed as rate per 1,000 person-years.
Minimally-adjusted model includes age, sex, race and central retinal artery equivalent (CRAE).
Multivariate model includes age, sex, race, CRAE, smoking status, usual alcohol consumption, physical activity, education level, antihypertensive medication, prevalent diabetes, waist circumference, and HDL cholesterol all assessed at visit 3, and mean systolic blood pressure assessed at visit 1 and 3.
Trend test was performed by assigning the median value of each quintile to corresponding individuals and treating it as a continuous variable in the model. HR1: HR per 1 standard deviation (16.7 μm) increment of CRVE.
HR1 for lacunar stroke not significantly different (p>0.05) from nonlacunar and cardioembolic by competing risks analysis.
Focal arteriolar narrowing and AV nicking were significantly and positively associated with lacunar stroke incidence (multivariate-adjusted HR for focal arteriolar narrowing: 2.22, 95% CI: 1.11-4.48, multivariate-adjusted HR for AV nicking: 2.38, 95% CI: 1.20-4.71, Table 5). Although retinopathy signs were significantly associated with lacunar stroke incidence in minimally-adjusted model, multivariate adjustment attenuated the association. This attenuation was explained by the adjustment for prevalent diabetes (43% reduction in the parameter estimate in the Cox model). In contrast, higher incidence of both nonlacunar thrombotic and cardioembolic stroke were significantly associated with retinopathy signs even after multivariate adjustment (HR for nonlacunar thrombotic stroke: 2.41, 95% CI: 1.47-3.95, HR for cardioembolic stroke: 2.25, 95% CI: 1.09-4.65). Despite these apparent differences among subtypes, the HR values were not significantly different. Results from sub-analyses with further adjustment for intima-media thickness were essentially the same.
Table 5.
Retinal microvascular abnormalities and incidence and hazard ratios (95% CIs) of ischemic stroke subtypes, ARIC, 1993-2005
| Retinal microvascular abnormalities |
||||||
|---|---|---|---|---|---|---|
| Focal narrowing |
Arteriovenous nicking |
Retinopathy signs‡ |
||||
| Ischemic stroke subtypes | Absent | Present | Absent | Present | Absent | Present |
| Lacunar | ||||||
| Number of cases | 56 | 10 | 56 | 10 | 53 | 9 |
| Person-years | 102,972 | 7,515 | 104,648 | 6,211 | 102,466 | 3,553 |
| Incidence rate* | 0.5 | 1.3 | 0.5 | 1.6 | 0.5 | 2.5 |
| Age-, sex-, race-adjusted | 1 (reference) | 2.69 (1.36-5.32)§ | 1 (reference) | 2.75 (1.40-5.40)§ | 1 (reference) | 3.62 (1.77-7.42)§ |
| Multivariate-adjusted † | 1 (reference) | 2.22 (1.11-4.48)§ | 1 (reference) | 2.38 (1.20-4.71)§ | 1 (reference) | 1.90 (0.88-4.100§ |
| Nonlacunar thrombotic | ||||||
| Number of cases | 164 | 25 | 173 | 19 | 159 | 21 |
| Person-years | 102,972 | 7,515 | 104,648 | 6,211 | 102,466 | 3,553 |
| Incidence rate* | 1.6 | 3.3 | 1.7 | 3.1 | 1.6 | 5.9 |
| Age-, sex-, race-adjusted | 1 (reference) | 1.77 (1.16-2.71) | 1 (reference) | 1.60 (1.00-2.57) | 1 (reference) | 3.21 (2.03-5.10) |
| Multivariate-adjusted † | 1 (reference) | 1.48 (0.96-2.29) | 1 (reference) | 1.43 (0.89-2.30) | 1 (reference) | 2.41 (1.47-3.95) |
| Cardioembolic | ||||||
| Number of cases | 72 | 8 | 75 | 5 | 65 | 10 |
| Person-years | 102,972 | 7,515 | 104,648 | 6,211 | 102,466 | 3,553 |
| Incidence rate* | 0.7 | 1.1 | 0.7 | 0.8 | 0.6 | 2.8 |
| Age-, sex-, race-adjusted | 1 (reference) | 1.46 (0.70-3.05) | 1 (reference) | 1.02 (0.41-2.53) | 1 (reference) | 3.67 (1.87-7.22) |
| Multivariate-adjusted † | 1 (reference) | 1.05 (0.49-2.22) | 1 (reference) | 0.85 (0.34-2.12) | 1 (reference) | 2.25 (1.09-4.65) |
Incidence rate is expressed as rate per 1,000 person-years.
Multivariate model includes age, sex, race, smoking status, usual alcohol consumption, physical activity, education level, antihypertensive medication, prevalent diabetes, waist circumference, HDL cholesterol all assessed at visit 3, and mean systolic blood pressure assessed at visit 1 and 3.
Retinopathy signs are any one of the following abnormalities: hemorrhages, microaneurysms, soft and hard exudates, intraretinal microvascular abnormalities, neovascularization at the disc and elsewhere, vitreous and preretinal hemorrhage, fibrovascular proliferation, macular edema, and laser photocoagulation scar.
HR for lacunar stroke not significantly different (p>0.05) from nonlacunar and cardioembolic by competing risks analysis.
Although subjects with hypertension and diabetes had about four times higher incidence rates of lacunar stroke than those without (1.2 vs. 0.3 and 1.8 vs. 0.4 per 1,000 person-years, respectively), there was no evidence hypertension or diabetes modified the associations of CRAE, CRVE, or other retinal microvascular abnormalities with stroke subtypes. However, the number of lacunar stroke events was low and power to detect interactions was limited.
Discussion
The main finding of this prospective population-based study is that narrower retinal arteriolar caliber and larger venular caliber, measured quantitatively from photographs, were associated significantly with incident lacunar stroke, independent of confounding variables such as SBP and diabetes. Similarly, focal arteriolar narrowing and AV nicking were also associated with lacunar stroke independent of the confounding variables, including SBP. However, one caveat is that formal statistical testing suggested that the HRs differed significantly between lacunar and other ischemic stroke subtypes only for CRAE. Yet, we consider it important that HRs for retinal findings were statistically significantly greater than 1.0 for lacunar stroke, despite the fewest number of events.
Our prospective population-based study demonstrating an association between retinal microvascular signs and incident lacunar stroke is consistent with a recent clinic-based cross-sectional study that showed acute lacunar stroke cases were more likely to have retinal focal arteriolar narrowing, AV nicking, generalized retinal arteriolar narrowing and generalized venular widening.10 The current finding further supports our previous analysis in an ARIC sub-sample in which MRI-defined subclinical cerebral infarct was associated positively with AV nicking and focal arteriolar narrowing, independent of blood pressure, diabetes and other factors.5 Taken in totality, these data provide further evidence that lacunar stroke is associated with retinal microvascular changes possibly reflecting similar pathological processes in the cerebrovasculature.
In contrast to retinal arteriolar narrowing, recent studies suggest that retinal venular widening is a risk marker of stroke, and is associated with measures of atherosclerosis, inflammation and dyslipidemia.18, 19 However, retinal venular widening remained independently associated with lacunar stroke incidence. Recently, retinal venular width was related to reduced arterial wall compliance in small arterial beds,20 which has been suggested to be a marker of endothelial dysfunction. Our finding would reinforce the view of possible involvement of endothelial dysfunction in the pathogenesis of lacunar stroke.21-23
In the present analysis, retinopathy signs were not significantly associated with lacunar stroke independent of diabetes status. Since the confidence interval was wide and the number of lacunar stroke cases with retinopathy signs was small (n=9), this finding might have arisen by chance. Further follow-up studies in patients with diabetes are probably needed to examine the significance of retinopathy signs for the prediction of lacunar stroke.
The present study provides further support that retinal imaging may yield significant predictive information for lacunar stroke that could not be obtained by measurements of conventional risk factors. There are several possible explanations for this finding. First, retinal microvascular signs may reflect susceptibility, vulnerability or damage of cerebral small vessels. Although such damage may be attributed to hypertension, retinal findings may be providing more information than just the degree of blood pressure elevation.24 Such vulnerability may be determined genetically,25 embryonically,26 or by other environmental factors;27 however, whether or not these factors modify the effect of blood pressure on retinal microcirculation has yet to be elucidated. Other confounding or mediating variables exist besides those related to blood pressure, which were not measured, such as markers of inflammation/immunity,28 autonomic function,29, 30 endothelial dysfunction,22 and lifestyle factors including diet.31
There are several limitations that warrant discussion. First, there was a certain degree of imprecision in the retinal measurements, which, however, would likely have attenuated the association toward the null. Second, the number of cases for each ischemic stroke subtype was small. To confirm the present findings, a larger study would be needed. A strength of the present study is that we analyzed the association of retinal microvascular variables and ischemic stroke subtypes using prospective population-based data, including both blacks and whites, with a relatively large numbers of events. There are no prior prospective studies that specifically addressed the association between these measures and ischemic stroke subtypes in detail.
In conclusion, retinal vessel calibers and the presence of retinal microvascular abnormalities, measured from photographs, were predictive of lacunar stroke. These data suggest that retinal imaging is useful in understanding the pathophysiology and mechanisms of cerebral small vessel disease.
Acknowledgements
The authors thank the staff and participants of the ARIC study for their important contributions.
Sources of Funding
The ARIC Study is carried out as a collaborative study supported by National Heart, Lung, and Blood Institute contracts N01-HC-55015, N01-HC-55016, N01-HC-55018, N01-HC-55019, N01-HC-55020, N01-HC-55021, and N01-HC-55022.
Footnotes
Disclosures
None.
References
- 1.Liew G, Wang JJ, Mitchell P, Wong TY. Retinal vascular imaging: A new tool in microvascular disease research. Circ Cardiovasc Imaging. 2008;1:156–161. doi: 10.1161/CIRCIMAGING.108.784876. [DOI] [PubMed] [Google Scholar]
- 2.Patton N, Aslam T, Macgillivray T, Pattie A, Deary IJ, Dhillon B. Retinal vascular image analysis as a potential screening tool for cerebrovascular disease: A rationale based on homology between cerebral and retinal microvasculatures. J Anat. 2005;206:319–348. doi: 10.1111/j.1469-7580.2005.00395.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kwa VI, van der Sande JJ, Stam J, Tijmes N, Vrooland JL. Retinal arterial changes correlate with cerebral small-vessel disease. Neurology. 2002;59:1536–1540. doi: 10.1212/01.wnl.0000033093.16450.5c. [DOI] [PubMed] [Google Scholar]
- 4.Wong TY, Klein R, Couper DJ, Cooper LS, Shahar E, Hubbard LD, Wofford MR, Sharrett AR. Retinal microvascular abnormalities and incident stroke: The Atherosclerosis Risk in Communities Study. Lancet. 2001;358:1134–1140. doi: 10.1016/S0140-6736(01)06253-5. [DOI] [PubMed] [Google Scholar]
- 5.Cooper LS, Wong TY, Klein R, Sharrett AR, Bryan RN, Hubbard LD, Couper DJ, Heiss G, Sorlie PD. Retinal microvascular abnormalities and mri-defined subclinical cerebral infarction: The Atherosclerosis Risk in Communities Study. Stroke. 2006;37:82–86. doi: 10.1161/01.STR.0000195134.04355.e5. [DOI] [PubMed] [Google Scholar]
- 6.Wong TY, Klein R, Sharrett AR, Couper DJ, Klein BE, Liao DP, Hubbard LD, Mosley TH. Cerebral white matter lesions, retinopathy, and incident clinical stroke. JAMA. 2002;288:67–74. doi: 10.1001/jama.288.1.67. [DOI] [PubMed] [Google Scholar]
- 7.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. doi: 10.1212/01.wnl.0000179177.15900.ca. [DOI] [PubMed] [Google Scholar]
- 8.Wong TY, Kamineni A, Klein R, Sharrett AR, Klein BE, Siscovick DS, Cushman M, Duncan BB. Quantitative retinal venular caliber and risk of cardiovascular disease in older persons: The Cardiovascular Health Study. Arch Intern Med. 2006;166:2388–2394. doi: 10.1001/archinte.166.21.2388. [DOI] [PubMed] [Google Scholar]
- 9.Ikram MK, de Jong FJ, Bos MJ, Vingerling JR, Hofman A, Koudstaal PJ, de Jong PT, Breteler MM. Retinal vessel diameters and risk of stroke: The Rotterdam Study. Neurology. 2006;66:1339–1343. doi: 10.1212/01.wnl.0000210533.24338.ea. [DOI] [PubMed] [Google Scholar]
- 10.Lindley RI, Wang JJ, Wong MC, Mitchell P, Liew G, Hand P, Wardlaw J, De Silva DA, Baker M, Rochtchina E, Chen C, Hankey GJ, Chang HM, Fung VS, Gomes L, Wong TY. Retinal microvasculature in acute lacunar stroke: A cross-sectional study. Lancet Neurol. 2009;8:628–634. doi: 10.1016/S1474-4422(09)70131-0. [DOI] [PubMed] [Google Scholar]
- 11.Konishi M, Iso H, Komachi Y, Iida M, Shimamoto T, Jacobs DR, Jr., Terao A, Baba S, Sankai T, Ito M. Associations of serum total cholesterol, different types of stroke, and stenosis distribution of cerebral arteries. The Akita Pathology Study. Stroke. 1993;24:954–964. doi: 10.1161/01.str.24.7.954. [DOI] [PubMed] [Google Scholar]
- 12.The Atherosclerosis Risk in Communities (ARIC) study: Design and objectives The ARIC investigators. Am J Epidemiol. 1989;129:687–702. [PubMed] [Google Scholar]
- 13.Hubbard LD, Brothers RJ, King WN, Clegg LX, Klein R, Cooper LS, Sharrett AR, Davis MD, Cai J. Methods for evaluation of retinal microvascular abnormalities associated with hypertension/sclerosis in the Atherosclerosis Risk in Communities Study. Ophthalmology. 1999;106:2269–2280. doi: 10.1016/s0161-6420(99)90525-0. [DOI] [PubMed] [Google Scholar]
- 14.Rosamond WD, Folsom AR, Chambless LE, Wang CH, McGovern PG, Howard G, Copper LS, Shahar E. Stroke incidence and survival among middle-aged adults: 9-year follow-up of the Atherosclerosis Risk in Communities (ARIC) cohort. Stroke. 1999;30:736–743. doi: 10.1161/01.str.30.4.736. [DOI] [PubMed] [Google Scholar]
- 15.Liew G, Sharrett AR, Kronmal R, Klein R, Wong TY, Mitchell P, Kifley A, Wang JJ. Measurement of retinal vascular caliber: Issues and alternatives to using the arteriole to venule ratio. Invest Ophthalmol Vis Sci. 2007;48:52–57. doi: 10.1167/iovs.06-0672. [DOI] [PubMed] [Google Scholar]
- 16.Baecke JA, Burema J, Frijters JE. A short questionnaire for the measurement of habitual physical activity in epidemiological studies. Am J Clin Nutr. 1982;36:936–942. doi: 10.1093/ajcn/36.5.936. [DOI] [PubMed] [Google Scholar]
- 17.Allison PD. Survival analysis using the SAS system. SAS Institute Inc.; Cary, NC: 1995. Competing risks. Chapter 6. [Google Scholar]
- 18.Ikram MK, de Jong FJ, Vingerling JR, Witteman JC, Hofman A, Breteler MM, de Jong PT. Are retinal arteriolar or venular diameters associated with markers for cardiovascular disorders? The Rotterdam Study. Invest Ophthalmol Vis Sci. 2004;45:2129–2134. doi: 10.1167/iovs.03-1390. [DOI] [PubMed] [Google Scholar]
- 19.McGeechan K, Liew G, Macaskill P, Irwig L, Klein R, Klein BE, Wang JJ, Mitchell P, Vingerling JR, de Jong PT, Witteman JC, Breteler MM, Shaw J, Zimmet P, Wong TY. 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. doi: 10.1093/aje/kwp306. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Cheung N, Islam FM, Jacobs DR, Jr., Sharrett AR, Klein R, Polak JF, Cotch MF, Klein BE, Ouyang P, Wong TY. Arterial compliance and retinal vascular caliber in cerebrovascular disease. Ann Neurol. 2007;62:618–624. doi: 10.1002/ana.21236. [DOI] [PubMed] [Google Scholar]
- 21.Knottnerus IL, Ten Cate H, Lodder J, Kessels F, van Oostenbrugge RJ. Endothelial dysfunction in lacunar stroke: A systematic review. Cerebrovasc Dis. 2009;27:519–526. doi: 10.1159/000212672. [DOI] [PubMed] [Google Scholar]
- 22.Kim JS, Lee HS, Park HY, Kim SS, Kang HG, Kim NH, Park JS, Kim Y. Endothelial function in lacunar infarction: A comparison of lacunar infarction, cerebral atherosclerosis and control group. Cerebrovasc Dis. 2009;28:166–170. doi: 10.1159/000226115. [DOI] [PubMed] [Google Scholar]
- 23.Arboix A. Retinal microvasculature in acute lacunar stroke. Lancet Neurol. 2009;8:596–598. doi: 10.1016/S1474-4422(09)70137-1. [DOI] [PubMed] [Google Scholar]
- 24.Sun C, Wang JJ, Mackey DA, Wong TY. Retinal vascular caliber: Systemic, environmental, and genetic associations. Surv Ophthalmol. 2009;54:74–95. doi: 10.1016/j.survophthal.2008.10.003. [DOI] [PubMed] [Google Scholar]
- 25.Sun C, Zhu G, Wong TY, Hewitt AW, Ruddle JB, Hodgson L, Montgomery GW, Young TL, Hammond CJ, Craig JE, Martin NG, He M, Mackey DA. Quantitative genetic analysis of the retinal vascular caliber: The Australian Twins Eye Study. Hypertension. 2009;54:788–795. doi: 10.1161/HYPERTENSIONAHA.109.132902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Mitchell P, Liew G, Rochtchina E, Wang JJ, Robaei D, Cheung N, Wong TY. Evidence of arteriolar narrowing in low-birth-weight children. Circulation. 2008;118:518–524. doi: 10.1161/CIRCULATIONAHA.107.747329. [DOI] [PubMed] [Google Scholar]
- 27.Nagaoka T, Kuo L, Ren Y, Yoshida A, Hein TW. C-reactive protein inhibits endothelium-dependent nitric oxide-mediated dilation of retinal arterioles via enhanced superoxide production. Invest Ophthalmol Vis Sci. 2008;49:2053–2060. doi: 10.1167/iovs.07-1387. [DOI] [PubMed] [Google Scholar]
- 28.Liew G, Sharrett AR, Wang JJ, Klein R, Klein BE, Mitchell P, Wong TY. Relative importance of systemic determinants of retinal arteriolar and venular caliber: The Atherosclerosis Risk in Communities Study. Arch Ophthalmol. 2008;126:1404–1410. doi: 10.1001/archopht.126.10.1404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Matz K, Tatschl C, Sebek K, Dachenhausen A, Brainin M. Dyslipidemia, elevated ldl cholesterol and reduced nocturnal blood pressure dipping denote lacunar strokes occurring during nighttime. Eur J Neurol. 2004;11:742–748. doi: 10.1111/j.1468-1331.2004.00811.x. [DOI] [PubMed] [Google Scholar]
- 30.Castilla-Guerra L, Espino-Montoro A, Fernandez-Moreno MC, Lopez-Chozas JM. Abnormal blood pressure circadian rhythm in acute ischaemic stroke: Are lacunar strokes really different? Int J Stroke. 2009;4:257–261. doi: 10.1111/j.1747-4949.2009.00314.x. [DOI] [PubMed] [Google Scholar]
- 31.Kaushik S, Wang JJ, Wong TY, Flood V, Barclay A, Brand-Miller J, Mitchell P. Glycemic index, retinal vascular caliber, and stroke mortality. Stroke. 2009;40:206–212. doi: 10.1161/STROKEAHA.108.513812. [DOI] [PubMed] [Google Scholar]