Journal of Glaucoma Publish Ahead of Print

DOI:10.1097/IJG.0000000000000829

Relationship between Open-Angle Glaucoma and

Stroke: A 2010-2012 Korea National Health and

Nutrition Examination Survey

Won June Lee, MD1,2 · Jin Wook Jeoung, MD,PhD1,2* · Kyeong Ik Na, MD1,2 ·

Young Kook Kim, MD1,2 · Chan Yun Kim, MD, PhD3 · Ki Ho Park, MD,PhD1,2*,

on behalf of the Epidemiologic Survey Committee of the Korean

Ophthalmologic Society
1
Department of Ophthalmology, Seoul National University College of Medicine, Seoul,

Korea
2
Department of Ophthalmology, Seoul National University Hospital, Seoul, Korea
3
Department of Ophthalmology, Institute of Vision Research, Yonsei University College of

Medicine, Seoul, South Korea

Meeting Presentation: None

Financial Support: None

Conflict of Interest: No conflicting relationship exists for any author.

Running Head: Open Angle Glaucoma and Stroke

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 *Both J.W.J. and K.H.P. contributed equally as corresponding authors.

Correspondence to:

Jin Wook Jeoung, MD, PhD

Department of Ophthalmology, Seoul National University Hospital, Seoul National

University College of Medicine, 101 Daehak-ro, Jongno-gu, Seoul 03080, Korea

Tel: +82-2-2072-2438, Fax: +82-2-741-3187

E-mail: neuroprotect@gmail.com

Ki Ho Park, MD, PhD

Department of Ophthalmology, Seoul National University Hospital, Seoul National

University College of Medicine, 101 Daehak-ro, Jongno-gu, Seoul 03080, Korea

Tel: +82-2-2072-3172, Fax: +82-2-741-3187

E-mail: kihopark@snu.ac.kr

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 ABSTRACT

Purpose: To determine the relationship between open-angle glaucoma (OAG) and stroke in

the Korean population.

Methods: This study was a population-based, cross-sectional study that used data from the

Korea National Health and Nutrition Examination Survey (KNHANES), a complex, stratified,

multistage, probability-cluster survey. We analyzed a total of 11,959 participants who were

>40 years old and were included in the fifth KNHANES database (2010-2012 data).

Weighted prevalence of OAG and stroke in various populations were estimated. Univariate

and multivariate logistic regressions were performed to examine potential risk factors for

stroke, including OAG. Multivariate adjusted odds ratios (OR) for stroke were examined in

subjects with and without OAG. Subjects were also examined by hypertension and diabetes

status.

Results: Stroke patients had a significantly higher OAG prevalence (8.5%) than non-stroke

patients (3.8%, p < 0.001). After adjusting for selected risk factors, the presence of OAG

tended to be associated with stroke (adjusted OR = 1.629), but this result was not statistically

significant (p = 0.053). Stroke and OAG were significantly associated with each other in

subjects with hypertension (OR = 2.059, p = 0.010) and diabetes (OR = 2.649, p = 0.040).

Additionally, the associations of stroke and other systemic diseases were stronger when OAG

was a comorbid condition.

Conclusions: Although there was no overall statistical significance, patients with OAG had

an increased risk of stroke, among those with systemic comorbidities, including hypertension

and diabetes. Our results may provide insight on the underlying mechanisms of OAG and

their association with stroke development.

Key wards: Open angle glaucoma; stroke; epidemiological study; KNHANES

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 Introduction

Glaucoma is a leading causes of irreversible blindness.1-3 The pathophysiology of glaucoma

is not yet fully understood, but 2 main theories have emerged. One theory is mechanical in

nature and suggests that an increased intraocular pressure (IOP) results in glaucomatous

damage. Another theory is vascular in nature and suggests that impaired vascular function

plays a role in the development of glaucomatous damage.3-5 Several large population-based

studies have demonstrated that systemic comorbidities, including diabetes,6, 7 hypertension,8-

11
stroke,11 and dyslipidemia11 are common risk factors of glaucoma.

A vascular mechanism for glaucoma development could result in similarities between

stroke (associated with abnormal cerebral vasculature) and OAG (associated with abnormal

ocular blood flow parameters) development. Unfortunately, only a few studies in the

literature have examined OAG as a stroke risk factor.12-14 However, these studies provide

preliminary evidence for an association between glaucoma and stroke pathophysiology.

We recently reported OAG prevalence and risk factors in Korea using data from the Korea

National Health and Nutrition Examination Survey (KNHANES).9 The KNHANES is a large,

population-based survey that can determine nationally-representative estimates of disease.

Here, we use KNHANES data to investigate the possible relationship between OAG and

stroke in a very large sample.

Materials and methods

This study adhered to the tenets of the Declaration of Helsinki; written informed consent was

obtained from all of the KNHANES participants. The survey protocol was approved by the

Institutional Review Board of the Korea Center for Disease Control and Prevention (KCDC).

As all of the KNHANES data is de-identified and opened to the public, the Institutional

Review Board of the Seoul National University Hospital determined that this study was

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 exempt from requiring their approval.

Study design and population

The KNHANES is an ongoing, population-based cross-sectional survey conducted in South

Korea by the KCDC and the Korean Ministry of Health and Welfare. Initiated in 1998, it has

been completed every year since 2007. The KNHANES uses a complex, stratified, multistage,

probability-cluster survey with rolling sampling designs to analyze a representative, civilian,

non-institutionalized South Korean population. By means of this multistage sampling design

(using strata, cluster, and weight), data from KNHANES may be considered to represent the

entire population of Korea. The detailed design of the KNHANES has been previously

described.9, 15, 16 The KNHANES selects sample households from among the included

districts annually using systematic sampling. Approximately 60 people from 20 households

were included per district in the KNHANES V. Within households, all family members aged

>1 year were included as eligible subjects and asked to participate in the survey.

The subjects included in our analysis met the following criteria: (1) aged ≥ 40 years; (2)

completed the health interview survey and health examination survey; and (3) had gradable

fundus photograph results for at least 1 eye.

Examination

The KNHANES consists of 3 parts: the health interview survey, the health examination

survey including a comprehensive ophthalmologic examination, and the nutrition survey. The

data from the first 2 surveys were used in this study. The health interview survey included

standardized questionnaires on demographic variables, as well as current and past medical

conditions, health-influencing behaviors and socioeconomic status. The health examination

survey included ophthalmologic examination. The detailed ophthalmologic examination was

conducted by ophthalmologists who had been periodically trained by the Korean

Ophthalmological Society National Epidemiologic Survey Committee. It included

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 measurement of visual acuity by Snellen chart, IOP by Goldmann applanation tonometry,

spherical equivalent (SE) using automatic refractometry (KR-8800; Topcon, Tokyo, Japan),

and slit-lamp biomicroscopy (Haag-Streit model BQ-900; Haag-Streit AG, Koeniz,

Switzerland). The anterior chamber and angle assessment was performed using the Van

Herick method, since gonioscopy was not included in the KNHANES. Participants aged ≥ 19

years underwent non-mydriatic fundus photography with a 45 degree field angle digital

fundus camera (TRC-NW6S; Topcon) in a dark room. All of the participants underwent

frequency doubling technology (FDT) perimetry with the screening program N30-1

(Humphrey Matrix FDT perimetry, Carl Zeiss Meditec Inc., Dublin, CA) if they had elevated

IOP of ≥ 22 mmHg or a glaucomatous optic disc appearance: (1) horizontal or vertical cup to-

disc (CD) ratio ≥ 0.5, (2) violation of the ISNT rule (the neuroretinal rim thickness order of

inferior > superior > nasal > temporal), (3) presence of optic disc hemorrhage, or (4) presence

of retinal nerve fiber layer (RNFL) defect.

Determination of glaucoma diagnosis

A diagnosis was made based on the non-mydriatic fundus photography and FDT perimetry

findings, in accordance with the International Society of Geographical and Epidemiological

Ophthalmology (ISGEO) criteria.17 Two discrete glaucoma reading committees were

established, each comprising glaucoma specialists from different institutes. The detailed

process has been previously described.9

The OAG group was defined as having an open angle using the Van Herick method

(peripheral anterior chamber depth > 1/4 peripheral corneal thickness) with any of the

following modified ISGEO criteria category I or II in one or two eyes. The category I criteria

were applied to subjects with FDT perimetry results with fixation error and false-positive

error ≤ 1. The glaucoma-diagnostic criteria were (1) loss of neuroretinal rim with vertical or

horizontal cup-to-disc (CD) ratio ≥ 0.7 or asymmetry of CD ratio ≥ 0.2 (both values

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 determined by ≥ 97.5th percentile for normal population in KNHANES), (2) presence of

optic disc hemorrhage, or (3) presence of RNFL defect. Additionally, the subjects had to

show an abnormal FDT testing result with at least 1 location of reduced sensitivity

compatible with optic disc appearance or RNFL defect. Criteria II were applied to those with

absence of FDT perimetry results or fixation error or false positive error ≥ 2 with (1) loss of

neuroretinal rim with vertical CD ratio ≥ 0.9 or asymmetry of vertical CD ratio ≥ 0.3 (both

values determined by ≥ 99.5th percentile for normal population in KNHANES) or (2)

presence of RNFL defect compatible with optic disc appearance.

The non-OAG group was defined as satisfying all of the following criteria in both eyes: (1)

IOP ≤ 21 mmHg, (2) presence of open angle, (3) non glaucomatous optic disc appearance, (4)

absence of optic disc hemorrhage or RNFL defect, and (5) optic disc not violating ISNT rule.

The subjects not included in either the OAG or non-OAG groups were categorized as the

“others” group. This included patients with ocular hypertension (high IOP without

glaucomatous damage), glaucoma with narrow or closed angle (primary angle closure), and

suspected glaucoma or preperimetric OAG (glaucomatous optic disc appearance with normal

FDT), and they were included in the total number of subjects when calculating the prevalence.

Definition of variables

The variables evaluated as risk factors were as follows: (1) demographics including age,
gender and smoking status; (2) systemic parameters including body mass index (BMI); (3)
history or presence of medical conditions including self-declared diabetes, hypertension,
dyslipidemia, coronary arterial disease (myocardial infarction and ischemic heart disease) and
cerebral stroke; was used for the analysis. BMI was calculated as the ratio of weight/height2
(kg/m2)
Statistical analyses

All 2010-2012 estimates were derived using sample weights statistically adjusted for

response rate, extraction rate, and Korean population distribution. Prevalence estimates for all

outcomes were estimated for the overall sample and are expressed as mean values with 95%

confidence intervals (CI).

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 Variables were categorized in risk factor analyses in several different ways and logistic

regression analysis reference groups are presented in data tables. Participants were

categorized into age (40–49 years, 50–59 years, 60–69 years, 70–79 years, and over 80 years),

smoking (ever smoker/current smoker), systemic comorbidities (positive history, negative

history), OAG (positive for OAG, negative for OAG), and stroke (stroke, non-stroke) groups.

Subject demographics in the stroke and non-stroke groups were compared using Student’s

t-tests for continuous variables and chi-square tests for categorical variables. Potential risk

factors for stroke were examined using univariate logistic regression analyses. All statistically

significant risk factors (p < 0.05) identified with univariate analyses were included in a

multivariate analysis to determine which factors were most strongly associated with stroke.

When there were multiple variables with high multicollinearity, the variables regarding the

presence of diseases were chosen as representative variables. Multivariate adjusted odds ratios

(aOR) for stroke were also examined by comparing data from the OAG and non-OAG groups

that had been stratified by hypertension and diabetes status. All statistical tests were

performed using PASW Statistics 18 (SPSS, Inc., Chicago, IL).

Results

A total of 25,534 subjects participated in the 2010-2012 KNHANES. Of these, 13,661

participants who were ≥40 years old were initially included in the present study. After
applying the inclusion criteria, 1702 participants who did not complete the health interview
and ophthalmic examination, including qualifying fundus photography, were excluded.
Ultimately, 11,959 subjects who were at least 40 years old and had a qualifying fundus
photograph in at least one eye were included in this study. In this study, 4,633 subjects
underwent the FDT test, which was performed based on the KNHANES protocol.
The 2010-2012 prevalence of open-angle glaucoma in Korea

The overall prevalence of OAG was 3.9% (95% CI: 3.5%–4.3%) in the 11,959 included

participants. A total of 338 participants were in the stroke group, which had an OAG

prevalence of 8.5% (95% CI: 5.5%–13.1%). This was significantly higher than that in the

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 non-stroke group (3.8% [95% CI: 3.4–4.3], n = 11,621 participants, p < 0.001).

Subject characteristics in the stroke and non-stroke groups

Table 1 summarizes the subject characteristics in the stroke and non-stroke groups. Patients

who had experienced a stroke were more likely to be older (p < 0.001) and have hypertension

(69.6% vs. 26.7%, p < 0.001), diabetes (25.1% vs. 9.7%, p < 0.001), coronary arterial disease

(10.8% vs. 3.3%, p < 0.001), dyslipidemia (31.3% vs. 13.2%, p < 0.001), and OAG (8.5% vs.

3.8%, p < 0.001) than patients who had not experienced a stroke. Patients who had not had a

stroke (22.5%) were more likely to be current smokers than patients who had experienced a

stroke (15.7%, p = 0.030).

Risk factors associated with stroke in Korean patients

Table 2 presents the risk factors for stroke, as determined by univariate and multivariate

logistic regression analyses. Multivariate analysis revealed that older age (Table 2) and the

presence of hypertension (aOR = 2.955, 95% CI: 2.139–4.082, p < 0.001), diabetes (aOR =

1.416, 95% CI: 1.012–1.981, p = 0.043), and dyslipidemia (aOR = 1.609, 95% CI: 1.181–

2.192, p = 0.003) were significantly associated with stroke. Additionally, the presence of

OAG tended to be associated with stroke (aOR = 1.629, 95% CI: 0.994–2.670), but the

increased risk was not statistically significant (p = 0.053).

Effect of comorbidities on the association between stroke and open-angle glaucoma

Figure 1 shows the prevalence of stroke when other systemic conditions were present. The

prevalence of stroke was 5.1%–6.9% in patients with only systemic disease and 6.6%–15.9%

in patients with systemic disease and OAG. Table 3 shows the aORs for stroke (obtained

using multivariate logistic regression) when each systemic disease was examined in the

presence of OAG. The aORs were greater when OAG and systemic disease were present than

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 when systemic disease was present on its own.

Patients were also striated into hypertension and diabetes groups (Table 4). Interestingly,

OAG was significantly associated with stroke in patients with hypertension (aOR = 2.059, 95%

CI: 1.186-3.575, p = 0.010), but not in patients without hypertension (aOR = 0.668, 95% CI:

0.202-2.206, p = 0.508). Analyses of patients with (aOR = 2.649, 95% CI = 1.047-6.703, p =

0.040) and without (aOR = 1.357, 95% CI: 0.777-2.370, p = 0.283) diabetes showed similar

results.

Discussion

In this large, nation-wide, population-based study, although there was no overall statistical

significance, we found that patients with OAG had an increased risk of stroke, especially

those with systemic comorbidities, including hypertension and diabetes. Stratified analyses

revealed an association between OAG and stroke in patients with hypertension and diabetes.

We found an overall OAG prevalence of 3.9% in the entire study population, which is

consistent with previous epidemiological studies in East Asia.18-20 However, analysis of

earlier KNHANES data (2008-2011) revealed an overall OAG prevalence of 4.7% (95% CI:

4.2%–5.1%).9 The current study showed a lower OAG prevalence even though similar data

and diagnostic criteria were used. This was not surprising because the earlier KNHANES

study classified participants into OAG and non-glaucomatous groups. However, some

subjects were excluded from the analyses because they did not qualify for either group (e.g.,

subjects who were glaucoma suspects in only one eye or who had angle closure glaucoma),

artificially elevating OAG prevalence. Our study included these participants in prevalence

calculations.

Our multivariate logistic regression analyses revealed that the risk of stroke was higher in

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 patients with systemic disease when OAG was also present. Furthermore, OAG and stroke

were more strongly associated in participants with hypertension and diabetes than in

participants without these conditions. These results suggest that OAG and stroke share

common pathophysiological mechanisms and provide support for a vascular OAG origin. It

may be that vascular changes occur in a similar manner in both the optic nerve head and

brain.4, 21 The anatomy of the posterior ciliary artery (PCA), the main blood supply to the

optic nerve head, supports this theory because the PCA branches off of the ophthalmic artery,

which branches off of the internal carotid artery.22 It may be that pathological changes in

optic nerve head vessels are indicative of similar changes in brain vessels.12 Furthermore,

both OAG and stroke have associations with vascular abnormalities, including narrowing of

retinal arterioles23-28 and pathological changes of the carotid artery.29, 30 Lastly, OAG and

stroke share several systemic risk factors. Large population-based studies have demonstrated

associations between OAG and diabetes,6, 7 hypertension,8-11 stroke,11 and dyslipidemia,11 all

of which have been associated with strokes.31

However, the association of stroke with OAG was statistically significant only when the

patient had hypertension or diabetes as comorbidities. This means that stroke is not directly

related to OAG. Rather, stroke could be associated only in the condition of having these

systemic comorbidities. This possibility should not be excluded in interpreting the result of

our study.

A cohort study previously evaluated the relationship between OAG and stroke in Taiwan

over a 5-year follow-up period.13 Some of our results are consistent with the Taiwanese study

findings, but other are not. The Taiwanese study only used the International Classification of

Diseases codes to determine the presence/absence of OAG.13 In contrast, we used clinical

data obtained with standardized examinations (e.g., glaucomatous optic neuropathy and

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 visual field testing results) to determine OAG status, as was previously done in a large

epidemiological study.17 Additionally, the Taiwanese study only included subjects who had

National Health Insurance and had visited the hospital. However, inclusion in our nation-

wide study population was independent of hospital visits.

Interestingly, patients without a history of stroke were more likely to be current smokers

than patients who had experienced a stroke (22.5 % vs 15.7%, p = 0.030). The patients with a

history of stroke were more likely to stop smoking following the stroke episode, which may

explain our results.

Our study had several limitations. First, the diagnosis of glaucoma was based only on non-
mydriatic fundus photographs and single FDT results, which may be a source of
misclassification. The statistically determined cut-offs as proposed by ISGEO may result in
misclassification as it is well known that there is a large overlap in CD ratio between glaucoma
patients and normal subjects. Although FDT perimetry has been used in many
epidemiological studies,9, 15, 16, 32, 33 the validity of FDT for the ISGEO criteria is unclear, and
FDT perimetry lacks both sensitivity and specificity as a means of screening for glaucoma.34
The use of the Van Herick test in place of gonioscopy may be another limitation. These issues
should be considered when interpreting our findings. Second, the cross-sectional analyses do
not provide information on temporal or causal relationships between OAG and stroke. A
previous cohort study performed in Taiwan examined stroke development over 5 years,13 but
we only examined stroke prevalence in our study population. Future studies are needed on
large cohorts to evaluate both the development and overall incidence of stroke. Third, our
non-stroke control group was not age- or gender-matched to our stroke group because data
were obtained from an already existing data set. Age in particular is a very important risk
factor for the pathologies evaluated in this study and this factor could influence the results.
The age distribution was significantly different between the stroke and non-stroke groups;
after the multivariate analysis, age still remained as an important associating factor. Deeper
statistical analysis to rule out the influence of age with age-matched samples should be
carried out. Fourth, the presence of stroke, hypertension, diabetes, coronary artery disease,
and dyslipidemia were self-reported by patients. Therefore, a recall bias was likely introduced
into our data. Fifth, since the presence of hypertension was determined using self-reported
questionnaire rather than actual blood pressure data, it may be difficult to explore the role of
anti-hypertension treatments as it relates to its impact on the blood pressure level and by that
the potential role of blood pressure status as a risk factor for OAG and as a confounder of the
OAG-stroke relationship. Sixth, we were unable to differentiate between ischemic or
hemorrhagic stroke. Knowing which type of stroke is associated with OAG would offer
further clues about this relationship. Seventh, using FDT with one sector abnormal as part of
the OAG definition may be problematic because stroke could cause a visual field defect that
might be confused with OAG. Lastly, our study population was made up of all Korean
subjects, limiting the application of our results to other racial/ethnic groups. Because the
prevalence of OAG with normal IOP (known as normal tension glaucoma) is high in Koreans,

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 the IOP criteria for additional FDT examination (IOP of ≥ 22 mmHg) might not be suitable
for screening for all OAG.
However, our study has the notable strength of being able to investigate the relationship

between OAG and stroke in a large, population-based sample by examining nationwide data.

Although there was no overall statistical significance, we found that patients with OAG were

at an increased risk of stroke, especially those with systemic comorbidities, including

hypertension and diabetes. Our results may provide insight on the underlying mechanisms of

OAG and their association with stroke development.

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 References

1. Congdon NG, Friedman DS, Lietman T. Important causes of visual impairment in the world today.
JAMA 2003;290:2057-2060.
2. Quigley HA, Broman AT. The number of people with glaucoma worldwide in 2010 and 2020. Br J
Ophthalmol 2006;90:262-267.
3. Weinreb RN, Aung T, Medeiros FA. The pathophysiology and treatment of glaucoma: a review. JAMA
2014;311:1901-1911.
4. Flammer J, Mozaffarieh M. What is the present pathogenetic concept of glaucomatous optic
neuropathy? Surv Ophthalmol 2007;52 Suppl 2:S162-173.
5. Flammer J, Orgul S, Costa VP, et al. The impact of ocular blood flow in glaucoma. Prog Retin Eye Res
2002;21:359-393.
6. Chopra V, Varma R, Francis BA, et al. Type 2 diabetes mellitus and the risk of open-angle glaucoma
the Los Angeles Latino Eye Study. Ophthalmology 2008;115:227-232 e221.
7. Mitchell P, Smith W, Chey T, Healey PR. Open-angle glaucoma and diabetes: the Blue Mountains eye
study, Australia. Ophthalmology 1997;104:712-718.
8. Dielemans I, Vingerling JR, Algra D, et al. Primary open-angle glaucoma, intraocular pressure, and
systemic blood pressure in the general elderly population. The Rotterdam Study. Ophthalmology 1995;102:54-
60.
9. Kim KE, Kim MJ, Park KH, et al. Prevalence, Awareness, and Risk Factors of Primary Open-Angle
Glaucoma: Korea National Health and Nutrition Examination Survey 2008-2011. Ophthalmology
2016;123:532-541.
10. Mitchell P, Lee AJ, Rochtchina E, Wang JJ. Open-angle glaucoma and systemic hypertension: the blue
mountains eye study. J Glaucoma 2004;13:319-326.
11. Newman-Casey PA, Talwar N, Nan B, et al. The relationship between components of metabolic
syndrome and open-angle glaucoma. Ophthalmology 2011;118:1318-1326.
12. Belzunce A, Casellas M. [Vascular risk factors in primary open angle glaucoma]. An Sist Sanit Navar
2004;27:335-344.
13. Ho JD, Hu CC, Lin HC. Open-angle glaucoma and the risk of stroke development: a 5-year
population-based follow-up study. Stroke 2009;40:2685-2690.
14. Klein R, Klein BE, Moss SE, Cruickshanks KJ. Association of ocular disease and mortality in a
diabetic population. Arch Ophthalmol 1999;117:1487-1495.
15. Yoon KC, Mun GH, Kim SD, et al. Prevalence of eye diseases in South Korea: data from the Korea
National Health and Nutrition Examination Survey 2008-2009. Korean J Ophthalmol 2011;25:421-433.
16. Yoon KC, Choi W, Lee HS, et al. An Overview of Ophthalmologic Survey Methodology in the 2008-
2015 Korean National Health and Nutrition Examination Surveys. Korean J Ophthalmol 2015;29:359-367.
17. Foster PJ, Buhrmann R, Quigley HA, Johnson GJ. The definition and classification of glaucoma in
prevalence surveys. Br J Ophthalmol 2002;86:238-242.
18. Wang YX, Xu L, Yang H, Jonas JB. Prevalence of glaucoma in North China: the Beijing Eye Study.
Am J Ophthalmol 2010;150:917-924.
19. Kim CS, Seong GJ, Lee NH, et al. Prevalence of primary open-angle glaucoma in central South Korea
the Namil study. Ophthalmology 2011;118:1024-1030.
20. Iwase A, Suzuki Y, Araie M, et al. The prevalence of primary open-angle glaucoma in Japanese: the
Tajimi Study. Ophthalmology 2004;111:1641-1648.
21. Gasser P, Flammer J. Blood-cell velocity in the nailfold capillaries of patients with normal-tension and
high-tension glaucoma. Am J Ophthalmol 1991;111:585-588.
22. Hayreh SS. The blood supply of the optic nerve head and the evaluation of it - myth and reality. Prog
Retin Eye Res 2001;20:563-593.
23. Wong TY, Kamineni A, Klein R, et al. Quantitative retinal venular caliber and risk of cardiovascular
disease in older persons: the cardiovascular health study. Arch Intern Med 2006;166:2388-2394.

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 24. Wong TY, Klein R, Couper DJ, et al. Retinal microvascular abnormalities and incident stroke: the
Atherosclerosis Risk in Communities Study. Lancet 2001;358:1134-1140.
25. Cooper LS, Wong TY, Klein R, et al. Retinal microvascular abnormalities and MRI-defined
subclinical cerebral infarction: the Atherosclerosis Risk in Communities Study. Stroke 2006;37:82-86.
26. Mitchell P, Leung H, Wang JJ, et al. Retinal vessel diameter and open-angle glaucoma: the Blue
Mountains Eye Study. Ophthalmology 2005;112:245-250.
27. Wang S, Xu L, Wang Y, et al. Retinal vessel diameter in normal and glaucomatous eyes: the Beijing
eye study. Clin Experiment Ophthalmol 2007;35:800-807.
28. Lee JY, Yoo C, Park JH, Kim YY. Retinal vessel diameter in young patients with open-angle glaucoma:
comparison between high-tension and normal-tension glaucoma. Acta Ophthalmol 2012;90:e570-571.
29. Gupta A, Baradaran H, Schweitzer AD, et al. Carotid plaque MRI and stroke risk: a systematic review
and meta-analysis. Stroke 2013;44:3071-3077.
30. Marmion VJ, Aldoori MI, Woodcock JP, Stephenson J. A cohort study of duplex Doppler
examinations of the carotid artery in primary open angle glaucoma. JRSM Open 2014;5:2054270414527933.
31. Elkind MS, Sacco RL. Stroke risk factors and stroke prevention. Semin Neurol 1998;18:429-440.
32. Iwase A, Tomidokoro A, Araie M, et al. Performance of frequency-doubling technology perimetry in a
population-based prevalence survey of glaucoma: the Tajimi study. Ophthalmology 2007;114:27-32.
33. Terry AL, Paulose-Ram R, Tilert TJ, et al. The methodology of visual field testing with frequency
doubling technology in the National Health and Nutrition Examination Survey, 2005-2006. Ophthalmic
Epidemiol 2010;17:411-421.
34. Boland MV, Gupta P, Ko F, et al. Evaluation of Frequency-Doubling Technology Perimetry as a
Means of Screening for Glaucoma and Other Eye Diseases Using the National Health and Nutrition
Examination Survey. JAMA Ophthalmol 2016;134:57-62.

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 Figure Legends

Figure 1. Prevalence of stroke in the presence/absence of systemic comorbidities. The

incidence of stroke tended to increase when both open-angle glaucoma (OAG) and systemic

disease were present.

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 Table 1. Characteristics of Korean subjects with and without a history of stroke. All subjects
were over 40 years of age.
Stroke (n=338) Non-stroke (n=11621) p-value

Gender (M) (%) 52.0 (45.4-58.5) 47.7 (46.9-48.6) 0.212

Age (years) <0.001

40-49 6.3 (3.5-11.1) 37.2 (35.5-38.8)

50-59 20.7 (15.1-27.6) 31.0 (29.8-32.3)

60-69 33.8 (27.7-40.6) 17.6 (16.7-18.5)

70-79 29.4 (23.9-35.5) 11.6 (10.9-12.4)

80- 9.8 (6.7-14.2) 2.6 (2.2-3.0)

Hypertension

Presence 69.6 (63.5-75.0) 26.7 (25.6-27.8) <0.001

Duration (years) 10.9 ( 9.5-12.3) 7.4 (7.1-7.8) <0.001

Under treatment 65.3 (59.1-71.1) 23.6 (22.5-24.7) <0.001

Diabetes

Presence 25.1 (19.9-31.3) 9.7 (9.1-10.4) <0.001

Duration (years) 10.8 (8.7-12.8) 8.0 (7.5-8.5) <0.001

Insulin treatment 3.0 (1.5-5.9) 0.8 (0.7-1.1) <0.001

Coronary arterial dz. (%) 10.8 (7.4-15.5) 3.3 (2.9-3.7) <0.001

Dyslipidemia (%) 31.3 (25.5-37.8) 13.2 (12.4-14.0) <0.001

Smoking

Ever smoker (%) 49.0 (42.6-55.4) 43.5 (42.5-44.6) 0.087

Current smoker (%) 15.7 (11.1-21.7) 22.5 (21.5-23.6) 0.030

Ex smoker (%) 33.9 (27.5-40.8) 22.8 (21.9-23.8) 0.001

BMI 24.27 (23.89-24.64) 24.05 (23.97-24.13) 0.269

OAG (%) 8.5 (5.5-13.1) 3.8 (3.4-4.3) <0.001

HTN; hypertension, DM; diabetes, BMI; body mass index, OAG; open angle glaucoma
Note: Values are expressed as mean (standard deviation) or weighted percentage (95%
confidence interval).
*A person who has smoked ≥100 cigarettes during the course of his life.

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 Table 2. Possible risk factors for stroke in the older (>40 years of age) Korean population.

Univariate Multivariate

OR (95% CI) p-value aOR (95% CI) p-value

Gender (M) 1.186 (0.907-1.551) 0.213

Age (years)

40-49 (Ref) (Ref)

50-59 3.951 (1.979-7.889) <0.001 2.723 (1.331-5.572) 0.006

60-69 11.407 (5.959-21.836) <0.001 5.708 (2.805-11.612) <0.001

70-79 14.986 (7.870-28.534) <0.001 6.790 (3.375-13.660) <0.001

80- 22.620 (10.919-46.858) <0.001 10.655 (4.817-23.566) <0.001

Hypertension

Presence 6.284 (4.766-8.285) <0.001 2.955 (2.139-4.082) <0.001

Duration† 1.047 (1.031-1.064) <0.001

Under treatment† 1.998 (1.002-3.983) 0.050

Diabetes

Presence 3.107 (2.278-4.236) <0.001 1.416 (1.012-1.981) 0.043

Duration† 1.041 (1.014-1.068) 0.003

Insulin Treatment† 1.340 (0.592-3.035) 0.482

Coronary arterial dz. 3.540 (2.279-5.499) <0.001 1.590 (0.999-2.533) 0.051

Dyslipidemia 2.998 (2.234-4.025) <0.001 1.609 (1.181-2.192) 0.003

Smoking

Ever smoker 1.258 (0.966-1.638) 0.088

Current smoker 0.643 (0.428-0.965) 0.033 0.981 (0.637-1.511) 0.930

Ex smoker* 2.234 (1.357-3.677) 0.002

BMI 1.021 (0.985-1.058) 0.258

OAG 2.683 (1.642-4.384) <0.001 1.629 (0.994-2.670) 0.053

OR; odds ratio, CI; confidence interval, aOR; adjusted odds ratio, BMI; body mass index, OAG; open angle
glaucoma
†
Considering that these variables showed high multicollinearity with pre-existing variables, the variables
regarding the presence of diseases were chosen as representative variables.
*
This variable showed high multicollinearity with current smokers, the current smoker was chosen as
representative variable.

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 Table 3. Association of stroke with open-angle glaucoma and systemic disease.

Combination of components Stroke Non-stroke Adjusted OR*

(n=338) (n=11621) (95 % CI) p-value

HTN + OAG 7.5 (4.6-12.0) 1.5 (1.3-1.7) 4.512 (2.396-8.496) <0.001

DM + OAG 3.4 (1.5-7.6) 0.6 (0.4-0.8) 2.882 (1.172-7.084) 0.021

CAD + OAG 0.6 (0.2-2.0) 0.2 (0.1-0.3) 1.515 (0.443-5.183) 0.507

Dyslipidemia + OAG 3.4 (1.6-7.3) 0.7 (0.6-0.9) 2.814 (1.253-6.321) 0.012

HTN + DM + OAG 3.2 (1.3-7.5) 0.4 (0.3-0.5) 9.649 (3.090-30.131) <0.001

HTN; hypertension. DM; diabetes, CAD; coronary arterial disease, OAG; open angle

glaucoma

Note: Values are expressed as mean (standard deviation) or weighted percentage (95%

confidence interval).

*Multivariate logistic regression analysis – the variables showed p <0.05 at the univariate

logistic regression analyses were used for adjustment. If the variables had high multicollinearity,

only representative variables were used for multivariate analysis.

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
 Table 4. Multivariate analysis adjusted odds ratios for stroke in subjects with and without

open-angle glaucoma. Subjects were striated into hypertension and diabetes subgroups.

Stroke

Pt with HTN Pt without HTN Pt with DM Pt without DM

Adjusted OR* Adjusted OR* Adjusted OR† Adjusted OR†

OAG 2.059 0.668 2.649 1.357

(1.186-3.575) (0.202-2.206) (1.047-6.703) (0.777-2.370)

Non-OAG (Ref) 1.00 1.00 1.00 1.00

P-value 0.010 0.508 0.040 0.283

Pt; patients, HTN; hypertension, DM; diabetes, OR; odds ratio, OAG; open angle glaucoma

*adjusted for gender, age, diabetes, dyslipidemia and coronary arterial disease

† adjusted for gender, age, hypertension, dyslipidemia and coronary arterial disease

Considering that the variables including HTN treatment, HTN duration, DM duration and

insulin treatment showed high multicollinearity with pre-existing variables, the variables

regarding the presence of diseases were chosen as representative variables.

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
olters Kluwer Health, Inc. Unauthorized reproduction of the