Prevalence and Systemic Risk Factors for Retinal Vein Occlusion in a General Japanese Population: The Hisayama Study

Abstract

Purpose.: To examine the prevalence of retinal vein occlusion (RVO) and its systemic relevant factors in a general Japanese population aged 40 years or older.

Methods.: In 1998, 1775 Hisayama residents consented to participate in the study. Each participant underwent a comprehensive examination that included ophthalmic testing. RVO was determined by grading color fundus photographs. Logistic regression analysis was performed to determine risk factors for RVO.

Results.: Of the 1775 subjects examined, 38 had RVO. The prevalence of RVO was 2.1% (2.0% for branch RVO and 0.2% for central RVO). After adjustment for age and sex, it was found that systolic and diastolic blood pressures, hypertension, and hematocrit were significantly associated with RVO. In multivariate analysis, age (per 10 years; odds ratio [OR], 1.47; 95% confidence interval [CI], 1.04–2.08), hypertension (OR, 4.25; 95% CI, 1.82–9.94), and hematocrit (per 10%; OR, 3.09; 95% CI, 1.10–1.22) remained independently significant risk factors for RVO. Both high-normal blood pressure and hypertension were significantly associated with RVO. Furthermore, compared with normotensive subjects without high hematocrit, the likelihood of RVO was markedly high in subjects having both high blood pressure and high hematocrit (age- and sex-adjusted OR, 36.0; 95% CI, 4.43–292).

Conclusions.: The findings suggest that the prevalence of RVO is higher in the Japanese than in other Asians or Caucasians and that older age, higher hematocrit, and both hypertension and high-normal blood pressure are significant risk factors for RVO in the Japanese.

Retinal vein occlusion (RVO) is a cause of significant loss of vision in elderly populations in developed countries. ¹ Despite the magnitude of this problem, the available treatment options remain limited. ^2,3 Furthermore, RVO has also been associated with increased risk of cardiovascular disease. ^4–6 In developing measures to prevent this disease, it is thus very important to determine the prevalence of RVO and to identify its systemic risk factors. To date, several population-based studies, ^6–11 mostly in Caucasian populations, have provided valuable information on the prevalence and systemic risk factors for RVO. These include hypertension, ^6–11 diabetes, ¹⁰ smoking habits, ¹⁰ dyslipidemia, ^7,9 and a history of angina. ⁹ However, there have been only a limited number of population-based epidemiologic studies on RVO in Japanese and other Asians. ^9,11,12  

The purpose of this article was to examine the prevalence of RVO and its systemic relevant factors in a cross-sectional study of a general Japanese population. 

Methods

Study Population

The Hisayama Study is an ongoing long-term prospective cohort study on cardiovascular disease and its risk factors in Hisayama, a town adjoining Fukuoka City, a metropolitan area in southern Japan. ^13,14 As a part of the follow-up survey, we performed a cross-sectional examination, including an eye examination, of Hisayama residents aged 40 years or older in 1998. ¹⁵ Among 4187 residents in that age group, 1775 (42.4%; 688 men and 1087 women) were enrolled in the present study. 

Ophthalmic Examination and Definition of RVO

The methods used for the ophthalmic examination have been published in detail. ¹⁵ Briefly, each participant underwent a comprehensive ophthalmic examination, including a stereoscopic fundus examination with indirect ophthalmoscopy and examination with a slit-lamp biomicroscope with a superfield lens (Volk, Mentor, OH), after pupil dilation with 1.0% tropicamide and 5% phenylephrine. Fundus photographs (45°) were taken of both eyes of each participant with a nonmydriatic fundus camera (TRC NW-5; Topcon, Tokyo, Japan) and slide film (Fujichrome Sensia II; Fujifilm, Tokyo, Japan). We photographed one field, centered at a point midway between the temporal edge of the optic disc and the fovea in both eyes. The presence of RVO was determined based on the grading of fundus examinations by indirect ophthalmoscopy and slit lamp and the color fundus photographs. All photographs were evaluated by retinal specialists (MY and TI) who were masked to the participants' data. The presence or absence of central or branch RVO was defined according to a standardized protocol. ^6,10,16 Recent central RVO was characterized by retinal edema, optic disc hyperemia or edema, scattered superficial and deep retinal hemorrhages, and venous dilation. Old central RVOs were characterized by occluded and sheathed retinal veins or vascular anastomosis at the optic disc. Branch RVOs involved a more localized area of the retina in the sector of the obstructed venule and were characterized by scattered superficial and deep retinal hemorrhages, venous dilation, intraretinal microvascular abnormalities, and occluded and sheathed retinal venules. The presence of any RVO was defined as the presence of branch or central RVO in either eye. 

Data Collection

Information on smoking habits, alcohol intake, and regular exercise during leisure time was obtained by trained interviewers using a standard questionnaire. Smoking habits and alcohol intake were classified as either current use or not. Those subjects who engaged in sports or other forms of exertion three or more times per week during their leisure time were designated as the regular exercise group. The questionnaire also investigated history of cardiovascular disease, including stroke and coronary heart disease. 

Blood pressure was measured three times in the sitting position after the subject had rested for at least 5 minutes. The average of the three measurements was used for the analysis. According to the 2007 European Society of Hypertension (ESH) and European Society of Cardiology (ESC) Practice Guidelines, ¹⁷ blood pressure levels were categorized as follows: optimal (systolic <120 mm Hg and diastolic <80 mm Hg), normal (120–129/80 to 84 mm Hg), high-normal (130–139/85 to 89 mm Hg), and hypertension (≥140/≥90 mm Hg or current use of antihypertensive medication). 

Plasma glucose levels were determined by the glucose-oxidase method, and diabetes mellitus was defined by a 75-g oral glucose tolerance test or by fasting (≥7.0 mM) or postprandial (≥11.1 mM) blood glucose level, or by the use of hypoglycemic agents. Serum total cholesterol, high-density lipoprotein (HDL) cholesterol, and triglyceride levels were determined enzymatically by using an autoanalyzer (TBA-80S; Toshiba Inc., Tokyo, Japan). White blood cell (WBC) and platelet counts and hematocrit levels were determined with a cell counter (STKS; Coulter Inc., Hialeah, FL). ECG abnormalities were defined as left ventricular hypertrophy (Minnesota code, ¹⁸ 3-1) or ST depression (4-1, 2, 3). Body height and weight were measured in light clothing without shoes, and the body mass index (in kilograms per square meter) was calculated. 

Statistical Methods

We considered the following 18 possible risk factors for RVO: age, sex, systolic and diastolic blood pressures, hypertension, total cholesterol, HDL cholesterol, triglycerides, body mass index, diabetes mellitus, WBC count, platelet count, hematocrit, ECG abnormalities, history of cardiovascular disease, smoking habits, alcohol intake, and regular exercise. Mean values were compared by Student's t-test, and frequencies by the χ² test. We estimated the age- and sex-adjusted and multivariate-adjusted odds ratio (OR) and 95% confidence interval (CI) for each potential risk factor by using logistic regression analysis (SAS software; SAS Institute, Cary, NC ¹⁹ ). A two-sided P < 0.05 was considered statistically significant. 

Ethical Considerations

This study was approved by the Human Ethics Review Committee of Kyushu University Graduate School of Medical Sciences and was performed in accordance with the Declaration of Helsinki. The study subjects provided written informed consent to participate in the study. 

Results

Table 1 shows the mean values or frequencies of potential risk factors according to the presence or absence of RVO. The geometric mean values and 95% prediction intervals of triglycerides are shown because of the skewed distribution. Subjects with RVO were older than those without RVO. Subjects with RVO had higher mean systolic and diastolic blood pressures and hematocrits, as well as a higher frequency of hypertension, whereas those without RVO had a lower mean platelet count. 

Table 1.

 

Mean Values or Frequencies of Risk Factors by Status of Retinal Vein Occlusion

Variable	Non-RVO (n = 1736)	RVO (n = 38)	P
Age, y	62 ± 11	67 ± 7	0.002
Sex (men), %	38.5	50.0	0.15
Systolic blood pressure, mm Hg	133 ± 21	147 ± 18	<0.0001
Diastolic blood pressure, mm Hg	77 ± 11	81 ± 10	0.02
Hypertension, %	43.7	81.6	<0.0001
Total cholesterol, mmol/L	5.3 ± 0.9	5.4 ± 1.1	0.73
High-density lipoprotein cholesterol, mmol/L	1.5 ± 0.4	1.6 ± 0.4	0.16
Triglycerides, mmol/L	1.20 (0.59–2.98)	1.02 (0.51–2.19)	0.14
Body mass index, kg/m2	23.1 ± 3.1	23.2 ± 3.5	0.77
Diabetes, %	12.6	10.5	0.70
White blood cells, ×103/mm3	5.8 ± 1.5	6.1 ± 1.6	0.15
Platelets, ×104/mm3	21.9 ± 5.2	19.8 ± 6.0	0.02
Hematocrit, %	40.1 ± 4.1	41.6 ± 3.6	0.03
ECG abnormalities, %	17.1	29.0	0.06
History of cardiovascular disease, %	2.7	5.3	0.42
Smoking habit (yes), %	17.5	18.4	0.88
Alcohol intake (yes), %	36.5	44.7	0.30
Regular exercise (yes), %	16.7	23.7	0.51

 Data are expressed as the mean ± SD or percentages. Geometric mean value and 95% prediction interval of triglycerides are shown because of the skewed distribution. RVO, retinal vein occlusion.

The age-specific prevalences of RVO are shown by sex in Table 2. Of the 1775 subjects examined, 38 (2.1%) had RVO. Of the subjects with RVO, 35 (92.1%) had branch RVO. The prevalence of branch RVO was slightly but not significantly higher in the men than in the women (2.3% vs. 1.8%). Central RVO was observed only in the men (0.4%). The prevalence of all RVO significantly increased with advancing age in all the subjects (P _trend = 0.005), whereas the prevalence of branch RVO significantly increased with advancing age only in the men (P _trend = 0.01). 

Table 2.

 

Age-Specific Prevalence of RVO by Sex

Age Range, y	Men			Women			All Subjects
	Subjects, n	Branch RVO, n (%)	Central RVO, n (%)	Subjects, n	Branch RVO, n (%)	Central RVO, n (%)	Subjects, n	Branch RVO, n (%)	Central RVO, n (%)	All RVO, n (%)
40–49	92	0 (0.0)	0 (0.0)	201	0 (0.0)	0 (0.0)	293	0 (0.0)	0 (0.0)	0 (0.0)
50–59	154	2 (1.3)	0 (0.0)	284	5 (1.8)	0 (0.0)	438	7 (1.6)	0 (0.0)	7 (1.6)
60–69	231	5 (2.2)	3 (1.3)	335	10 (3.0)	0 (0.0)	566	15 (2.7)	3 (0.5)	18 (3.2)
70–79	178	7 (3.9)	0 (0.0)	212	2 (0.9)	0 (0.0)	390	9 (2.3)	0 (0.0)	9 (2.3)
80+	33	2 (6.1)	0 (0.0)	55	2 (3.6)	0 (0.0)	88	4 (4.6)	0 (0.0)	4 (4.6)
Total	688	16 (2.3)	3 (0.4)	1087	19 (1.8)	0 (0.0)	1775	35 (2.0)	3 (0.2)	38 (2.1)
P trend		0.01	0.87		0.15			0.005	0.66	0.005

 RVO, retinal vein occlusion.

The results of age- and sex-adjusted and multivariate-adjusted logistic regression analyses of relevant factors for RVO are presented in Table 3. After adjusting for age and sex, we found that systolic and diastolic blood pressures, hypertension, and hematocrit were significantly associated with RVO. In multivariate analysis, age (per 10 years; OR, 1.47; 95% CI, 1.04–2.08), hypertension (OR, 4.25; 95% CI, 1.82– 9.94), and hematocrit (per 10%) (OR, 3.09; 95% CI, 1.10–1.22) remained independently significant relevant factors for RVO. 

Table 3.

 

Age- and Sex-Adjusted and Multivariate-Adjusted OR of Relevant Factors of RVO

Association	Age- and Sex-Adjusted		Multivariate-Adjusted
	OR	95% CI	OR	95% CI
Age, per 10 years			1.47*	1.04–2.08
Sex (men), %			0.93	0.42–2.07
Systolic blood pressure, per 10 mm Hg	1.23†	1.07–1.41
Diastolic blood pressure, per 10 mm Hg	1.46*	1.09–1.97
Hypertension	4.53†	1.94–10.6	4.25†	1.82–9.94
Total cholesterol, per 1 mmol/L	1.20	0.83–1.74
High-density lipoprotein cholesterol, per 1 mmol/L	2.22	0.94–5.25
Triglycerides, per 1 mmol/L	0.63	0.36–1.10
Body mass index, per 1 kg/m2	1.04	0.94–1.15
Diabetes	0.65	0.23–1.87
White blood cells, per 103/mm3	1.15	0.94–1.40
Platelets, per 104/mm3	0.94	0.88–1.01
Hematocrit, per 10 %	3.09*	1.13–8.46	1.10*	1.00–1.22
ECG abnormalities	1.57	0.76–3.26
History of cardiovascular disease	0.91	0.21–3.91
Smoking habit	0.95	0.39–2.34
Alcohol intake	1.42	0.67–3.01
Regular exercise	1.24	0.58–2.68

 RVO, retinal vein occlusion; OR, odds ratio; CI confidence interval.

*   P < 0.05.

†   P < 0.01.

Table 4 demonstrates the age- and sex-adjusted OR of RVO according to blood pressure levels and quartiles of hematocrit. The age- and sex-adjusted OR of RVO significantly increased with elevated blood pressure levels (P _trend < 0.001). Compared with those with optimal or normal blood pressure, the OR of RVO was significantly higher, not only in the subjects with hypertension (age- and sex-adjusted OR, 11.9; 95% CI, 2.78–50.9), but also in the subjects with high-normal blood pressure (age- and sex-adjusted OR, 6.81; 95% CI, 1.30–35.6). The age- and sex-adjusted OR of RVO also significantly increased with rising hematocrit levels (P _trend = 0.003): the likelihood of RVO was significantly higher in the fourth quartile than in the first (age- and sex-adjusted OR, 6.03; 95% CI, 1.85–19.7). 

Table 4.

 

Age- and Sex-Adjusted OR of RVO According to Blood Pressure Levels and Quartiles of Hematocrit

Risk Factor Level	Subjects, n	Cases, n	Age- and Sex-Adjusted OR (95% CI)	P trend
Blood pressure level
Optimal	469	1	1.00 (reference)	<0.001
Normal	276	1
High-normal	240	5	6.81 (1.30–35.6)*
Hypertension	790	31	11.9 (2.78–50.9)†
Hematocrit
First quartile, <37.7	436	5	1.00 (reference)	0.004
Second quartile, 37.7–39.9	447	7	1.40 (0.44–4.46)
Third quartile, 40.0–42.6	445	8	1.81 (0.58–5.70)
Fourth quartile, ≥42.7	446	18	6.03 (1.85–19.7)*

 RVO, retinal vein occlusion; OR, odds ratio; CI, confidence interval.

*   P < 0.05.

†   P < 0.01.

Further, we examined both the combined and separate effects of high blood pressure and elevated hematocrit levels on RVO in the groups according to the presence or absence of high blood pressure (high normal blood pressure or hypertension) and high hematocrit level (fourth quartile, ≥42.7%). As shown in Table 5, compared with normotensive subjects without high hematocrit, the OR of RVO was significantly increased in subjects with high blood pressure alone (age- and sex-adjusted OR, 11.9; 95% CI, 1.57–90.9), whereas the OR of RVO was slightly but not significantly increased in subjects with high hematocrit alone (age- and sex-adjusted OR, 4.81; 95% CI, 0.28–82.2). Furthermore, the OR of RVO was markedly high in subjects having both high blood pressure and high hematocrit (age- and sex-adjusted OR, 36.0; 95% CI, 4.43–292). However, the interaction between high blood pressure and high hematocrit level was not significant (P = 0.35). 

Table 5.

 

Age- and Sex-Adjusted OR of RVO According to the Presence or Absence of High Blood Pressure and High Hematocrit

	Subjects, n	Cases, n	Age- and Sex-Adjusted OR (95% CI)	P
Normal blood pressure + low hematocrit	595	1	1.00 (reference)
Normal blood pressure + high hematocrit	150	1	4.81 (0.28–82.2)	0.28
High blood pressure + low hematocrit	742	20	11.9 (1.57–90.9)	0.02
High blood pressure + high hematocrit	288	16	36.0 (4.43–292)	<0.01

 Normal blood pressure: optimal + normal; high blood pressure: high normal + hypertension; low hematocrit: first to third quartiles (<42.7%); high hematocrit: fourth quartiles (≥42.7%).

Discussion

In a cross-sectional examination of a general Japanese population, we demonstrated that the prevalence of RVO was 2.1% and that age, high blood pressure, and elevation of hematocrit levels were independent relevant risk factors for RVO. In addition, the likelihood of RVO increased significantly in subjects having both high blood pressure and high hematocrit. 

The prevalence of RVO has also been estimated in several other population-based studies (Table 6). The disease prevalence was reported to be 1.6% in the Blue Mountains Eye Study in Australia ¹⁶ and 1.1% in the Multiethnic Study of Atherosclerosis in the United States. ⁷ A study on a Chinese population, the Beijing Eye Study, reported an RVO prevalence of 1.2%, ¹² and a study of a Malay population, the Singapore Malay Eye Study, reported a prevalence of 0.7%. ⁹ The prevalence of RVO in the present study (2.1%) seemed to be somewhat higher than those in the previous studies. Although the variation in disease prevalence among these studies could be due to differences in the characteristics of subjects and in the methodologies, our findings of a higher prevalence suggest that RVO is more common among the Japanese population than among other Asian or Western populations, since the same grading protocols and RVO definitions were used in most of those studies. ^7,9,12,16 Indeed, some studies have shown racial differences in the prevalence of RVO. ^9,10 The reason for such differences remains uncertain, although genetic or environmental factors could contribute to the discrepancy. 

Table 6.

 

Prevalence of RVO in the Hisyama Study and Other Population-Based Studies

Study	Country	Subjects, n	Age	n (Prevalence %)
Blue Mountains Eye Study 16	Australia	3654	49	59 (1.6)
Multiethnic Study of Atherosclerosis 7	United States	6147	45	65 (1.1)
Beijing Eye Study 12	China	4439	40	58 (1.3)
Singapore Malay Eye Study 9	Singapore	3280	40	22 (0.7)
Hisayama Study 15	Japan	1775	40	38 (2.1)

In the present study, we found that the prevalence of RVO increased significantly with advancing age. The etiology and pathogenesis of RVO are largely unknown. The consistent association with increasing age found in this study is in accordance with the findings in many others, ^6,7,9 confirming the age-related nature of the disease. 

Our data indicated a clear association between hypertension and RVO, which is consistent with clinical knowledge and the findings of other population-based studies. ^{6–8,10–12} Our results also showed that not only hypertension but also high-normal blood pressure was significantly associated with RVO. The Framingham Heart Study indicated that the risk of cardiovascular disease is significantly increased in patients with high-normal blood pressure and higher blood pressure levels. ²⁰ Based on these findings, it may be reasonable to suppose that high-normal blood pressure promotes systemic arteriosclerosis, including retinal vascular changes, and thereby causes RVO. Therefore, subjects with high-normal blood pressure should be considered at high risk for RVO. Strict control of elevated blood pressure may be important in preventing the disease. 

We found that a higher hematocrit level was associated with RVO, independent of age, sex, and hypertension. A previous case–control study also indicated that hematocrit was significantly higher in a branch RVO group than in the control subjects. ²¹ Moreover, another study reported a significantly higher prevalence of elevated hematocrit in subjects with central RVO than in control subjects. ²² RVO is caused by thrombosis of the vein, but the role played by various hematologic abnormalities in its etiology and pathogenesis remains unclear and controversial. It is known that elevated hematocrit increases blood viscosity. ²² Therefore, increased hematocrit may augment the risk of RVO through the increase in blood viscosity. 

The present study showed an extremely increased likelihood of RVO in subjects who had both hypertension and a higher hematocrit level. Although the mechanism underlying this phenomenon is not clearly understood, a possible explanation is that hypertension is a strong risk factor for systemic arteriosclerosis, including retinal arteriosclerosis, ^5,8 and sclerotic arteriolar walls in the retina may compress the underlying veins at arteriovenous crossings, leading to reduced blood flow, which in turn could facilitate the development of a thrombus and downstream venous occlusion. It is therefore speculated that increased hematocrit levels markedly enhance the likelihood of RVO by hyperviscosity in people whose retinal vessel walls have already been damaged by hypertension. 

This study has several limitations. First, we ascertained RVO cases by using one photographic field per eye, whereas in most previous population-based studies, two to six photographic fields were taken per eye. This difference could have resulted in underestimation of the prevalence of RVO if peripheral lesions were overlooked. Second, the number of our RVO cases is relatively small, and therefore the CIs around the prevalence and ORs are very wide. It might be misleading to compare the prevalence in this study with that in other population-based studies, and there is a possibility that the ORs are inflated due to the small samples. The estimates of our study should be interpreted with caution. Third, because of the cross-sectional design of this study, it is still unclear how risk factors are related to the onset of RVO. Further prospective investigation would help to clarify this issue. 

In conclusion, the results of this study suggest that RVO is more common among the Japanese than among other Asians or Caucasians and that older age, higher hematocrit, and not only hypertension but also high-normal blood pressure are risk factors for RVO in the Japanese. In addition, among subjects who have both high blood pressure and higher hematocrit, the likelihood of RVO was substantially increased. Therefore, patients having both high blood pressure and higher hematocrit should be considered a population at high risk for RVO and continued preventive efforts should be made in these patients to reduce the burden of the disease. 

Footnotes

 Disclosure: M. Yasuda, None; Y. Kiyohara, None; S. Arakawa, None; Y. Hata, None; K. Yonemoto, None; Y. Doi, None; M. Iida, None; T. Ishibashi, None