Retinal Vascular Caliber, Blood Pressure, and Cardiovascular Risk Factors in an Asian Population: The Singapore Malay Eye Study

Abstract

purpose. To describe the relationship of retinal vascular caliber with cardiovascular risk factors in an Asian population.

methods. The Singapore Malay Eye Study is a population-based, cross-sectional survey that included 3280 (78.7% response) persons aged 40 to 80 years. Retinal vascular caliber was measured from digital retinal photographs using a validated standardized protocol. Data on major cardiovascular risk factors were collected from all participants.

results. Of the 3019 participants with retinal vascular caliber data available, the mean retinal arteriolar caliber (CRAE) was 139.5 ± 15.7 μm (SD), and mean venular caliber (CRVE) was 219.3 ± 22.2 μm (SD). Smaller retinal arteriolar caliber was associated with higher current mean blood pressure and male sex (P < 0.001 for both). Larger retinal venular caliber was associated with younger age, current cigarette smoking, greater body mass index, higher glycosylated hemoglobin level, and lower HDL cholesterol (P = 0.012 for glycosylated hemoglobin level and P < 0.001 for other risk factors). The association of retinal arteriolar narrowing and blood pressure was stronger in younger people than in older people and in men than in women (P for interaction < 0.001 for both).

conclusions. In this Asian population, smaller retinal arteriolar caliber was associated with hypertension and larger retinal venular caliber with cigarette smoking, dyslipidemia, hyperglycemia, and higher body mass index. The pattern of these associations is similar to that in white populations.

The retina offers a unique opportunity for the direct, noninvasive study of the human microvasculature and its relationship to cardiovascular disease risk factors (e.g., hypertension and diabetes). ^{1 2} Recent developments in retinal imaging techniques have made quantitative measurement of retinal vascular caliber possible. Using such techniques, many large population-based studies have described the relationship of various vascular risk factors and cardiovascular outcomes (e.g., stroke, heart disease) with retinal vascular caliber. ^{3 4 5 6 7 8 9 10 11} Findings in these studies have indicated an important role of microvascular disease in the pathogenesis of major cardiovascular diseases. These data also support potentially novel therapeutic agents that specifically target the microcirculation ¹² and are useful for physicians when considering treatment strategies for cardiovascular diseases. For example, physicians may consider choosing antihypertensive medications that have direct beneficial effects on microvascular structure, which could thereby reduce damage to the microvasculature beyond the reduction resulting from lowered blood pressure (e.g., angiotensin converting enzyme inhibitors). ^{1 12 13}  

However, most studies have been conducted exclusively in whites ^{3 4 7 8 9 14} or predominantly white populations, ⁵ with only one study in an exclusively Asian population. ⁶ There are well-recognized ethnic differences in the relationship between cardiovascular risk factors and large vessel atherosclerosis. ¹⁵ In contrast, there are limited data on associations of cardiovascular risk factors with microvascular disease in nonwhite populations. ⁶ Furthermore, few studies have examined the potential age and sex interactions of these relationships. Some have suggested that the role of microvascular disease may be more prominent in middle-aged persons (age, 43–69 years) than in older people (age, ≥70 years). ¹⁶ For example, in the Blue Mountains Eye Study, the association of higher blood pressure with arteriolar narrowing diminished with increasing age, ¹⁰ suggesting that aging may lead to local changes in vessel wall elasticity and reactivity that prevent the microcirculation from responding to changes brought on by blood pressure. However, whether the microvascular response to other cardiovascular risk factors also diminishes with aging is unknown. There is also a suggestion that retinal arteriolar caliber may be narrower in men than in women, ⁵ which could be due to sex-specific effects of cardiovascular risk factors (e.g., diabetes), ¹⁷ sex hormonal influence, or sex-related structural and functional differences in the vasculature. ¹⁸  

The purposes of this study were to describe the distribution of retinal vascular caliber and its associations with cardiovascular risk factors in an Asian population. We also examined potential age and sex interactions of retinal vascular caliber with blood pressure. 

Methods

Study Population

The Singapore Malay Eye Study (SiMES) is a population-based, cross-sectional study of 3280 Malay men and women, aged 40 to 80 years, living in Singapore. ¹⁹ The study design and details of sample recruitment have been described elsewhere. ¹⁹ In brief, an age-stratified, random-sampling procedure was used to select people of Malay ethnicity, as defined by Singapore Census criteria, aged 40 to 80 years living in the southwestern part of Singapore. From the sampling frame of 4168 persons, 3280 Malays were examined from August 2004 to June 2006, giving an overall response rate of 78.7%. Nonparticipants were older but did not differ by sex or possession of a home telephone (data not shown). The study followed the principles of the Declaration of Helsinki, and ethics approval was granted by the Institutional Review Board of the Singapore Eye Research Institute. Written informed consent in either Malay or English was obtained from each participant. 

Digital fundus photography was performed for 3264 (99.5% of the total study population) participants, of which 3193 (97.8%) had retinal photographs suitable for measurement of retinal vascular caliber. Of these, we excluded 174 participants with ungradable photographs of either eye, leaving 3019 participants who were considered eligible for this analysis (92.0% of the total 3280 participants). 

Measurement of Retinal Vascular Caliber

All participants had their pupils dilated except those with known allergy to mydriatic eye drops and those with closed anterior chamber angles and intraocular pressure >21 mm Hg. ¹⁹ After pupil dilation with tropicamide 1% and phenylephrine hydrochloride 2.5%, color fundus photographs were taken of both eyes in each participant, with a digital retinal camera (CR-DGi with a 10D SLR backing; Canon, Tokyo, Japan). Two retinal images of each eye were obtained, one centered on the optic disc and another centered on the fovea (Early Treatment of Diabetic Retinopathy Study [ETDRS] standard fields 1 and 2). 

Retinal vascular caliber was measured at the Retinal Vascular Imaging Centre, the University of Melbourne with image-analysis software, as used in previous population-based studies. ^{20 21} A trained grader, masked to participant characteristics, performed the vessel measurements on the optic disc–centered image of the right eye for most participants and of the left eye in those without gradable right eye images. The largest six arterioles and venules coursing through a zone between 0.5 and 1 disc diameter from the optic disc margin were measured. Images were considered ungradable if the largest six arterioles and venules could not be measured. ²² Estimates were summarized as central arteriolar equivalent (CRAE) and central retinal venular equivalent (CRVE), representing the average diameter of arterioles and venules of the eye, respectively, by using a modification of the Parr-Hubbard formulas ²¹ as described by Knudtson et al. ²² These revised formulas have demonstrated clear superiority over the previous Parr-Hubbard formulas, providing more robust measurements that are independent of the number and the scale of retinal vessels measured. ²² They use measurements from the same software (IVAN, University of Wisconsin, Madison, WI) as in previous population-based studies, ^{20 21} but combine individual vessel measures differently, by an iterative procedure of pairing the largest vessels with the smallest and repeating until a single value, the central retinal vessel equivalent, is obtained. ²²  

Intragrader reliability was assessed in 200 randomly selected retinal photographs, and the intraclass correlation coefficient (95% CI) was 0.99 (0.98–0.99) for CRAE and 0.94 (0.92–0.96) for CRVE. 

Assessment of Cardiovascular Risk Factors

Participants underwent a standardized interview, systemic and ocular examination, and laboratory investigations. A detailed interviewer-administered questionnaire based on the Blue Mountains Eye Study was used to collect information about medical history (e.g., hypertension, diabetes, angina, myocardial infarction, and stroke), cigarette smoking (defined as current, past, or never), alcohol consumption (defined as ever or never), and current medications used (e.g., aspirin, oral steroids). ^{4 23} Socioeconomic variables assessed included education, income levels, housing type, and marital status. Education was recorded as the highest number of years of schooling completed and was categorized as primary or lower (≤6 years), secondary (7–10 years), or postsecondary (≥11 years, including university education). Income level was defined as individual monthly income in US dollars (USD) and was divided into two categories: (1) low (from <USD680 to USD2040 or retired) and (2) high (>USD2040). The majority of Singaporeans live in government-built high-rise buildings, with the remainder in private estates. As a result, housing status was classified as one- to two-room, three- to four-room, or five-room public housing or private housing. Marital status was categorized as married, divorced or separated, never married, or widowed. 

Height was measured in centimeters using a wall-mounted measuring tape. Weight was measured in kilograms with a digital scale (SECA, model 782-2321009; Vogel & Halke, Hamburg, Germany). Body mass index (BMI) was calculated as kg/m². Blood pressure was taken with the participant seated and after 5 minutes of rest by using a digital automatic blood pressure monitor (Dinamap model Pro Series DP110X-RW, 100V2; GE Medical Systems Information Technologies, Inc., Milwaukee, WI) according to a protocol similar to that used in the Multi-ethnic Study of Atherosclerosis. ²⁴ Blood pressure was measured twice, 5 minutes apart. A third measurement was made if the systolic blood pressure differed by more than 10 mm Hg or the diastolic by more than 5 mm Hg. The mean between the two closest readings was then taken as the blood pressure of that individual. Mean arterial blood pressure (MABP) was calculated as two thirds of the diastolic blood pressure plus one third of the systolic value. Hypertension was defined as systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg at examinations or physician diagnosis. Diabetes mellitus was defined as nonfasting glucose ≥11.1 mmol/L, use of diabetic medication, or self-reported history of diabetes. ²⁵  

Nonfasting venous blood samples were collected and used to assess levels of serum lipids (total, HDL, and LDL cholesterol), glycosylated hemoglobin (HbA1_C), serum creatinine, and glucose at the National University Hospital Reference Laboratory. ¹⁹  

Statistical Analysis

Retinal vascular caliber (CRAE and CRVE) was analyzed as continuous variables. We used analysis of covariance (ANCOVA) to estimate mean retinal vascular caliber associated with the presence versus absence of categorical variables (e.g., diabetes) or increasing quartiles of continuous variables (e.g., glucose levels) adjusted for age and sex. Tests for trend were determined by treating categorical risk factors (e.g., quartiles of glucose) as continuous ordinal variables in multiple linear regression models. We analyzed age both continuously and categorically (age groups: 40–49, 50–59, 60–69, and 70–80 years). 

We constructed two multivariate models. First, in model 1, we constructed a model for CRAE initially including variables that were significantly associated with CRAE in age and sex analyses. Final variables were selected based on a stepwise backward approach, adjusting for age. We then constructed a similar model for CRVE based on variables significantly associated with CRVE in age and sex analyses. Second, in model 2, we preformed supplementary analysis by additional adjustment for the fellow retinal vessel caliber (i.e., CRVE was included as an independent variable in the model for CRAE, and vice versa), as previously described by Liew et al. ²⁶ to control for potential confounding from fellow vessel diameter. We calculated sequential R ² to indicate the contribution of each independent variable to the model. Data presented here are based on results from model 1, as results from model 2 were largely similar. 

Finally, we investigated the potential interaction of risk factors with age and sex. Where interactions were statistically significant (P < 0.01), stratified analyses were performed by subgroup. All analyses were conducted with commercial software (Stata ver. 9.0; Stata Corp., College Station, TX). 

Results

Selected characteristics of the study sample with retinal vascular caliber data stratified by sex are shown in Table 1 . The mean age of the study sample was 57.7 ± 10.6 (SD) years with 48.2% males. High prevalence rates for hypertension (n = 2024, 67.0%) and diabetes (n = 690, 22.9%) were observed in this study population. The age-specific prevalence of diabetes is comparable to that in the Singapore National Health Survey conducted in 1992 for Malays aged 40 to 70 years. ²⁷ Apart from serum glucose and self-reported stroke, there were also significant differences between men (n = 1455) and women (n = 1564) in the frequency and distribution of most risk factors examined. 

Mean retinal arteriolar caliber was 139.5 ±15.7 μm (SD) and mean venular caliber was 219.3 ±22.2 μm. The distributions were approximately normal (Fig. 1) . 

Table 2shows cardiovascular risk factor associations of retinal vascular caliber, adjusted for age and sex. Smaller retinal arteriolar caliber was associated with increasing age, male sex, increasing quartiles of systolic and diastolic blood pressure, hypertension, higher quartiles of total cholesterol, BMI, and serum creatinine. Larger retinal arteriolar caliber was associated with diabetes and current cigarette smoking. 

Larger venular caliber was associated with younger age, diabetes, higher quartiles of glucose, BMI, and HbA1c, current cigarette smoking, lower quartiles of HDL cholesterol, and lower income (Table 2) . Other risk factors (e.g., LDL cholesterol, education level, housing and marital status, and self reported stroke and myocardial infarction) were not associated with retinal vascular caliber in age- and sex-adjusted analysis (data not shown). 

Table 3shows multivariate analyses (model 1) for retinal vascular caliber. Smaller retinal arteriolar caliber was associated with male sex and elevated MABP. Larger arteriolar caliber was associated with current cigarette smoking. In supplementary analyses (model 2), the associations of male sex and MABP with narrower retinal arteriolar caliber remained significant after further adjustment for retinal venular caliber, but the association of current cigarette smoking with wider arteriolar caliber was no longer present (data not shown). Larger retinal venular caliber was associated with younger age, current cigarette smoking, greater body mass index, higher HbA1c level, lower HDL cholesterol and lower income levels (model 1). After further adjusting for retinal arteriolar caliber, all associations remained significant except for the association of lower income with wider retinal venular caliber (model 2, P = 0.07, data not shown). 

Table 3also shows the contributions of each risk factor to arteriolar and venular caliber models. The major contributions to arteriolar caliber model were MABP, male sex, and current cigarette smoking; the major contributions to venular caliber model were age, HDL cholesterol, HbA1c, current cigarette smoking, BMI, and income. 

Table 4presents the results showing significant interaction of age and sex with MABP in relation to retinal arteriolar caliber. The relationship of narrower retinal arteriolar caliber to higher MABP was weaker with increasing age; with a decreasing trend in β-coefficients for MABP (−5.26, −4.44, −3.04, and −1.27, across the age groups of 40–49, 50–59, 60–69, and 70–80 years, respectively). Among persons with MABP levels <100 mm Hg, retinal arteriolar caliber narrowed with increasing age, but in persons with MABP >110 mm Hg, retinal arteriolar caliber significantly widened with increasing age. The relationship of narrower retinal arteriolar caliber and higher MABP was stronger in the men than in the women, with trend β-coefficient across four MABP groups of −4.35 for the men and −2.83 for the women, respectively. 

Discussion

In this population-based study among Asians, we demonstrated two distinct patterns of association for retinal arteriolar and venular caliber. Retinal arteriolar caliber is strongly related to blood pressure, and retinal venular caliber is related to cigarette smoking, lower HDL cholesterol level, higher HbA1c level, and greater BMI. Our current findings are consistent with previous reports on the patterns of association of retinal vascular caliber with blood pressure and cardiovascular risk factors from predominantly white persons ^{3 4 5 8 9 14} and from a Japanese population, the only other reported study of an Asian ethnic group. ⁶  

The association of retinal arteriolar narrowing and hypertension is well documented. ^{1 2} In our study, the observed magnitude of the association between elevated blood pressure and smaller arteriolar caliber was similar to that demonstrated in other population-based studies of white persons. ^{3 4 5 6 28 29} Besides blood pressure, we found few associations with arteriolar caliber. The association of current cigarette smoking with larger arteriolar caliber became nonsignificant after further adjustment for the fellow vascular caliber, suggesting the possibility of a confounding effect of retinal venular caliber on this association. ²⁶  

We have shown that larger retinal venular caliber predicts incident stroke and coronary heart disease. ⁹ The relationship of larger retinal venular caliber to current cigarette smoking, lower HDL cholesterol level, higher glycosylated hemoglobin, and higher BMI, all major cardiovascular risk factors, is consistent with previous data in white and black populations ^{5 14 30} and supports the notion that larger venular caliber is a biomarker of adverse vascular risk in persons with diabetes and prediabetes. ^{31 32 33} The finding that lower income was associated with larger venular caliber, independent of other risk factors, is interesting. Lower income, reflecting lower socioeconomic status, has been shown to affect health behavior and lifestyle patterns and is considered a surrogate for cardiovascular risk factors and other exposures. ^{34 35}  

We have described an interaction between age, blood pressure, and retinal arteriolar caliber. The weaker association of higher blood pressure and narrower retinal arteriolar caliber in the older compared with the younger participants is consistent with previous observations among white persons in the Beaver Dam and Blue Mountains Eye studies. ^{3 10} In addition, in persons with lower blood pressure, we showed that increasing age was associated with smaller arteriolar caliber; however, in persons with higher blood pressure, increasing age was independently associated with wider arteriolar caliber. This arteriolar “ballooning” effect associated with increasing blood pressure in older individuals has been described previously ²⁸ and may be related to cumulative fibrotic or atrophic arteriolar wall changes from both aging (and its related conditions) and blood pressure, with subsequent passive dilation. ³⁶  

Our finding that retinal arteriolar caliber is wider in women than in men is similar to reports of some previous studies in whites and other ethnicities, ^{4 5} although this has not been a consistent finding. ³⁷ We also found a stronger effect of blood pressure on retinal arteriolar narrowing in men than in women. We speculate that these findings reflect a protective effect of sex hormones on the microcirculation, consistent with a similar effect of estrogen on macrovascular disease. ³⁸ We should note, however, that previous studies lend little support to an effect of hormone replacement therapy or female reproductive parameters on retinal arteriolar caliber in elderly white women. ^{39 40}  

Taken together, the range of cardiovascular risk factors considered in our analysis explained only 8.7% and 4.2%, respectively, of the variance in retinal arteriolar caliber and venular caliber. This suggests that many unmeasured and unknown factors (e.g., genetic factors and inflammatory markers) affect retinal vascular caliber. ^{26 41}  

The strengths of our study include its population-based sample, the use of standardized retinal vascular caliber measurement from digital photographs, and detailed information on a range of risk factors and potential confounders. The findings in our study should be interpreted within the context of several potential limitations. First, the cross-sectional nature of our study and others does not provide temporal information or support for the associations. There are current plans for a prospective study. Second, blood glucose, HbA1c, and lipids were measured from nonfasting blood samples. Third, retinal vascular caliber measured using our imaging software may be influenced by the pulse cycle, although this variation is likely to bias associations toward the null. ⁴² We obtained estimates of central retinal vascular caliber using the improved Knudtson formulas, which have been shown to be more robust than the previous Parr-Hubbard formulas, and this may partially explain some differences between our findings and those of other studies in which the Parr-Hubbard formulas were used. ²² Finally, some variables (e.g., alcohol consumption, smoking, and education level) were obtained from self-reported questionnaires based on validated questionnaires used in other population-based studies, such as the Blue Mountains Eye Study, and are unlikely to be a major source of bias. 

In conclusion, findings from this large Asian Malay population confirm observations in white, blacks, and Japanese populations that there are two distinct patterns of association for retinal arteriolar and venular caliber. Retinal arteriolar caliber is strongly related to blood pressure. Retinal venular caliber, in contrast, is related to cigarette smoking, lower HDL cholesterol, higher HbA1c levels, and greater BMI. Stronger effects of blood pressure on retinal arteriolar caliber were seen in men compared with women and in younger compared with older people. These findings add to the growing body of literature on the effects of blood pressure and cardiovascular factors on retinal vascular changes in diverse populations. 

 Supported by the National Medical Research Council Grants 0796/2003, 0863/2004 and CSI/0002/2005; Biomedical Research Council Grant No 501/1/25-5; and the Singapore Tissue Network and the Ministry of Health, Singapore.

 Submitted for publication November 11, 2007; revised January 6, 2008; accepted March 13, 2008.

 Disclosure: C. Sun, None; G. Liew, None; J.J. Wang, None; P. Mitchell, None; S.M. Saw, None; T. Aung, None; E.S. Tai, None; T.Y. Wong, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Characteristics of the SiMES Study Population

	All Persons (n = 3019)	Males (n = 1455)	Females (n = 1564)	P *
Age (y)	57.7 (10.6)	58.3 (11.0)	57.1 (10.3)	0.001
Systolic blood pressure (mm Hg)	145.9 (23.1)	143.9 (21.8)	147.8 (24.2)	<0.001
Diastolic blood pressure (mm Hg)	79.6 (11.1)	81.3 (10.8)	78.0 (11.1)	<0.001
Serum glucose (mmol/dL)	6.8 (3.7)	6.7 (3.6)	6.8 (3.7)	0.47
Body mass index (kg/m2)	26.4 (5.1)	25.3 (4.4)	27.5 (5.5)	<0.001
HDL cholesterol (mg/dL)	1.4 (0.3)	1.2 (0.3)	1.5 (0.3)	<0.001
Serum creatinine (mg/dL)	91.9 (48.4)	105.6 (47.3)	79.1 (45.8)	<0.001
Sex, male	1455 (48.2)	—	—	—
Education, primary/lower ≤6 y	2221 (73.6)	997 (68.8)	1224 (78.4)	<0.001
Hypertension	2024 (67.0)	928 (63.8)	1096 (70.1)	<0.001
Diabetes	690 (22.9)	303 (20.8)	387 (24.7)	0.001
Cigarette smoker, current	624 (20.7)	583 (40.2)	41 (2.6)	<0.001
Alcohol consumption, ever	52 (1.7)	45 (3.1)	7 (0.5)	<0.001
Myocardial infarction	190 (6.3)	131 (9.0)	59 (3.8)	<0.001
Stroke	70 (2.3)	33 (2.3)	37 (2.4)	0.60
Income ≤USD2040	2908 (96.3)	1380 (94.9)	1528 (97.7)	<0.001
House 1–2 room public flat	430 (14.3)	177 (12.2)	253 (16.2)	<0.001

 Data in the top section are expressed as the mean (±SD) and in the bottom section as n (% of distribution).

*  Adjusted for age (except for age).

Figure 1.

 

Distribution of retinal arteriolar caliber (CRAE; left) and retinal venular caliber (right; CRVE) in the SiMES.

Table 2.

 

Relationship of Cardiovascular Risk Factors with Retinal Arteriolar and Venular Caliber

	Arterial Caliber (CRAE)					Venular Caliber (CRVE)
		n	M (SE), μm*	P , †	n	M (SE), μm*	P , †
Age
40–49 y	805	141.0 (0.55)	0.001	805	223.0 (0.78)	<0.001
50–59 y	939	139.4 (0.51)		939	219.9 (0.72)
60–69 y	723	139.0 (0.58)		723	217.3 (0.82)
70–80 y	552	138.3 (0.66)		552	215.4 (0.94)
Sex
Male	1455	137.5 (0.41)	<0.001	1455	219.40 (0.58)	0.78
Female	1564	141.4 (0.39)		1564	219.16 (0.56)
Hypertension
Absent	994	144.4 (0.52)	<0.001	994	220.0 (0.75)	0.29
Present	2024	137.2 (0.35)		2024	218.9 (0.51)
Income (monthly)
≤USD2040	2908	139.6 (0.29)	0.12	2908	219.5 (0.41)	0.01
>USD2040	111	137.2 (1.50)		111	214.0 (2.12)
Systolic BP (mm Hg)
1st quartile, <129.0	756	144.82 (0.58)	<0.001	756	219.73 (0.84)	0.28
2nd quartile, 129.0–143.5	764	140.93 (0.55)		764	220.07 (0.80)
3rd quartile, 144.0–161.5	759	136.84 (0.55)		759	218.23 (0.80)
4th quartile, >161.5	736	135.46 (0.59)		736	218.97 (0.85)
Diastolic BP (mm Hg)
1st quartile, <71.5	797	144.05 (0.54)	<0.001	797	220.19 (0.79)	0.32
2nd quartile, 71.5–78.0	723	141.35 (0.56)		723	218.81 (0.82)
3rd quartile, 78.5–86.5	757	138.19 (0.55)		757	218.94 (0.80)
4th quartile, >86.5	738	134.28 (0.56)		738	218.99 (0.81)
Diabetes
Absent	2329	139.1 (0.32)	0.004	2329	218.7 (0.45)	0.01
Present	690	141.1 (0.60)		690	221.1 (0.85)
Serum glucose (mg/dL)
1st quartile, <4.8	802	140.0 (0.55)	0.74	802	217.6 (0.78)	0.005
2nd quartile, 4.8–5.4	667	138.9 (0.60)		667	219.9 (0.85)
3rd quartile, 5.4–7.1	729	138.2 (0.58)		729	218.1 (0.81)
4th quartile, ≥7.1	704	140.6 (0.59)		704	221.5 (0.83)
HbA1c (%)
1st quartile, <5.7	845	140.6 (0.54)	0.43	845	217.7 (0.76)	0.001
2nd quartile, 5.7–5.9	644	139.2 (0.61)		644	219.1 (0.86)
3rd quartile, 6.0–6.6	754	137.7 (0.57)		754	218.0 (0.80)
4th quartile, >6.6	707	140.6 (0.59)		707	222.3 (0.83)
Cigarette smoking
Past/never	2387	138.9 (0.33)	<0.001	2387	218.1 (0.46)	<0.001
Current	624	142.1 (0.70)		624	223.7 (0.99)
BMI (kg/m2)
1st quartile, <23.1	751	140.9 (0.57)	<0.001	751	216.9 (0.80)	0.009
2nd quartile, 23.1–26.0	750	140.3 (0.57)		750	220.8 (0.80)
3rd quartile, 26.0–29.4	750	139.3 (0.57)		750	218.7 (0.80)
4th quartile, ≥29.4	750	137.8 (0.58)		750	220.8 (0.82)
Total cholesterol, mg/dL
1st quartile, <4.9	746	142.1 (0.57)	0.002	746	218.7 (0.81)	0.12
2nd quartile, 4.9–5.5	742	139.3 (0.57)		742	218.2 (0.81)
3rd quartile, 5.5–6.3	738	139.0 (0.57)		738	219.7 (0.81)
4th quartile, ≥6.3	738	138.6 (0.57)		738	220.1 (0.81)
HDL cholesterol, mg/dL
1st quartile, <1.1	743	139.6 (0.59)	0.77	743	222.2 (0.83)	<0.001
2nd quartile, 1.1–1.3	746	139.3 (0.57)		746	219.5 (0.81)
3rd quartile, 1.3–1.5	736	140.2 (0.57)		736	218.9 (0.81)
4th quartile, ≥1.5	738	139.1 (0.59)		738	215.9 (0.81)
Serum creatinine, mg/dL
1st quartile, <4.8	741	140.8 (0.64)	0.005	741	219.1 (0.91)	0.66
2nd quartile, 4.8–5.4	730	139.9 (0.58)		730	220.2 (0.83)
3rd quartile, 5.4–7.1	741	138.8 (0.60)		741	218.8 (0.85)
4th quartile, ≥7.1	691	138.2 (0.64)		691	218.8 (0.91)

*  Mean (SE) for arteriolar and venular caliber, adjusted for age and sex (sex only adjusted for age; and age only adjusted for sex).

†  Test for trend.

Table 3.

 

Relationship of Cardiovascular Risk Factors for Retinal Arteriolar And Venular Caliber, Model 1

	Unit Change	Arteriolar Caliber (CRAE)*					Venular Caliber (CRVE), †
				β (SE)	P	R 2	β (SE)	P	R 2
Risk Factors
Age	Per 10-y increase	0.15 (0.29)	0.60	0.005	−2.86 (0.42)	<0.001	0.017
Sex	Male vs. female	−4.80 (0.63)	<0.001	0.015	—	—	—
MABP	Per SD increase (14 mm Hg)	−4.01 (0.28)	<0.001	0.063	—	—	—
Cigarette smoking	Current vs. never/past	2.89 (0.78)	<0.001	0.004	4.00 (1.05)	<0.001	0.005
BMI	Per SD increase (5.1 kg/m2)	—	—	—	1.23 (0.42)	0.003	0.004
HbA1c level	Per SD increase (1.5%)	—	—	—	1.41 (0.41)	0.001	0.005
HDL cholesterol	Per SD increase (0.33 mg/dL)	—	—	—	−1.71 (0.42)	<0.001	0.009
Income	<USD2040 vs. ≥USD2040				5.58 (2.13)	0.009	0.002
Total R 2				0.087			0.042

*  Model 1 for CRAE, adjusted for age, sex, mean arterial blood pressure (MABP) and current smoking status.

†  Model 1 for CRVE, adjusted for age, blood HbA1c, BMI, HDL cholesterol, current smoking status, and income.

Table 4.

 

Interaction between Age and Sex with MABP in Relation to Retinal Arterial Caliber (CRAE)

	n	<100* (n = 1505)	100–109* (n = 686)	110–119* (n = 492)	>120* (n = 336)	P Trend, †	β (SE) for MABP, †
Age group
40–49	805	143.78 (0.61)	138.55 (1.12)	134.11 (1.44)	126.98 (2.06)	<0.001	−5.26 (0.54)	Age–MABP interaction P < 0.001
50–59	939	143.79 (0.69)	137.10 (0.95)	134.57 (1.26)	131.00 (1.46)	<0.001	−4.44 (0.47)
60–69	723	142.48 (0.86)	138.41 (1.21)	137.68 (1.21)	132.38 (1.50)	<0.001	−3.04 (0.51)
70–80	552	140.59 (1.23)	137.15 (1.47)	136.74 (1.68)	137.32 (1.87)	0.12	−1.27 (0.67)
	P trend, †	0.03	0.59	0.05	<0.001
	β (SE) for increasing age, †	−0.79 (0.37)	−0.31 (0.57)	1.28 (0.66)	3.14 (0.91)
Sex
Male	1455	140.78 (0.61)	135.73 (0.88)	132.22 (0.99)	127.36 (1.35)	<0.001	−4.35 (0.39)	Sex–MABP interaction P < 0.001
Female	1564	144.88 (0.56)	139.91 (0.89)	138.80 (1.03)	134.89 (1.35)	<0.001	−2.83 (0.37)
	P trend, †	<0.001	0.001	<0.001	0.001
	β (SE) for males vs. females, †	−4.10 (0.89)	−4.19 (1.31)	−6.58 (1.47)	−7.53 (2.01)

 Data are the mean arterial blood pressure (mm Hg ± SE).

*  Mean retinal arteriolar caliber (CRAE) in micrometers adjusted for age (per 10 years), sex, MABP (per 10 mm Hg), and current smoking status.

†  Trend across four different age groups (for increasing age); trend across four MABP groups (for increasing MABP) and trend for males vs. females.

The authors thank the staff and participants in the SiMES study for their important contributions.