Retinal Arteriolar Diameter and Risk for Hypertension

Context Does narrowing of small arterioles lead to or result from hypertension? Contribution This large prospective study measured diameters of small retinal vessels using digitized photographs in people without preexisting hypertension. After 3 years, more people with narrowed arterioles at baseline had hypertension than did people without any arteriolar narrowing. Implications Smaller arteriolar diameters are independently associated with development of hypertension. Cautions Rather than leading to hypertension, reduced arteriolar diameters at baseline might have reflected elevated blood pressure in persons who did not yet meet diagnostic criteria for hypertension. The Editors Hypertension affects up to 50 million people in the United States and is the single most important modifiable risk factor for stroke (1). Despite extensive research, much remains to be elucidated about the risk factors and pathogenesis of hypertension. A key characteristic of hypertension is the presence of narrowing and vasoconstriction of the small arteries and arterioles in the peripheral circulation (2-4). However, it is uncertain whether reduced arteriolar caliber, by increasing peripheral vascular resistance, contributes to the subsequent development of hypertension (5-7). Prospective clinical data demonstrating a link between smaller arteriolar caliber and risk for hypertension are unavailable, largely because the microcirculation is difficult to evaluate outside of experimental settings (5-7). As a result, the value of specific antihypertensive treatment targeted at the peripheral microcirculation remains questionable (7). The retinal arterioles offer a unique opportunity to noninvasively investigate the relation of arteriolar characteristics to the development of cardiovascular disease (8). We recently developed a method to quantify retinal arteriolar diameters from digitized retinal photographs (9). In this study, we examine whether retinal arteriolar narrowing is related to incident hypertension in a cohort of middle-aged normotensive persons. Methods Study Sample The Atherosclerosis Risk in Communities (ARIC) study is a population-based cohort study with 4 examinations (10). The ARIC study examined 15792 participants 45 to 64 years of age at baseline from 1987 to 1989 (10). The study sample was selected by probability sampling from 4 U.S. communities: Forsyth County, North Carolina; Jackson, Mississippi; suburbs of Minneapolis, Minnesota; and Washington County, Maryland. The Jackson sample included African-American persons only; in the other field centers, samples were representative of the populations in these communities (that is, mostly white persons in the suburbs of Minneapolis and Washington County and about 15% African-American persons in Forsyth County). Initial participation rates were 46% in Jackson and approximately 65% in the other communities. Participants were examined every 3 years; the second examination was done between 1990 and 1992 (n = 14348 [93% of 15440 survivors]), the third examination was done between 1993 and 1995 (n = 12887 [86% of 14944 survivors]), and the fourth examination was done between 1996 and 1998 (n = 11656 [81% of 14485 survivors]). Retinal photographs were taken at the third examination. Of the 12887 persons who returned for this examination, we excluded 38 whose race was neither African American nor white and 42 nonwhite residents in the suburbs of Minneapolis and Washington County, 1009 with no photographs or ungradable photographs, and 1434 who did not participate in the fourth examination. We further excluded 4464 persons with prevalent hypertension diagnosed at the first, second, or third examination and 272 persons with missing hypertension data. The remaining cohort consisted of 5628 normotensive persons at the third examination (Figure 1). Excluded participants (n = 7259) were older and more likely to be African American; had higher systolic and diastolic blood pressures, body mass indexes, waist-to-hip ratios, and fasting glucose and triglyceride levels; had lower high-density lipoprotein cholesterol levels; and were more likely to currently smoke and drink alcohol compared with participants included in the study (data not shown). Figure 1. Study design and population. Institutional review boards at each study site and at the Fundus Photograph Reading Center at the University of Wisconsin, Madison, Wisconsin, approved the study. Informed consent was obtained from all participants. Measurement of Retinal Arteriolar Diameters The retinal photography procedure has been reported in detail (9). Briefly, photographs of the retina of one randomly selected eye were taken after 5 minutes of dark adaptation. To estimate a generalized reduction in arteriolar diameters (referred to as generalized retinal arteriolar narrowing in this paper), the photographs were digitized, and the diameters of all arterioles and venules coursing through a specified area surrounding the optic disc were measured on the computer by graders who were blinded to participant identity. The individual arteriolar and venular diameters were combined into summary measures (in m) and combined as an arteriole-to-venule ratio on the basis of formulas described elsewhere (9). The arteriole-to-venule ratio accounts for magnification differences among photographs and is distributed normally. An arteriole-to-venule ratio of 1.0 indicates that retinal arteriolar diameters were, on average, the same as venular diameters, whereas a smaller ratio represents narrower arterioles, because venular diameters vary little (9). Figure 2 shows examples of retinas with low and high arteriole-to-venule ratios. The intra- and intergrader reliability coefficients for the arteriole-to-venule ratio were 0.84 and 0.79, respectively (9). Figure 2. Retinal photographs with arteriole-to-venule ratios. Top. Bottom. Trained graders who were blinded to retinal vessel measurements also evaluated photographs for the presence of localized areas of arteriolar constriction (referred to as focal retinal arteriolar narrowing) as well as other retinal microvascular characteristics (arteriovenous nicking, microaneurysms, and retinal hemorrhages) by using a standard protocol. Intra- and intergrader statistics ranged from 0.61 to 1.00 (9). Definition of Incident Hypertension Trained technicians performed a standard evaluation of blood pressure at each examination (11). Blood pressure was taken with a random-zero sphygmomanometer, and the mean of the last 2 measurements was used. Hypertension was defined as systolic blood pressure of 140 mm Hg or higher, diastolic blood pressure of 90 mm Hg or higher, or use of antihypertensive medication during the previous 2 weeks (11). Persons without preexisting hypertension at the first, second, or third examination who met these criteria at the fourth examination were defined as having incident hypertension. To examine the effect of current and previous blood pressure on these associations, we defined a person's 6-year average systolic and diastolic blood pressures as the mean of the blood pressure measurements taken at the first, second, and third examinations. The 6-year average systolic and diastolic blood pressures were then included as covariates in the assessment of the independence of retinal arteriolar narrowing with incident hypertension. We categorized a person as having normal blood pressure for systolic values averaging less than 130 mm Hg and diastolic values averaging less than 85 mm Hg and as having high normal blood pressure for systolic values averaging 130 to 139 mm Hg or diastolic values of 85 to 90 mm Hg, according to the Joint National Committee on the Prevention, Detection, Evaluation, and Treatment of High Blood Pressure classification (12). Definition of Other Variables Height and weight were measured with participants dressed in scrub suits. Body mass index, defined as weight/height2, was then computed. We calculated waist-to-hip ratio as the circumference of the waist (umbilical level) divided by the circumference of the hips (maximum circumference of the buttocks). We characterized physical activity by using a sports index; scores ranged from 1 (low) to 5 (high) (13). Diabetes mellitus was defined as a fasting glucose level of 7.0 mmol/L (126 mg/dL) or higher, a nonfasting glucose level of 11.1 mmol/L (200 mg/dL) or higher, or a history of physician-diagnosed diabetes or treatment for diabetes. Blood collection and processing for total and high-density lipoprotein cholesterol levels, triglyceride level, and fasting glucose level are described elsewhere (14). All variables were based on data from the third examination. Statistical Analysis For analysis of generalized retinal arteriolar narrowing, the arteriole-to-venule ratio was categorized into quintiles (with the first quintile indicating the largest arteriolar diameters and the fifth representing the smallest diameters). We also analyzed the ratio as a continuous variable (per SD reduction). Focal retinal arteriolar narrowing and other lesions were defined as binary variables. We used analysis of covariance to compare the arteriole-to-venule ratio and its components (summary measures of retinal arteriolar and venular diameters) between persons who did and did not subsequently develop hypertension. We used multiple logistic regression to calculate the odds ratio of incident hypertension by comparing a given arteriole-to-venule ratio quintile with the first quintile and the presence versus the absence of focal narrowing and other lesions. In these models, we adjusted for age, sex, race, field center, 6-year average systolic and diastolic blood pressures (mm Hg), body mass index (kg/m2), waist-to-hip ratio, sports activity index (1 to 5), diabetes (yes or no), cigarette smoking and alcohol consumption (current, former, or never), total cholesterol level, high-density lipoprotein cholesterol and triglyceride levels (mmol/L), and fasting glucose