- Email: [email protected]
Does refractive error influence the association of blood pressure and retinal vessel diameters? the blue mountains eye study
Does Refractive Error Influence the Association of Blood Pressure and Retinal Vessel Diameters? The Blue Mountains Eye Study TIEN YIN WONG, FRCSE, MPH, PHD, JIE JIN WANG, MMED, PHD, ELENA ROCHTCHINA, MAPPLSTAT, RONALD KLEIN, MD, MPH, AND PAUL MITCHELL, MD, PHD
● PURPOSE: To determine if refractive errors influence the association of blood pressure and retinal vessel diameters. ● DESIGN: Population-based, cross-sectional study. ● METHODS: Retinal photographs from the right eyes of participants (n ⴝ 3,654, aged 49ⴙ years) in the Blue Mountains Eye Study taken during baseline examinations (1992 to 1994) were digitized. The diameter of all retinal vessels located half to one disk diameter from the disk margin was measured using a computer-assisted imaging program. These measurements were combined to provide the average diameters of retinal arterioles and venules of that eye, and the ratio of their diameters, the arteriole-to-venule ratio (AVR). The association of blood pressure and retinal vessel diameters was analyzed before and after correction for refraction using the Bengtsson formula. ● RESULTS: Before correction, each 10-mm Hg increase in mean arterial blood pressure was associated with a 3.7-m (95% confidence interval [CI], 3.2– 4.3) decrease in arteriolar diameter and a 0.9-m (95% CI, Biosketch and/or additional material at www.ajo.com. Accepted for publication Jan 6, 2004. From the Centre for Eye Research Australia, University of Melbourne, East Melbourne, Australia (T.Y.W.); the Singapore Eye Research Institute, National University of Singapore (T.Y.W.); the Centre for Vision Research, Department of Ophthalmology, University of Sydney, Westmead, Australia (J.J.W., E.R., P.M.); and the Department of Ophthalmology, University of Wisconsin, Madison, Wisconsin (R.K.). This study was supported by grants 153948, 991407, and 211069 from the Australian National Health and Medical Research Council, Canberra, Australia (P.M.), and by the Ethicon Foundation Fund Travel Grant from the Royal College of Surgeons of Edinburgh, UK and the Young Investigator Award from the Biomedical Research Council, Singapore (T.Y.W.). Inquiries to Paul Mitchell, MD, PhD, FRANZCO, University of Sydney Department of Ophthalmology, Westmead Hospital, Hawkesbury Rd, Westmead, NSW, Australia, 2145. Telephone: 61 2 9845 7960 Fax: 61 2 9845 6117, E-mail: [email protected]
1050
©
2004 BY
0.3– 0.9) decrease in venular diameter. After correction for refraction, each 1-mm Hg increase in mean arterial blood pressure was associated with a 3.7-m (95% CI, 3.2– 4.2) decrease in arteriolar diameter and a 0.8-m (95% CI, 0.3– 0.9) decrease in venular diameter. Refraction was not associated with the AVR and had no effect on the association of blood pressure and AVR. ● CONCLUSION: Refraction had no appreciable effect on the association of blood pressure and retinal vessel diameters or on the AVR. Correction for refraction is important for quantifying absolute retinal vessel caliber, but may not be particularly important in epidemiologic studies investigating the association of generalized retinal arteriolar narrowing and hypertension. (Am J Ophthalmol 2004;137:1050 –1055. © 2004 by Elsevier Inc. All rights reserved.)
T
HE RETINAL BLOOD VESSELS OF THE HUMAN EYE ARE
accessible to direct and repeated observations in vivo. Characteristics of its arterioles and venules in health and disease may provide unique information applicable to the study of various ocular and systemic vascular disorders. One such characteristic, generalized narrowing of the retinal arterioles, has recently been quantified via a computer-assisted method in two large population-based studies, the Atherosclerosis Risk in Communities (ARIC) Study and the Cardiovascular Health Study (CHS).1,2 Generalized retinal arteriolar narrowing was found to be associated with both current and past blood pressure2,3 and, independent of blood pressure, with risk of stroke,4 coronary heart disease,5 and diabetes.6 There is therefore increasing interest in the development of quantitative methods to assess retinal vessel diameters from fundus photographs for hypertension and cardiovascular risk assessment.
ELSEVIER INC. ALL
RIGHTS RESERVED.
0002-9394/04/$30.00 doi:10.1016/j.ajo.2004.01.035
An important methodologic consideration in the assessment of the retinal vessel diameters (and other fundus structures) from photographic images is the effect of optical magnification error associated with variation in refraction.7–11 From an epidemiologic perspective, variation in refraction would not be expected to bias the reported cardiovascular associations of generalized arteriolar narrowing. However, for risk prediction at an individual patient level (for example, identifying patients at risk of cardiovascular disease based on severity of arteriolar narrowing), understanding the influence of refractive error on retinal vessel diameter may be important. To address this issue, we examined the effect of refractive errors on the association of retinal vessel diameters and blood pressure in a population-based study of older people in Australia.
METHODS ● STUDY POPULATION:
The Blue Mountains Eye Study was a population-based study of eye diseases in an urban population aged 49 years or older.12 Baseline participants (n ⫽ 3,654) represented 82.4% of eligible potential participants living in two postcode areas in the Blue Mountains, Australia. This study was conducted according to the recommendations of the Declaration of Helsinki and was approved by the Western Sydney Area Human Ethics Committee. Written, informed consent was obtained from all participants. At the baseline examinations (1992 to 1994), dilated, 30-degree stereoscopic retinal photographs of the macula, optic disk, and other retinal fields of both eyes were taken, using a Zeiss FF3 fundus camera (Carl Zeiss, Oberkochen, Germany). In this report, retinal photographs from right eyes of 3,355 of the 3,654 participants were included, after excluding 299 participants who had no retinal photographs, had ungradable photographs, or had retinal pathologies that confounded the measurement of retinal vessel width.
● RETINAL VESSEL DIAMETER MEASUREMENTS: Details of image digitization and grading protocols were previously described.13,14 In brief, a trained grader masked to participant characteristics retrieved a fundus photograph image from the network, identified each vessel as an arteriole or venule using the original color photographs for reference, and selected a segment of the vessel within a specified area (half to one disk diameter surrounding the optic disk) for measurement. A computer-imaging program (Retinal Analysis, Optimate, Madison, Wisconsin) performed the measurements. The branches of arterioles were also measured if the trunk measures were ⱖ85 m. The retinal vessel measurements were summarized using formulas by Parr and Hubbard.13,14 The formulas allow individual vessel diameters to be combined into the summary indices reflecting the arteriolar (referred to as the central retinal arteriolar equivalent) and venular (central retinal venular
VOL. 137, NO. 6
REFRACTIVE ERRORS
AND
equivalent) diameter of that eye, taking into account branching patterns. These summary indices were further expressed as the arteriole-to-venule ratio (AVR). An AVR of 1.0 suggests that arteriolar diameters were, on average, the same as venular diameters in that eye, while a smaller AVR suggests narrower arterioles compared with venules.13 As previously reported, intragrader and intergrader grading reliability was high, with quadratic weighted values ranging from 0.80 to 0.93.14 ● MEASUREMENT OF REFRACTION: Refraction was measured using a standard protocol described elsewhere.15 Following objective refraction with an autorefractor, a subjective refraction was performed according to a modification of the Early Treatment Diabetic Retinopathy Study protocol using logMAR chart. For this study, baseline refraction data were converted to spherical equivalent refraction (SER), calculated by the algebraic addition of the best-corrected spherical refraction and half the cylindrical refraction. Emmetropia was defined as SER between ⫹1.00 diopter and –1.00 diopter inclusive, hyperopia as more than ⫹1.00 diopter, and myopia as less than –1.00 diopter. Myopia was further classified as low (⫺1.00 diopter to ⫺3.25 diopters), moderate (⫺3.50 diopters to –5.75 diopters), and high (⫺6.00 diopters or less). Hyperopia was also divided into low (⫹1.00 diopter to ⫹1.75 diopters), moderate (⫹2.00 diopters to ⫹3.75 diopters), and high (⫹4.00 diopters or greater). ● CORRECTION FOR REFRACTION: A correction factor (1– 0.017 SER) described by Bengtsson7 was used to account for the effects of magnification error resulting from refraction for the Zeiss fundus cameras. A comprehensive derivation of this factor is described elsewhere.7,8 This correction factor was also used in our previous studies on the assessment of optic disk size.16,17 ● BLOOD PRESSURE: Blood pressure was measured after participants had been comfortably seated for at least 5 minutes. A single measure of systolic and diastolic blood pressure using a mercury sphygmomanometer was recorded from the first and fifth Korotkoff sounds. Mean arterial blood pressure was defined as (0.33 ⫻ systolic blood pressure) ⫹ (0.67 ⫻ diastolic blood pressure). ● STATISTICAL METHODS: We performed analyses with SAS (Version 8.0; SAS Institute Inc, Cary, North Carolina). Summary indices of retinal arteriolar and venular diameters and the AVR were analyzed as continuous variables. Refractive error was analyzed as categoric variables defined above. Results were similar when refraction was analyzed as a continuous variable (data not shown). We used two models for analysis. First, we used analysis of covariance to compare the mean retinal vessel diameter in eyes with different categories of refractive errors, before and after correction of refraction with the Bengtsson
RETINAL VESSEL DIAMETER
1051
TABLE 1. Mean Retinal Vessel Diameter in Association With Categories of Refractive Error, Right Eyes Arteriolar diameters, m Correction
Refraction
No refractive correction
Myopia
Emmetropia Hyperopia
With refractive correction
Myopia
Emmetropia Hyperopia
Refractive Error
Number
High (⫺6.00 D or less) Mod (⫺3.50 D to ⫺5.75 D) Low (⫺1.25 D to ⫺3.25 D) (⫹1.00 D to ⫺1.00 D) Low (1.25 D to 1.75 D) Mod (2.00 D to 3.75 D) High (4.00 D or higher)
33 74 226 1,473 666 608 83
High (⫺6.00 D or less) Mod (⫺3.50 D to ⫺5.75 D) Low (⫺1.25 D to ⫺3.25 D) (⫹1.00 D to ⫺1.00 D) Low (1.25 D to 1.75 D) Mod (2.00 D to 3.75 D) High (4.00 D or higher)
33 74 226 1,473 666 608 83
Mean*
SE
162.5 3.3 175.9 2.2 183.6 1.3 193.2 0.5 197.0 0.7 198.8 0.8 204.7 2.1 P for trend ⬍.001 190.1 3.3 190.9 2.2 189.9 1.3 192.5 0.5 192.3 0.7 190.5 0.8 188.9 2.1 P for trend .14
Venular diameters, m Mean*
SE
195.9 3.3 206.9 2.2 214.2 1.3 224.4 0.5 228.3 0.7 231.6 0.8 238.9 2.1 P for trend ⬍.001 229.5 3.3 224.4 2.2 221.5 1.3 223.6 0.5 222.9 0.7 222.0 0.8 220.5 2.1 P for trend .13
Retinal AVR Mean*
SE
0.83 0.014 0.85 0.009 0.86 0.005 0.86 0.002 0.87 0.003 0.86 0.003 0.86 0.009 P for trend .23 — — — — — — —
— — — — — — — —
AVR ⫽ arteriole-to-venule ratio. *Mean arteriolar diameters, venular diameters, and AVR, in analysis of covariance models, adjusted for age and male or female sex.
formula. The Mantel extension test was used to test for linear trends,18 and the median values for each refractive category were included in the model as a continuous variable. Next, we constructed linear regression models of retinal vessel diameters and blood pressure, before and after correction for refraction. Models were initially adjusted for age and male or female sex. In multivariable models, we further adjusted for diabetes, fasting glucose and total cholesterol measurements, cigarette smoking, body mass index, and presence of cardiovascular disease.
refraction, no associations between refraction and vessel diameters were seen. Refraction was not associated with the AVR. Correction for refraction had no impact on the association between blood pressure and retinal vessel diameters (Table 2). As previously reported, increasing blood pressure was significantly associated with smaller retinal arteriolar and venular diameter and lower AVR.13 For example, before correction, each 10-mm Hg increase in mean arterial blood pressure was associated with a 3.7-m decrease in arteriolar diameter and a 0.9-m decrease in venular diameter. After correction for refraction, each 10-mm Hg increase in mean arterial blood pressure was associated with a similar 3.7-m decrease in arteriolar diameter and a 0.8-m decrease in venular diameter. Table 3 shows multivariable models of arteriolar diameters and systolic blood pressure, before and after correction for refraction, simultaneously controlling for age, male or female sex, diabetes, fasting glucose, total cholesterol, cigarette smoking, body mass index, and presence of cardiovascular disease. The additional adjustment of these factors did not substantially improve the adjusted R2 of the model between arteriolar diameter and systolic blood pressure (from 0.10 in Table 2 to 0.11 in Table 3). As in Table 2, correction for refraction had marginal impact on the association between increasing systolic blood pressure and smaller retinal arteriolar diameters. There was a 1.82-m decrease in arteriolar diameter per 10-mm Hg
RESULTS THE MEAN AGE OF THE STUDY POPULATION WAS 65.0 YEARS
(SD, 9.1). The percentages of people who had hypertension, diabetes, a history of cardiovascular disease, and who were current cigarette smokers were 44.9%, 7.1%, 18.3%, and 15.2%, respectively. The mean systolic, diastolic, and arterial blood pressures were 145.6 mm Hg, 83.5 mm Hg, and 104.0 mm Hg, respectively. Before correction for refraction, eyes with a myopic refraction had smaller mean retinal arteriolar and venular diameters than eyes with a hyperopic refraction (Table 1). Retinal arteriolar and venular diameters were approximately 20% smaller (40 m) comparing highly myopic eyes (⫺6.0 diopters or less) with highly hyperopic eyes (4.0 diopters or more). After correction for 1052
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JUNE 2004
TABLE 2. Effect of Refraction on the Relationship Between Blood Pressure and Retinal Vessel Diameter, Right Eyes Arteriolar diameters
Blood pressure per 10 mm Hg
Systolic blood pressure No refractive correction With refractive correction Diastolic blood pressure No refractive correction With refractive correction Mean arterial blood pressure No refractive correction With refractive correction
Venular diameters
Retinal AVR
Mean Change (95% CI) (m)*
Adjusted R2
Mean Change (95% CI) (m)*
Adjusted R2
Mean Change (95% CI)*
Adjusted R2
⫺1.9 (⫺2.2, ⫺1.6) ⫺1.9 (⫺2.2, ⫺1.6)
0.10 0.14
⫺0.5 (⫺0.9, ⫺0.2) ⫺0.5 (⫺0.9, ⫺0.2)
0.04 0.08
⫺0.006 (⫺0.008, ⫺0.005) —
0.05
⫺4.3 (⫺5.0, ⫺3.7) ⫺4.3 (⫺4.9, ⫺3.7)
0.11 0.15
⫺0.8 (⫺1.5, ⫺0.1) ⫺0.7 (⫺1.4, ⫺0.1)
0.04 0.08
⫺0.016 (⫺0.019, ⫺0.014) —
0.06
⫺3.7 (⫺4.3, ⫺3.2) ⫺3.7 (⫺4.2, ⫺3.2)
0.11 0.15
⫺0.9 (⫺1.4, ⫺0.3) ⫺0.8 (⫺1.4, ⫺0.3)
0.04 0.08
⫺0.013 (⫺0.016, ⫺0.011) —
0.06
AVR ⫽ arteritole-to-venule ratio. *Mean change in arteriolar diameters, venular diameters, and AVR, for each 10-mm Hg increase in systolic, diastolic, and mean arterial blood pressure, adjusted for age and male or female sex.
TABLE 3. Effect of Refraction on the Relationship Between Blood Pressure and Retinal Arteriolar Diameter, Adjusted for Cardiovascular Risk Factors No refractive correction Mean Change (SE) in Arteriolar Diameter (m)*
Variables
With refractive correction
P value
⫺1.82 (0.18) ⬍.001 ⫺0.40 (0.04) ⬍.001 ⫺4.22 (0.76) ⬍.001 1.78 (1.75) .31 0.50 (0.30) .10 0.30 (0.35) .38 7.03 (1.05) ⬍.001 ⫺0.08 (0.08) .34 0.96 (0.98) .33 Adjusted R2 ⫽ 0.11
Systolic blood pressure, 10 mm Hg Age, yr Sex, female vs male Diabetes, yes vs no Fasting glucose, mg/dl Total cholesterol, mg/dl Cigarette smoking, current vs never Body mass index, kg/m2 Cardiovascular disease, yes vs no
Mean Change (SE) in Arteriolar Diameter (m)*
P value
⫺1.79 (0.17) ⬍.001 ⫺0.56 (0.04) ⬍.001 ⫺3.73 (0.72) ⬍.001 1.89 (1.68) .26 0.60 (0.30) .04 0.28 (0.33) .40 6.62 (1.00) ⬍.001 ⫺0.10 (0.08) .22 0.79 (0.93) .39 Adjusted R2 ⫽ 0.15
*Mean change in arteriolar diameters, for a unit change in independent variables (for example, 10-mm Hg increase in systolic blood pressure).
increase in systolic blood pressure before correction, and after (1.79 m) correction for refraction.
DISCUSSION OUR STUDY DEMONSTRATES THE FOLLOWING. FIRST, WE
showed that refraction was significantly associated with retinal vessel diameters as measured from fundus photographs, but had no effect on the retinal AVR, an indicator of the relative diameter of the arterioles compared with venules. After correction for magnification using the Bengtsson formula, the association of refraction and retinal vessel diameter was no longer present. Second, we demonstrated that refraction did not influence the relationship VOL. 137, NO. 6
REFRACTIVE ERRORS
AND
between higher blood pressure and smaller retinal vessel diameter. Our findings are consistent with the theoretical effect of refractive error and optical dimension on the image size of fundus photographs.10,19 Rudnicka and associates10 showed that for a telecentric camera, such as the Zeiss FF3 fundus camera used in our study, the image of an object on the fundus camera (or photograph) is directly proportional to the true size of that object on the retina and inversely proportional to the optical dimension of the eye. Thus, assuming myopia is only associated with longer axial lengths, the apparent diameter of a retinal blood vessel measured from fundus photographs would be reduced by a factor proportional to the increase in the optical dimension of the myopic eye. For example, when compared with RETINAL VESSEL DIAMETER
1053
from digitized fundus photographs. There are several noteworthy limitations. First, unknown sources of variability cannot be excluded, despite the overall high reproducibility of the retinal vessel measurements.14 Potential errors may have been introduced from variability of the timing of retinal photography, since “physiologic” variations in cardiac cycle25 and autonomic nervous system activity26 may influence retinal vessel caliber. In addition, grader variability may have led to further errors in retinal vessel measurement. However, we have no reason to believe these types of variability are differential between myopic and hyperopic eyes. Second, there were few people with extremes in refraction (severe myopia or hyperopia). It is unclear if these data are applicable to other populations with a higher prevalence of myopia or hyperopia. Third, the Bengtsson formula provides only an approximate magnification correction using spherical equivalent refraction. Future studies with axial length data will be important to determine more precisely the impact of magnification error on the absolute measurements of retinal vessel caliber and its relationship to hypertension and other cardiovascular disorders. In summary, we demonstrate that, at a population level, refraction affects the absolute magnitude of retinal vessel diameters measured from fundus photographs. However, refraction does not appear to affect the magnitude of the AVR or influence the direction or strength of association between retinal vessel diameters and blood pressure. Thus, correction for refraction may not be critical in epidemiologic studies examining the association of generalized retinal arteriolar narrowing with hypertension. Future studies with axial length data will be important to determine more precisely the impact of refractive error on the relationship of retinal vessel caliber to hypertension and other cardiovascular disorders.
a hyperopic eye with an axial length of 22 mm, a myopic eye with an axial length of 26 mm could be expected to have retinal vessel diameters that appeared 18% (4 mm divided 22 mm) smaller. This is compatible with our observation that retinal vessel diameters were approximately 20% smaller comparing myopic eyes ⫺6.0 diopters or less with hyperopic eyes 4.0 diopters or more. Correction using the Bengtsson formula eliminated the association of refraction with retinal vessel diameter size. Because refraction was not measured in the ARIC study or the CHS,1– 6 we had used the AVR as an indicator of generalized retinal arteriolar narrowing. In fact, the AVR had been proposed by Wagener and associates20 in 1947 as a practical method to quantify generalized arteriolar narrowing from direct ophthalmoscopy. Others have similarly hypothesized that the AVR would minimize the effect of potential optical magnification error introduced by variation in refraction, since retinas with artificially magnified (or minified) arterioles would likely have similarly magnified (or minified) venules.21–23 We now demonstrate that refraction has no effect on the magnitude of the AVR, which further supports its usefulness as a summary indicator of the relative arteriolar-to-venular diameter, particularly when refraction data are unavailable. In any case, our study shows that refraction does not appear to alter either the pattern or the strength of association between elevated blood pressure and smaller retinal vessel diameters. This finding suggests that association between retinal arteriolar narrowing and hypertension described in the ARIC study and the CHS is unlikely to be biased because of a lack of refraction data.1,2 Our results are also consistent with the Beaver Dam Eye Study, which used an identical computer-assisted method to measure retinal vessel diameters from digitized fundus photographs.24 In that study, after controlling for age and male or female sex, each 10-mm Hg increase in mean arterial blood pressure was associated with a 5.1-m (95% confidence intervals [CI], 4.6 –5.6) decrease in arteriolar diameter. This association was not altered with further adjustment for SER (5.0 m, 95% CI, 4.5–5.5 decrease in arteriolar diameter per 10-mm Hg increase in mean arterial blood pressure). We note that the statistical power to detect associations of retinal vessel diameters appears to be increased by the greater precision in vessel diameter data after correction for refraction, since less of the total variation is a result of measurement error (R2 was higher in the regression models of retinal vessel diameters corrected for refraction). Thus, correction for refraction may increase the chance of finding significant associations of retinal vessel diameters if the effect size is small. This should be taken into consideration in future studies with smaller sample sizes. Strengths of the current study include a populationbased sample less susceptible to selection biases, a standardized assessment of refraction, and a quantitative, computer-assisted measurement of retinal vessel diameters 1054
AMERICAN JOURNAL
REFERENCES 1. Couper DJ, Klein R, Hubbard LD, et al. Reliability of retinal photography in the assessment of retinal microvascular characteristics: the Atherosclerosis Risk in Communities Study. Am J Ophthalmol 2002;133:78 –88. 2. Wong TY, Hubbard LD, Klein R, et al. Retinal microvascular abnormalities and blood pressure in older people: the Cardiovascular Health Study. Br J Ophthalmol 2002;86:1007– 1013. 3. Sharrett AR, Hubbard LD, Cooper LS, et al. Retinal arteriolar diameters and elevated blood pressure: the Atherosclerosis Risk in Communities Study. Am J Epidemiol 1999;150: 263–270. 4. Wong TY, Klein R, Couper DJ, et al. Retinal microvascular abnormalities and incident strokes: the Atherosclerosis Risk in Communities Study. Lancet 2001;358:1134 –1140. 5. Wong TY, Klein R, Sharrett AR, et al. Retinal arteriolar narrowing and incident coronary heart disease in men and women: the Atherosclerosis Risk in Communities Study. JAMA 2002;287:1153–1159. 6. Wong TY, Klein R, Sharrett AR, et al. Retinal arteriolar OF
OPHTHALMOLOGY
JUNE 2004
7. 8. 9. 10. 11. 12. 13.
14. 15. 16.
narrowing and risk of diabetes in middle-aged persons. JAMA 2002;287:2528 –2533. Bengtsson B. The variation and covariation of cup and disc diameters. Acta Ophthalmol (Copenh) 1976;54:804 –818. Bengtsson B, Krakau CE. Correction of optic disc measurements on fundus photographs. Graefes Arch Clin Exp Ophthalmol 1992;230:24 –28. Littmann H. Zur Bestimmung der wahren Gro¨ sse eines Objektes auf dem Hintergrund des lebenden Auges. Klin Monatsbl Augenheilkd 1982;180:286 –289. Rudnicka AR, Burk RO, Edgar DF, Fitzke FW. Magnification characteristics of fundus imaging systems. Ophthalmology 1998;105:2186 –2192. Heier H, Brinchmann-Hansen O. Reliable measurements from fundus photographs in the presence of focusing errors. Invest Ophthalmol Vis Sci 1989;30:674 –677. Attebo K, Mitchell P, Smith W. Visual acuity and the causes of visual loss in Australia: the Blue Mountains Eye Study. Ophthalmology 1996;103:357–364. Leung H, Wang JJ, Rochtchina E, et al. Relationships between age, blood pressure, and retinal vessel diameters in an older population. Invest Ophthalmol Vis Sci 2003;44: 2900 –2904. Sherry L, Wang JJ, Rochtchina E, et al. The reliability of computer-assisted retinal vessel measurement in a population. Clin Exp Ophthalmol 2002;30:179 –182. Attebo K, Ivers RQ, Mitchell P. Refractive errors in an older population: the Blue Mountains Eye Study. Ophthalmology 1999;106:1066 –1072. Mitchell P, Smith W, Attebo K, Healey PR. Prevalence of open-angle glaucoma in Australia: the Blue Mountains Eye Study. Ophthalmology 1996;103:1661–1669.
VOL. 137, NO. 6
REFRACTIVE ERRORS
AND
17. Healey PR, Mitchell P. Optic disk size in open-angle glaucoma: the Blue Mountains Eye Study. Am J Ophthalmol 1999;128:515–517. 18. Mantel N. Chi-square test with one degree of freedom: extensions of the Mantel Haenszel procedure. J Am Stat Assoc 1963;58:690 –700. 19. Arnold JV, Gates JW, Taylor KM. Possible errors in the measurement of retinal lesions. Invest Ophthalmol Vis Sci 1993;34:2576 –2580. 20. Wagener HP, Clay GE, Gipner JF. Classification of retinal lesions in the presence of vascular hypertension. Trans Am Ophthalmol Soc 1947;45:57–73. 21. Stokoe NL, Turner RW. Normal retinal vascular pattern: arteriovenous ratio as a measure of arterial calibre. Br J Ophthalmol 1966;50:21–40. 22. Houben AJ, Canoy MC, Paling HA, et al. Quantitative analysis of retinal vascular changes in essential and renovascular hypertension. J Hypertens 1995;13:1729 –1733. 23. Stanton AV, Mullaney P, Mee F, et al. A method of quantifying retinal microvascular alterations associated with blood pressure and age. J Hypertens 1995;13:41–48. 24. Wong TY, Knudtson MD, Klein R, et al. Computer-assisted retinal vessel diameters in the Beaver Dam Eye Study: methodology, correlation between eyes and effect of refractive error. Ophthalmology (in press). 25. Knudtson MD, Klein BEK, Klein R, et al. Variation associated with retinal vessel diameter measurements controlling for known sources of variability. Br J Ophthalmol 2004;88:57– 61. 26. Lanigan LP, Clark CV, Hill DW. Retinal circulation responses to systemic autonomic nerve stimulation. Eye 1988; 2:412–417.
RETINAL VESSEL DIAMETER
1055
● PURPOSE: To determine if refractive errors influence the association of blood pressure and retinal vessel diameters. ● DESIGN: Population-based, cross-sectional study. ● METHODS: Retinal photographs from the right eyes of participants (n ⴝ 3,654, aged 49ⴙ years) in the Blue Mountains Eye Study taken during baseline examinations (1992 to 1994) were digitized. The diameter of all retinal vessels located half to one disk diameter from the disk margin was measured using a computer-assisted imaging program. These measurements were combined to provide the average diameters of retinal arterioles and venules of that eye, and the ratio of their diameters, the arteriole-to-venule ratio (AVR). The association of blood pressure and retinal vessel diameters was analyzed before and after correction for refraction using the Bengtsson formula. ● RESULTS: Before correction, each 10-mm Hg increase in mean arterial blood pressure was associated with a 3.7-m (95% confidence interval [CI], 3.2– 4.3) decrease in arteriolar diameter and a 0.9-m (95% CI, Biosketch and/or additional material at www.ajo.com. Accepted for publication Jan 6, 2004. From the Centre for Eye Research Australia, University of Melbourne, East Melbourne, Australia (T.Y.W.); the Singapore Eye Research Institute, National University of Singapore (T.Y.W.); the Centre for Vision Research, Department of Ophthalmology, University of Sydney, Westmead, Australia (J.J.W., E.R., P.M.); and the Department of Ophthalmology, University of Wisconsin, Madison, Wisconsin (R.K.). This study was supported by grants 153948, 991407, and 211069 from the Australian National Health and Medical Research Council, Canberra, Australia (P.M.), and by the Ethicon Foundation Fund Travel Grant from the Royal College of Surgeons of Edinburgh, UK and the Young Investigator Award from the Biomedical Research Council, Singapore (T.Y.W.). Inquiries to Paul Mitchell, MD, PhD, FRANZCO, University of Sydney Department of Ophthalmology, Westmead Hospital, Hawkesbury Rd, Westmead, NSW, Australia, 2145. Telephone: 61 2 9845 7960 Fax: 61 2 9845 6117, E-mail: [email protected]
1050
©
2004 BY
0.3– 0.9) decrease in venular diameter. After correction for refraction, each 1-mm Hg increase in mean arterial blood pressure was associated with a 3.7-m (95% CI, 3.2– 4.2) decrease in arteriolar diameter and a 0.8-m (95% CI, 0.3– 0.9) decrease in venular diameter. Refraction was not associated with the AVR and had no effect on the association of blood pressure and AVR. ● CONCLUSION: Refraction had no appreciable effect on the association of blood pressure and retinal vessel diameters or on the AVR. Correction for refraction is important for quantifying absolute retinal vessel caliber, but may not be particularly important in epidemiologic studies investigating the association of generalized retinal arteriolar narrowing and hypertension. (Am J Ophthalmol 2004;137:1050 –1055. © 2004 by Elsevier Inc. All rights reserved.)
T
HE RETINAL BLOOD VESSELS OF THE HUMAN EYE ARE
accessible to direct and repeated observations in vivo. Characteristics of its arterioles and venules in health and disease may provide unique information applicable to the study of various ocular and systemic vascular disorders. One such characteristic, generalized narrowing of the retinal arterioles, has recently been quantified via a computer-assisted method in two large population-based studies, the Atherosclerosis Risk in Communities (ARIC) Study and the Cardiovascular Health Study (CHS).1,2 Generalized retinal arteriolar narrowing was found to be associated with both current and past blood pressure2,3 and, independent of blood pressure, with risk of stroke,4 coronary heart disease,5 and diabetes.6 There is therefore increasing interest in the development of quantitative methods to assess retinal vessel diameters from fundus photographs for hypertension and cardiovascular risk assessment.
ELSEVIER INC. ALL
RIGHTS RESERVED.
0002-9394/04/$30.00 doi:10.1016/j.ajo.2004.01.035
An important methodologic consideration in the assessment of the retinal vessel diameters (and other fundus structures) from photographic images is the effect of optical magnification error associated with variation in refraction.7–11 From an epidemiologic perspective, variation in refraction would not be expected to bias the reported cardiovascular associations of generalized arteriolar narrowing. However, for risk prediction at an individual patient level (for example, identifying patients at risk of cardiovascular disease based on severity of arteriolar narrowing), understanding the influence of refractive error on retinal vessel diameter may be important. To address this issue, we examined the effect of refractive errors on the association of retinal vessel diameters and blood pressure in a population-based study of older people in Australia.
METHODS ● STUDY POPULATION:
The Blue Mountains Eye Study was a population-based study of eye diseases in an urban population aged 49 years or older.12 Baseline participants (n ⫽ 3,654) represented 82.4% of eligible potential participants living in two postcode areas in the Blue Mountains, Australia. This study was conducted according to the recommendations of the Declaration of Helsinki and was approved by the Western Sydney Area Human Ethics Committee. Written, informed consent was obtained from all participants. At the baseline examinations (1992 to 1994), dilated, 30-degree stereoscopic retinal photographs of the macula, optic disk, and other retinal fields of both eyes were taken, using a Zeiss FF3 fundus camera (Carl Zeiss, Oberkochen, Germany). In this report, retinal photographs from right eyes of 3,355 of the 3,654 participants were included, after excluding 299 participants who had no retinal photographs, had ungradable photographs, or had retinal pathologies that confounded the measurement of retinal vessel width.
● RETINAL VESSEL DIAMETER MEASUREMENTS: Details of image digitization and grading protocols were previously described.13,14 In brief, a trained grader masked to participant characteristics retrieved a fundus photograph image from the network, identified each vessel as an arteriole or venule using the original color photographs for reference, and selected a segment of the vessel within a specified area (half to one disk diameter surrounding the optic disk) for measurement. A computer-imaging program (Retinal Analysis, Optimate, Madison, Wisconsin) performed the measurements. The branches of arterioles were also measured if the trunk measures were ⱖ85 m. The retinal vessel measurements were summarized using formulas by Parr and Hubbard.13,14 The formulas allow individual vessel diameters to be combined into the summary indices reflecting the arteriolar (referred to as the central retinal arteriolar equivalent) and venular (central retinal venular
VOL. 137, NO. 6
REFRACTIVE ERRORS
AND
equivalent) diameter of that eye, taking into account branching patterns. These summary indices were further expressed as the arteriole-to-venule ratio (AVR). An AVR of 1.0 suggests that arteriolar diameters were, on average, the same as venular diameters in that eye, while a smaller AVR suggests narrower arterioles compared with venules.13 As previously reported, intragrader and intergrader grading reliability was high, with quadratic weighted values ranging from 0.80 to 0.93.14 ● MEASUREMENT OF REFRACTION: Refraction was measured using a standard protocol described elsewhere.15 Following objective refraction with an autorefractor, a subjective refraction was performed according to a modification of the Early Treatment Diabetic Retinopathy Study protocol using logMAR chart. For this study, baseline refraction data were converted to spherical equivalent refraction (SER), calculated by the algebraic addition of the best-corrected spherical refraction and half the cylindrical refraction. Emmetropia was defined as SER between ⫹1.00 diopter and –1.00 diopter inclusive, hyperopia as more than ⫹1.00 diopter, and myopia as less than –1.00 diopter. Myopia was further classified as low (⫺1.00 diopter to ⫺3.25 diopters), moderate (⫺3.50 diopters to –5.75 diopters), and high (⫺6.00 diopters or less). Hyperopia was also divided into low (⫹1.00 diopter to ⫹1.75 diopters), moderate (⫹2.00 diopters to ⫹3.75 diopters), and high (⫹4.00 diopters or greater). ● CORRECTION FOR REFRACTION: A correction factor (1– 0.017 SER) described by Bengtsson7 was used to account for the effects of magnification error resulting from refraction for the Zeiss fundus cameras. A comprehensive derivation of this factor is described elsewhere.7,8 This correction factor was also used in our previous studies on the assessment of optic disk size.16,17 ● BLOOD PRESSURE: Blood pressure was measured after participants had been comfortably seated for at least 5 minutes. A single measure of systolic and diastolic blood pressure using a mercury sphygmomanometer was recorded from the first and fifth Korotkoff sounds. Mean arterial blood pressure was defined as (0.33 ⫻ systolic blood pressure) ⫹ (0.67 ⫻ diastolic blood pressure). ● STATISTICAL METHODS: We performed analyses with SAS (Version 8.0; SAS Institute Inc, Cary, North Carolina). Summary indices of retinal arteriolar and venular diameters and the AVR were analyzed as continuous variables. Refractive error was analyzed as categoric variables defined above. Results were similar when refraction was analyzed as a continuous variable (data not shown). We used two models for analysis. First, we used analysis of covariance to compare the mean retinal vessel diameter in eyes with different categories of refractive errors, before and after correction of refraction with the Bengtsson
RETINAL VESSEL DIAMETER
1051
TABLE 1. Mean Retinal Vessel Diameter in Association With Categories of Refractive Error, Right Eyes Arteriolar diameters, m Correction
Refraction
No refractive correction
Myopia
Emmetropia Hyperopia
With refractive correction
Myopia
Emmetropia Hyperopia
Refractive Error
Number
High (⫺6.00 D or less) Mod (⫺3.50 D to ⫺5.75 D) Low (⫺1.25 D to ⫺3.25 D) (⫹1.00 D to ⫺1.00 D) Low (1.25 D to 1.75 D) Mod (2.00 D to 3.75 D) High (4.00 D or higher)
33 74 226 1,473 666 608 83
High (⫺6.00 D or less) Mod (⫺3.50 D to ⫺5.75 D) Low (⫺1.25 D to ⫺3.25 D) (⫹1.00 D to ⫺1.00 D) Low (1.25 D to 1.75 D) Mod (2.00 D to 3.75 D) High (4.00 D or higher)
33 74 226 1,473 666 608 83
Mean*
SE
162.5 3.3 175.9 2.2 183.6 1.3 193.2 0.5 197.0 0.7 198.8 0.8 204.7 2.1 P for trend ⬍.001 190.1 3.3 190.9 2.2 189.9 1.3 192.5 0.5 192.3 0.7 190.5 0.8 188.9 2.1 P for trend .14
Venular diameters, m Mean*
SE
195.9 3.3 206.9 2.2 214.2 1.3 224.4 0.5 228.3 0.7 231.6 0.8 238.9 2.1 P for trend ⬍.001 229.5 3.3 224.4 2.2 221.5 1.3 223.6 0.5 222.9 0.7 222.0 0.8 220.5 2.1 P for trend .13
Retinal AVR Mean*
SE
0.83 0.014 0.85 0.009 0.86 0.005 0.86 0.002 0.87 0.003 0.86 0.003 0.86 0.009 P for trend .23 — — — — — — —
— — — — — — — —
AVR ⫽ arteriole-to-venule ratio. *Mean arteriolar diameters, venular diameters, and AVR, in analysis of covariance models, adjusted for age and male or female sex.
formula. The Mantel extension test was used to test for linear trends,18 and the median values for each refractive category were included in the model as a continuous variable. Next, we constructed linear regression models of retinal vessel diameters and blood pressure, before and after correction for refraction. Models were initially adjusted for age and male or female sex. In multivariable models, we further adjusted for diabetes, fasting glucose and total cholesterol measurements, cigarette smoking, body mass index, and presence of cardiovascular disease.
refraction, no associations between refraction and vessel diameters were seen. Refraction was not associated with the AVR. Correction for refraction had no impact on the association between blood pressure and retinal vessel diameters (Table 2). As previously reported, increasing blood pressure was significantly associated with smaller retinal arteriolar and venular diameter and lower AVR.13 For example, before correction, each 10-mm Hg increase in mean arterial blood pressure was associated with a 3.7-m decrease in arteriolar diameter and a 0.9-m decrease in venular diameter. After correction for refraction, each 10-mm Hg increase in mean arterial blood pressure was associated with a similar 3.7-m decrease in arteriolar diameter and a 0.8-m decrease in venular diameter. Table 3 shows multivariable models of arteriolar diameters and systolic blood pressure, before and after correction for refraction, simultaneously controlling for age, male or female sex, diabetes, fasting glucose, total cholesterol, cigarette smoking, body mass index, and presence of cardiovascular disease. The additional adjustment of these factors did not substantially improve the adjusted R2 of the model between arteriolar diameter and systolic blood pressure (from 0.10 in Table 2 to 0.11 in Table 3). As in Table 2, correction for refraction had marginal impact on the association between increasing systolic blood pressure and smaller retinal arteriolar diameters. There was a 1.82-m decrease in arteriolar diameter per 10-mm Hg
RESULTS THE MEAN AGE OF THE STUDY POPULATION WAS 65.0 YEARS
(SD, 9.1). The percentages of people who had hypertension, diabetes, a history of cardiovascular disease, and who were current cigarette smokers were 44.9%, 7.1%, 18.3%, and 15.2%, respectively. The mean systolic, diastolic, and arterial blood pressures were 145.6 mm Hg, 83.5 mm Hg, and 104.0 mm Hg, respectively. Before correction for refraction, eyes with a myopic refraction had smaller mean retinal arteriolar and venular diameters than eyes with a hyperopic refraction (Table 1). Retinal arteriolar and venular diameters were approximately 20% smaller (40 m) comparing highly myopic eyes (⫺6.0 diopters or less) with highly hyperopic eyes (4.0 diopters or more). After correction for 1052
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JUNE 2004
TABLE 2. Effect of Refraction on the Relationship Between Blood Pressure and Retinal Vessel Diameter, Right Eyes Arteriolar diameters
Blood pressure per 10 mm Hg
Systolic blood pressure No refractive correction With refractive correction Diastolic blood pressure No refractive correction With refractive correction Mean arterial blood pressure No refractive correction With refractive correction
Venular diameters
Retinal AVR
Mean Change (95% CI) (m)*
Adjusted R2
Mean Change (95% CI) (m)*
Adjusted R2
Mean Change (95% CI)*
Adjusted R2
⫺1.9 (⫺2.2, ⫺1.6) ⫺1.9 (⫺2.2, ⫺1.6)
0.10 0.14
⫺0.5 (⫺0.9, ⫺0.2) ⫺0.5 (⫺0.9, ⫺0.2)
0.04 0.08
⫺0.006 (⫺0.008, ⫺0.005) —
0.05
⫺4.3 (⫺5.0, ⫺3.7) ⫺4.3 (⫺4.9, ⫺3.7)
0.11 0.15
⫺0.8 (⫺1.5, ⫺0.1) ⫺0.7 (⫺1.4, ⫺0.1)
0.04 0.08
⫺0.016 (⫺0.019, ⫺0.014) —
0.06
⫺3.7 (⫺4.3, ⫺3.2) ⫺3.7 (⫺4.2, ⫺3.2)
0.11 0.15
⫺0.9 (⫺1.4, ⫺0.3) ⫺0.8 (⫺1.4, ⫺0.3)
0.04 0.08
⫺0.013 (⫺0.016, ⫺0.011) —
0.06
AVR ⫽ arteritole-to-venule ratio. *Mean change in arteriolar diameters, venular diameters, and AVR, for each 10-mm Hg increase in systolic, diastolic, and mean arterial blood pressure, adjusted for age and male or female sex.
TABLE 3. Effect of Refraction on the Relationship Between Blood Pressure and Retinal Arteriolar Diameter, Adjusted for Cardiovascular Risk Factors No refractive correction Mean Change (SE) in Arteriolar Diameter (m)*
Variables
With refractive correction
P value
⫺1.82 (0.18) ⬍.001 ⫺0.40 (0.04) ⬍.001 ⫺4.22 (0.76) ⬍.001 1.78 (1.75) .31 0.50 (0.30) .10 0.30 (0.35) .38 7.03 (1.05) ⬍.001 ⫺0.08 (0.08) .34 0.96 (0.98) .33 Adjusted R2 ⫽ 0.11
Systolic blood pressure, 10 mm Hg Age, yr Sex, female vs male Diabetes, yes vs no Fasting glucose, mg/dl Total cholesterol, mg/dl Cigarette smoking, current vs never Body mass index, kg/m2 Cardiovascular disease, yes vs no
Mean Change (SE) in Arteriolar Diameter (m)*
P value
⫺1.79 (0.17) ⬍.001 ⫺0.56 (0.04) ⬍.001 ⫺3.73 (0.72) ⬍.001 1.89 (1.68) .26 0.60 (0.30) .04 0.28 (0.33) .40 6.62 (1.00) ⬍.001 ⫺0.10 (0.08) .22 0.79 (0.93) .39 Adjusted R2 ⫽ 0.15
*Mean change in arteriolar diameters, for a unit change in independent variables (for example, 10-mm Hg increase in systolic blood pressure).
increase in systolic blood pressure before correction, and after (1.79 m) correction for refraction.
DISCUSSION OUR STUDY DEMONSTRATES THE FOLLOWING. FIRST, WE
showed that refraction was significantly associated with retinal vessel diameters as measured from fundus photographs, but had no effect on the retinal AVR, an indicator of the relative diameter of the arterioles compared with venules. After correction for magnification using the Bengtsson formula, the association of refraction and retinal vessel diameter was no longer present. Second, we demonstrated that refraction did not influence the relationship VOL. 137, NO. 6
REFRACTIVE ERRORS
AND
between higher blood pressure and smaller retinal vessel diameter. Our findings are consistent with the theoretical effect of refractive error and optical dimension on the image size of fundus photographs.10,19 Rudnicka and associates10 showed that for a telecentric camera, such as the Zeiss FF3 fundus camera used in our study, the image of an object on the fundus camera (or photograph) is directly proportional to the true size of that object on the retina and inversely proportional to the optical dimension of the eye. Thus, assuming myopia is only associated with longer axial lengths, the apparent diameter of a retinal blood vessel measured from fundus photographs would be reduced by a factor proportional to the increase in the optical dimension of the myopic eye. For example, when compared with RETINAL VESSEL DIAMETER
1053
from digitized fundus photographs. There are several noteworthy limitations. First, unknown sources of variability cannot be excluded, despite the overall high reproducibility of the retinal vessel measurements.14 Potential errors may have been introduced from variability of the timing of retinal photography, since “physiologic” variations in cardiac cycle25 and autonomic nervous system activity26 may influence retinal vessel caliber. In addition, grader variability may have led to further errors in retinal vessel measurement. However, we have no reason to believe these types of variability are differential between myopic and hyperopic eyes. Second, there were few people with extremes in refraction (severe myopia or hyperopia). It is unclear if these data are applicable to other populations with a higher prevalence of myopia or hyperopia. Third, the Bengtsson formula provides only an approximate magnification correction using spherical equivalent refraction. Future studies with axial length data will be important to determine more precisely the impact of magnification error on the absolute measurements of retinal vessel caliber and its relationship to hypertension and other cardiovascular disorders. In summary, we demonstrate that, at a population level, refraction affects the absolute magnitude of retinal vessel diameters measured from fundus photographs. However, refraction does not appear to affect the magnitude of the AVR or influence the direction or strength of association between retinal vessel diameters and blood pressure. Thus, correction for refraction may not be critical in epidemiologic studies examining the association of generalized retinal arteriolar narrowing with hypertension. Future studies with axial length data will be important to determine more precisely the impact of refractive error on the relationship of retinal vessel caliber to hypertension and other cardiovascular disorders.
a hyperopic eye with an axial length of 22 mm, a myopic eye with an axial length of 26 mm could be expected to have retinal vessel diameters that appeared 18% (4 mm divided 22 mm) smaller. This is compatible with our observation that retinal vessel diameters were approximately 20% smaller comparing myopic eyes ⫺6.0 diopters or less with hyperopic eyes 4.0 diopters or more. Correction using the Bengtsson formula eliminated the association of refraction with retinal vessel diameter size. Because refraction was not measured in the ARIC study or the CHS,1– 6 we had used the AVR as an indicator of generalized retinal arteriolar narrowing. In fact, the AVR had been proposed by Wagener and associates20 in 1947 as a practical method to quantify generalized arteriolar narrowing from direct ophthalmoscopy. Others have similarly hypothesized that the AVR would minimize the effect of potential optical magnification error introduced by variation in refraction, since retinas with artificially magnified (or minified) arterioles would likely have similarly magnified (or minified) venules.21–23 We now demonstrate that refraction has no effect on the magnitude of the AVR, which further supports its usefulness as a summary indicator of the relative arteriolar-to-venular diameter, particularly when refraction data are unavailable. In any case, our study shows that refraction does not appear to alter either the pattern or the strength of association between elevated blood pressure and smaller retinal vessel diameters. This finding suggests that association between retinal arteriolar narrowing and hypertension described in the ARIC study and the CHS is unlikely to be biased because of a lack of refraction data.1,2 Our results are also consistent with the Beaver Dam Eye Study, which used an identical computer-assisted method to measure retinal vessel diameters from digitized fundus photographs.24 In that study, after controlling for age and male or female sex, each 10-mm Hg increase in mean arterial blood pressure was associated with a 5.1-m (95% confidence intervals [CI], 4.6 –5.6) decrease in arteriolar diameter. This association was not altered with further adjustment for SER (5.0 m, 95% CI, 4.5–5.5 decrease in arteriolar diameter per 10-mm Hg increase in mean arterial blood pressure). We note that the statistical power to detect associations of retinal vessel diameters appears to be increased by the greater precision in vessel diameter data after correction for refraction, since less of the total variation is a result of measurement error (R2 was higher in the regression models of retinal vessel diameters corrected for refraction). Thus, correction for refraction may increase the chance of finding significant associations of retinal vessel diameters if the effect size is small. This should be taken into consideration in future studies with smaller sample sizes. Strengths of the current study include a populationbased sample less susceptible to selection biases, a standardized assessment of refraction, and a quantitative, computer-assisted measurement of retinal vessel diameters 1054
AMERICAN JOURNAL
REFERENCES 1. Couper DJ, Klein R, Hubbard LD, et al. Reliability of retinal photography in the assessment of retinal microvascular characteristics: the Atherosclerosis Risk in Communities Study. Am J Ophthalmol 2002;133:78 –88. 2. Wong TY, Hubbard LD, Klein R, et al. Retinal microvascular abnormalities and blood pressure in older people: the Cardiovascular Health Study. Br J Ophthalmol 2002;86:1007– 1013. 3. Sharrett AR, Hubbard LD, Cooper LS, et al. Retinal arteriolar diameters and elevated blood pressure: the Atherosclerosis Risk in Communities Study. Am J Epidemiol 1999;150: 263–270. 4. Wong TY, Klein R, Couper DJ, et al. Retinal microvascular abnormalities and incident strokes: the Atherosclerosis Risk in Communities Study. Lancet 2001;358:1134 –1140. 5. Wong TY, Klein R, Sharrett AR, et al. Retinal arteriolar narrowing and incident coronary heart disease in men and women: the Atherosclerosis Risk in Communities Study. JAMA 2002;287:1153–1159. 6. Wong TY, Klein R, Sharrett AR, et al. Retinal arteriolar OF
OPHTHALMOLOGY
JUNE 2004
7. 8. 9. 10. 11. 12. 13.
14. 15. 16.
narrowing and risk of diabetes in middle-aged persons. JAMA 2002;287:2528 –2533. Bengtsson B. The variation and covariation of cup and disc diameters. Acta Ophthalmol (Copenh) 1976;54:804 –818. Bengtsson B, Krakau CE. Correction of optic disc measurements on fundus photographs. Graefes Arch Clin Exp Ophthalmol 1992;230:24 –28. Littmann H. Zur Bestimmung der wahren Gro¨ sse eines Objektes auf dem Hintergrund des lebenden Auges. Klin Monatsbl Augenheilkd 1982;180:286 –289. Rudnicka AR, Burk RO, Edgar DF, Fitzke FW. Magnification characteristics of fundus imaging systems. Ophthalmology 1998;105:2186 –2192. Heier H, Brinchmann-Hansen O. Reliable measurements from fundus photographs in the presence of focusing errors. Invest Ophthalmol Vis Sci 1989;30:674 –677. Attebo K, Mitchell P, Smith W. Visual acuity and the causes of visual loss in Australia: the Blue Mountains Eye Study. Ophthalmology 1996;103:357–364. Leung H, Wang JJ, Rochtchina E, et al. Relationships between age, blood pressure, and retinal vessel diameters in an older population. Invest Ophthalmol Vis Sci 2003;44: 2900 –2904. Sherry L, Wang JJ, Rochtchina E, et al. The reliability of computer-assisted retinal vessel measurement in a population. Clin Exp Ophthalmol 2002;30:179 –182. Attebo K, Ivers RQ, Mitchell P. Refractive errors in an older population: the Blue Mountains Eye Study. Ophthalmology 1999;106:1066 –1072. Mitchell P, Smith W, Attebo K, Healey PR. Prevalence of open-angle glaucoma in Australia: the Blue Mountains Eye Study. Ophthalmology 1996;103:1661–1669.
VOL. 137, NO. 6
REFRACTIVE ERRORS
AND
17. Healey PR, Mitchell P. Optic disk size in open-angle glaucoma: the Blue Mountains Eye Study. Am J Ophthalmol 1999;128:515–517. 18. Mantel N. Chi-square test with one degree of freedom: extensions of the Mantel Haenszel procedure. J Am Stat Assoc 1963;58:690 –700. 19. Arnold JV, Gates JW, Taylor KM. Possible errors in the measurement of retinal lesions. Invest Ophthalmol Vis Sci 1993;34:2576 –2580. 20. Wagener HP, Clay GE, Gipner JF. Classification of retinal lesions in the presence of vascular hypertension. Trans Am Ophthalmol Soc 1947;45:57–73. 21. Stokoe NL, Turner RW. Normal retinal vascular pattern: arteriovenous ratio as a measure of arterial calibre. Br J Ophthalmol 1966;50:21–40. 22. Houben AJ, Canoy MC, Paling HA, et al. Quantitative analysis of retinal vascular changes in essential and renovascular hypertension. J Hypertens 1995;13:1729 –1733. 23. Stanton AV, Mullaney P, Mee F, et al. A method of quantifying retinal microvascular alterations associated with blood pressure and age. J Hypertens 1995;13:41–48. 24. Wong TY, Knudtson MD, Klein R, et al. Computer-assisted retinal vessel diameters in the Beaver Dam Eye Study: methodology, correlation between eyes and effect of refractive error. Ophthalmology (in press). 25. Knudtson MD, Klein BEK, Klein R, et al. Variation associated with retinal vessel diameter measurements controlling for known sources of variability. Br J Ophthalmol 2004;88:57– 61. 26. Lanigan LP, Clark CV, Hill DW. Retinal circulation responses to systemic autonomic nerve stimulation. Eye 1988; 2:412–417.
RETINAL VESSEL DIAMETER
1055