RESUMO
Epidemiological studies suggest retinal microvascular abnormalities predict cardiac events. This study examined microvascular features associated with coronary artery abnormalities. This was a single-centre, cross-sectional, observational study of 144 consecutive subjects undergoing coronary angiography for clinical indications. Their angiograms were deidentified and graded for disease (Leaman score, LAD stenosis ≥ 70%, number of vessels stenosed ≥ 70%), and Thrombolysis in Myocardial Infarction (TIMI) blush score. Subjects also underwent retinal photography (KOWA non-mydriatic camera, Japan), and their deidentified retinal images were graded for hypertensive microvascular retinopathy (Wong and Mitchell classification), vessel calibre using a computer-assisted method (IVAN, U Wisconsin), and diabetic retinopathy (modified Airlie House scheme) independently by a trained grader and an ophthalmologist. Retinal abnormalities were compared between subjects with high and low angiography scores using one way ANOVA, Chi squared and logistic regression analysis (StataCorp, Texas). Subjects had a mean age of 61 years (range 32-88), and included 101 males (70%). Seventeen (12%) had Leaman scores > 10.5, 46 (32%) had LAD stenosis, 13 (9%) had ≥ 3 arteries stenosed, and 20 (14%) had TIMI blush scores < 1. Twenty-six subjects (18%) had a retinal hemorrhage, and 115 (74%) a mild or moderate hypertensive retinopathy. Fifty-five (38%) had diabetes, and 24 (17%) a background (n = 20) or proliferative (n = 4) diabetic retinopathy. A retinal hemorrhage (p = 0.046), moderate microvascular retinopathy (p = 0.08) and proliferative diabetic retinopathy (p = 0.04) were all associated with a higher Leaman score. Venular calibre was increased with triple vessel disease (205.7 ± 21.6 µm, and 193.7 ± 22.3 µm in normals, p = 0.03). Diabetic retinopathy correlated with an increased TIMI blush score (p = 0.01). Retinal microvascular imaging warrants further evaluation in identifying the presence, extent and nature of coronary artery disease.
