Purchase this article with an account.
P. Mitchell, J.J. Wang, W. Smith, R. Klein, T.Y. Wong, S.R. Leeder, Blue Mountains Eye Study; Retinal Vascular Changes Predict Subsequent Cerebrovascular Events: Blue Mountains Eye Study Findings . Invest. Ophthalmol. Vis. Sci. 2003;44(13):2053.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: To assess the relationship between retinal vascular changes assessed from baseline photographs and subsequent stroke and cerebrovascular death in an older population. Methods: Blue Mountains Eye Study participants (n=3654, aged 49+ years, seen during 1992-94, were re-examined during 1997-99. At baseline, 35mm retinal slides were assessed using a Donaldson viewer and light-box to grade presence of focal arteriolar narrowing, arteriovenous (A/V) nicking and retinopathy lesions (microaneurysms or haemorrhages, in non-diabetic subjects). A computer-assisted method was used to assess retinal vessel diameters from the retinal photographs from each participant’s right eye. Summarized estimates of central retinal arteriolar equivalent (CRAE), representing the average arteriolar diameter and arteriole-to-venule ratio (AVR), were calculated. The lowest quintile of CRAE represents generalized arteriolar narrowing. Deaths and causes of death were obtained from Australian National Death Index. Surviving incident stroke cases were confirmed from medical records (data on clinical presentation or CT scans). Deaths and incident strokes were considered incident cerebrovascular events. Results: A marked age-related increase in the baseline prevalence of retinal vascular lesions was found. As age increased from <60 years to 80+ years, the prevalence of focal arteriolar narrowing increased from 2.0% to 22%; A/V nicking prevalence increased from 5.5% to 10.2%; prevalence of retinopathy lesions in non-diabetic subjects increased from 7.4% to 22% and the narrowest quintile of CRAE increased from 15% to 32%. Presence at baseline of any of these vascular changes doubled the 5-year risk of subsequent cerebrovascular events: age adjusted relative risks were 2.3 (95% confidence intervals 1.4-3.7) for focal narrowing, 2.0 (1.1-3.6) for AV nicking, 2.9 (1.8-4.8) for presence of retinopathy lesions and 2.8 (1.0-8.2) for the narrowest quintile of CRAE. AVR was not associated with subsequent cerebrovascular events in this population. Further adjustment for sex, systolic blood pressure, smoking, self-rated health and diabetes in Cox regression models did not alter the associations. Conclusions: These findings support previous findings from other populations that retinal vascular changes predict subsequent cerebrovascular events, independent of age, blood pressure, smoking and diabetes.
This PDF is available to Subscribers Only