April 2010
Volume 51, Issue 13
ARVO Annual Meeting Abstract  |   April 2010
Association of Retinal Vascular Fractal Dimension and Retinopathy Lesions in Older People Without Diabetes
Author Affiliations & Notes
  • P. Mitchell
    Ophthalmology, University of Sydney, Sydney, Australia
  • E. Rochtchina
    Ophthalmology, University of Sydney, Sydney, Australia
  • G. Liew
    Ophthalmology, University of Sydney, Sydney, Australia
  • H. Li
    Ophthalmology, University of Sydney, Sydney, Australia
  • T. Y. Wong
    Singapore Eye Research Institute, National University of Singapore, Singapore, Singapore
  • J. Wang
    Ophthalmology, University of Sydney, Sydney, Australia
  • Footnotes
    Commercial Relationships  P. Mitchell, None; E. Rochtchina, None; G. Liew, None; H. Li, None; T.Y. Wong, None; J. Wang, None.
  • Footnotes
    Support  Aust NHMRC ####
Investigative Ophthalmology & Visual Science April 2010, Vol.51, 1242. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      P. Mitchell, E. Rochtchina, G. Liew, H. Li, T. Y. Wong, J. Wang; Association of Retinal Vascular Fractal Dimension and Retinopathy Lesions in Older People Without Diabetes. Invest. Ophthalmol. Vis. Sci. 2010;51(13):1242.

      Download citation file:

      © ARVO (1962-2015); The Authors (2016-present)

  • Supplements

Purpose: : To assess whether retinopathy in people without diabetes is associated with retinal vascular fractal dimension.

Methods: : Of 3654 baseline participants aged 49+, 2335/ 3111 (75% of survivors) were examined after 5 yrs, and 1935/ 2581 (75.0% of survivors) after 10 yrs. Retinopathy lesions (hemorrhages, microaneurysms, hard/ soft exudates) were assessed from stereo retinal photos. Diabetes (defined by history/ fasting glucose) was excluded. Retinal vascular fractal dimension (Df) measures vascular bed density (or fullness), using image analysis software and disc-centered retinal photos. In right-eye data, associations between Df and prevalent or incident retinopathy were examined.

Results: : 197/ 3041 (6.5%) non-diabetic participants had isolated retinopathy lesions (right eye) at baseline. After excluding these and people with ungradable photos, 1588 were at risk of incident retinopathy at 5 or 10 yrs. Overall cumulative retinopathy incidence was 9.9% in right eyes (similar by gender), increasing from 6.9% for ages <60 to 32.6% for ages 80+. Prevalent retinopathy increased from 4.4% to 12.6% at these ages. Df decreased with age (i.e. vascular bed became less dense); the proportion with the lowest Df quartile increased from 12.5% (age <60) to 63.0% (80+). After adjusting for age, sex, body mass, smoking, mean arterial BP and retinal vessel caliber, we found a U-shaped association between Df and either prevalent or incident retinopathy. Compared to those in middle 2 Df quartiles, both those in lowest (least dense) and highest (most dense) quartiles had increased odds of prevalent retinopathy (odds ratio, OR, 2.2, 95% confidence interval, CI 1.5-3.2 for the lowest, and OR 1.7, 95% CI 1.1-2.5 for the highest quartile). They also had an increased risk of incident retinopathy over 10 yrs (OR 1.4, 95% CI 0.9-2.2 for the lowest, and OR 1.8, 95% CI 1.2-2.7 for the highest quartile). These relationships with prevalent and incident retinopathy were stronger in people with hypertension or obesity.

Conclusions: : Either low or high retinal vascular fractal dimension was associated with isolated retinopathy in non-diabetic persons, cross-sectionally and longitudinally. This suggests that Df may measure microcirculatory optimality or suscepibility. A similar U-shaped relationship was found between Df and incident coronary-heart disease mortality over 13 yrs in the BMES.

Keywords: clinical (human) or epidemiologic studies: risk factor assessment • blood supply • imaging/image analysis: clinical 

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.