Abstract
Purpose::
To assess whether baseline retinal arteriolar narrowing predicts incident open angle glaucoma in an older population.
Methods::
The Blue Mountains Eye Study (BMES) examined 3,654 participants aged 49+ years at baseline (1992-4). Retinal photographs were digitised and retinal arteriolar and venular diameters measured. Open angle glaucoma (OAG) was diagnosed from matching visual field defects and optic disc cupping, without reference to intra-ocular pressure (IOP) level. Associations between incident OAG and arteriolar or venular diameter were assessed after adjusting for baseline age, gender, IOP, use of glaucoma medications, vertical disc diameter, cup-to-disc ratio and baseline disc haemorrhage. Vessel diameter was analysed continuously and by tertile.
Results::
After excluding subjects with glaucoma or optic disc abnormalities at baseline or those with incomplete data, 2328 participants seen at the 5- and/or 10-year visits were considered at risk of incident OAG. This developed in 99 eyes of 78 participants (3.4%), at either the 5- or 10-year follow-up exam. After adjusting for age and sex, narrower baseline arteriolar and venular diameter was associated with an increased risk of incident OAG (narrowest vs widest tertile, odds ratio, OR 2.32, 95% confidence interval, CI, 1.28- 4.20, p for trend 0.004 for arterioles and OR 1.76, CI 1.01- 3.05, p for trend 0.04 for venules). Each standard deviation (SD) decrease in vessel diameter was associated with an increased risk of incident OAG (OR 1.36 CI 1.07- 1.73 for arterioles and OR 1.28 CI 1.07- 1.73 for venules). After further adjusting for IOP, use of glaucoma medications, vertical disc diameter, cup-to-disc ratio and baseline disc haemorrhage, the associations with arteriolar narrowing remained but became borderline non-significant (OR 1.86, CI 1.00- 3.61, p for trend 0.057 for the lowest vs highest tertile, and OR 1.24, CI 0.94- 1.65 per SD decrease). The association with narrowed venular diameter was no longer present.
Conclusions::
These findings suggest that retinal arteriolar narrowing may precede the development of OAG. This adds to previously reported BMES data which demonstrated a strong cross-sectional association between OAG and retinal arteriolar narrowing.
Keywords: clinical (human) or epidemiologic studies: prevalence/incidence