Abstract
Purpose :
To investigate the predictive value of retinal nerve fiber layer defects (RNFLDs) as a risk factor for an incident stroke in 5 years.
Methods :
The Kailuan Cohort Study is a community-based study, performed in Tangshan, China. A total of 14,441 participants (male: 9835 (68.1%); mean age: 53 ± 13 years (range: 21 - 108 years)) have been enrolled since the baseline in 2015,. They underwent a series of ophthalmic and systematic examinations. Using fundus photographs a localized RNFLD was defined as a wedge-shape, not spindle-shaped, defect in the retinal nerve fiber layer, running in direction to and not necessarily reaching the optic disc border. Incident cerebrovascular events including stroke were recorded till December 2019. Cox proportional hazards models were used to assess the relationship between the RNFLD prevalence at baseline and the incidence of stroke in a multivariable model.
Results :
Within a follow-up of 5 years (mean: 44 ± 12 months), 291 participants developed a cerebral stroke (2.0% ± 0.1%). After excluding those individuals with glaucoma, optic neuropathy, diffuse RNFL loss, and those with unevaluable fundus photographs, 215 participants with incident stroke were compared with a randomly selected control group (n = 2147). Incident stroke occurred in the 22nd month as mean (95% confidence interval (CI): 20 - 24 months) after enrollment. The RNFLD prevalence at baseline was significantly more frequently observed in the group with incident stroke than in the control group (83/215 (38.6% ± 3.3%) versus 296/2147 (13.8% ± 0.7%); Odds ratio: 3.93; P < 0.001). Cox proportional hazards model demonstrated a statistically increased risk of stroke in participants with RNFLDs (P < 0.001), with a hazards ratio (HR) of 3.87 (95% CI: 2.45 - 6.13), after controlling of age, sex, and comorbidity of diabetes or hypertension.
Conclusions :
This study demonstrated the presence of RNFLDs as a biomarker for an increased risk of future cerebral strokes.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.