Abstract
Purpose :
To test the hypothesis that diabetic retinopathy severity is an independent risk factor for stroke in diabetic patients.
Methods :
The original cohort from the Framingham Heart Study (FHS) was screened for possible eye disease, and 2,675 participants completed eye exams in the Framingham Eye Study (FES). Of those, 2,194 completed eye exams in 1973-1975, within a year of examination 13, had all of the components of the FSRP, and were stroke-free. Diabetic retinopathy (DR) severity was classified based on modified Early Treatment for Diabetic Retinopathy Study (ETDRS) criteria. Incident strokes were documented for the ensuing ten years. Cox proportional hazards analyses were used to relate DR severity to stroke incidence, adjusted for age and sex and for the Framingham Stroke Risk Profile (FSRP).
Results :
The mean age of participants was 65 years (range 51-85); 59% were female. Of these participants, 9.1% (199/2194) developed diabetes, 1.32% (29/2194) had DR, and 6.5% (142/2194) suffered a stroke within the ten-year epoch after the eye exam. The percent who developed stroke in ten years increased from 6.3% (136/2169) in participants with absent or mild non-proliferative DR (NPDR) to 11.1% (2/18) in those with moderate to severe NPDR and 57% (4/17) in those with proliferative DR (PDR). Among participants with diabetes, those with PDR were significantly more likely to have a stroke within ten years compared to those with absent or mild NPDR, after adjustment for age and sex (HR=5.08; P=0.010; 95% CI=[1.48-17.48]). After adjustment for the FSRP score, the hazards ratio was no longer statistically significant (P=0.074).
Conclusions :
Increased severity of diabetic retinopathy may be a marker for stroke as it was associated with increased incidence of stroke in diabetic participants in the original FHS cohort. After adjusting for the FSRP, this association was much smaller and not statistically significant; however, the incidence of any DR during this examination was low (29 cases). Additional eye exams or more sensitive metrics for microvascular changes might lead to a higher incidence of DR and improved ability to detect its effect on stroke risk. If the above association of DR severity with stroke is true, then diabetic patients with more severe DR might benefit from more aggressive stroke prevention.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.