Diabetes mellitus is recognized as a stroke risk factor in atrial fibrillation (AF). Patients with diabetes with retinopathy have an increased risk for systemic cardiovascular complications, and severe diabetic retinopathy predisposes to ocular bleeding. We hypothesized that patients with diabetes, retinopathy, and AF have increased stroke/thromboembolism (TE) and severe bleeding risks when compared with patients with diabetes and AF who do not have retinopathy or to patients with AF and without diabetes.
We tested our hypothesis in a large “real-world” cohort of individuals with AF from the Loire Valley Atrial Fibrillation project.
Of 8,962 patients with AF in our dataset, 1,409 (16%) had documented diabetes mellitus. Of these, 163 (1.8% of the whole cohort) were patients with diabetic retinopathy. After a follow-up of 31 ± 36 months, when compared with patients without diabetes, the risk of stroke/TE in patients with diabetes with no retinopathy increased 1.3-fold (relative risk [RR], 1.30; 95% CI, 1.07-1.59; P = .01); in patients with diabetes with retinopathy, the risk of stroke/TE was increased 1.58-fold (RR, 1.58; 95% CI, 1.07-2.32; P = .02). There was no significant difference when patients with diabetes with no retinopathy were compared with patients with diabetes with retinopathy (RR, 1.21; 95% CI, 0.80-1.84; P = .37). A similar pattern was seen for mortality and severe bleeding. On multivariate analysis, the presence of diabetic retinopathy did not emerge as an independent predictor for stroke/TE or severe bleeding.
Crude rates of stroke/TE increased in a stepwise fashion when patients without diabetes and with AF were compared with patients with diabetes with no retinopathy and patients with diabetes with retinopathy. However, we have shown for the first time, to our knowledge, that the presence of diabetic retinopathy did not emerge as an independent predictor for stroke/TE or severe bleeding on multivariate analysis.