Retinal microvascular findings and risk of incident peripheral artery disease: An analysis from the Atherosclerosis Risk in Communities (ARIC) Study. (February 2020)

• Record Type: Journal Article
• Authors: Yang, Chao; Kwak, Lucia; Ballew, Shoshana H.; Jaar, Bernard G.; Deal, Jennifer A.; Folsom, Aaron R.; Heiss, Gerardo; Sharrett, A. Richey; Selvin, Elizabeth; Sabanayagam, Charumathi; Coresh, Josef; Matsushita, Kunihiro
• Is Part Of: Atherosclerosis. Volume 294(2020)
• Page Start: 62
• Page End: 71
• Publication Date: 2020-02
• Subjects: Peripheral artery disease -- Critical limb ischemia -- Retinopathy -- Microvascular disease; Arteriosclerosis -- Periodicals; Electronic journals; 616.136
• DOI: 10.1016/j.atherosclerosis.2019.10.012
• Languages: English
• ISSNs: 0021-9150

Abstract

Abstract: Background and aims: Lower-extremity peripheral artery disease (PAD) is usually considered large artery disease. Interestingly, retinal microvascular findings were shown to predict PAD progression in diabetes. However, it is unknown whether retinal microvascular parameters are associated with incident PAD and its severe form, critical limb ischemia (CLI), in a community-based cohort. Methods: Among 9371 ARIC participants (aged 49–72 years) free of a history of PAD, we quantified the associations of several retinal measures by retinal photography during the period 1993–1995 with PAD risk using Cox models. Incident PAD was defined as the first hospitalization with PAD diagnosis or leg revascularization (considered CLI if an additional diagnosis of ulcer, gangrene, or amputation). Results: During a median follow-up of 18.8 years, 303 participants developed PAD (including 91 CLI cases). Although generalized retinal arteriolar narrowing was not associated with PAD, most measures of retinopathy demonstrated strong associations with PAD beyond potential confounders including diabetes, with adjusted hazard ratios (HR) of 3.26 (95% CI 2.18–4.90) for blot-shaped hemorrhages, 3.11 (1.83–5.29) for hard exudates, and 2.18 (1.62–2.95) for any retinopathy. Adjusted HRs were significantly greater for CLI (ranging from 3.2 to 5.9) than for PAD (all p -values <0.05). Retinopathy measures showed particularly strong associations in participants with diabetes ( p -value for interaction [ vs. those without diabetes] <0.001). Conclusions: Several retinopathy measures were strongly associated with PAD, especially with CLI and in diabetes. Our results support the contribution of microvascular abnormalities to the development and progression of PAD and would have implications on its preventive and therapeutic approaches. Graphical abstract: Image 1 Highlights: Retinopathy measures (e.g., hemorrhage and exudates) were robustly associated with incident PAD. The associations were particularly evident for critical limb ischemia (CLI). Retinopathy measures improved risk prediction of PAD and CLI, especially in persons with diabetes.