Retinal Arteriolar Narrowing and Risk of Diabetes Mellitus in Middle-aged Persons
- Tien Yin Wong, MD, MPH;
- Ronald Klein, MD, MPH;
- A. Richey Sharrett, MD, DrPH;
- Maria I. Schmidt, MD;
- James S. Pankow, PhD;
- David J. Couper, PhD;
- Barbara E. K. Klein, MD, MPH;
- Larry D. Hubbard, MAT;
- Bruce B. Duncan, MD, PhD;
- for the ARIC Investigators
- Author Affiliations: Department of Ophthalmology, University of Wisconsin-Madison (Drs Wong, R. Klein, B. Klein, and Mr Hubbard); Singapore National Eye Center and Department of Ophthalmology, National University of Singapore (Dr Wong); National Heart, Lung, and Blood Institute, Bethesda, Md (Dr Sharrett); Social Medicine Department, School of Medicine, Federal University of Rio Grande do Sul, Porto Alegre RS, Brazil (Drs Schmidt and Duncan); Division of Epidemiology, University of Minnesota, Minneapolis (Dr Pankow); and Department of Biostatistics, University of North Carolina, Chapel Hill (Dr Couper).
Abstract
Context Microvascular processes have been hypothesized to play a role in the pathogenesis of type 2 diabetes mellitus, but prospective clinical data regarding this hypothesis are unavailable.
Objective To examine the relation of retinal arteriolar narrowing, a marker of microvascular damage from aging, hypertension, and inflammation, to incident diabetes in healthy middle-aged persons.
Design, Setting, and Participants The Atherosclerosis Risk in Communities Study, an ongoing population-based, prospective cohort study in 4 US communities that began in 1987-1989. Included in this analysis were 7993 persons aged 49 to 73 years without diabetes, of whom retinal photographs were taken during the third examination (1993-1995).
Main Outcome Measures Incident diabetes (defined as fasting glucose levels of ≥126 mg/dL [7.0 mmol/L], casual levels of ≥200 mg/dL [11.1 mmol/L], diabetic medications use, or physician diagnosis of diabetes at the fourth examination) by quartile of retinal arteriole-to-venule ratio (AVR).
Results After a median follow-up of 3.5 years, 291 persons (3.6%) had incident diabetes. The incidence of diabetes was higher in persons with lower AVR at baseline (2.4%, 3.1%, 4.0%, and 5.2%, from highest to lowest AVR quartile; P for trend<.001). After controlling for fasting glucose and insulin levels, family history of diabetes, adiposity, physical activity, blood pressure, and other factors, persons in the lowest quartile of AVR were 71% more likely to develop diabetes than those in the highest quartile (odds ratio [OR], 1.71; 95% confidence interval [CI], 1.13-2.57; P for trend = .002). This association persisted with different diagnostic criteria (OR, 1.92; 95% CI, 1.10-3.36; P for trend = .01, using a fasting glucose level of ≥141 mg/dL [7.8 mmol/L] as a cutoff), and was seen even in people at lower risk of diabetes, including those without a family history of diabetes, without impaired fasting glucose, and with lower measures of adiposity.
Conclusions Retinal arteriolar narrowing is independently associated with risk of diabetes, supporting a microvascular role in the development of clinical diabetes.
