The association between retinal vascular geometry changes and diabetic retinopathy and their role in prediction of progression – an exploratory study

BMC Ophthalmology The association between retinal vascular geometry changes and diabetic retinopathy and their role in prediction of progression - an exploratory study Maged S Habib Bashir Al-Diri Andrew Hunter David HW Steel Background: The study describes the relationship of retinal vascular geometry (RVG) to severity of diabetic retinopathy (DR), and its predictive role for subsequent development of proliferative diabetic retinopathy (PDR). Methods: The research project comprises of two stages. Firstly, a comparative study of diabetic patients with different grades of DR. (No DR: Minimal non-proliferative DR: Severe non-proliferative DR: PDR) (10:10: 12: 19). Analysed RVG features including vascular widths and branching angles were compared between patient cohorts. A preliminary statistical model for determination of the retinopathy grade of patients, using these features, is presented. Secondly, in a longitudinal predictive study, RVG features were analysed for diabetic patients with progressive DR over 7 years. RVG at baseline was examined to determine risk for subsequent PDR development. Results: In the comparative study, increased DR severity was associated with gradual vascular dilatation (p = 0.000), and widening of the bifurcating angle (p = 0.000) with increase in smaller-child-vessel branching angle (p = 0.027). Type 2 diabetes and increased diabetes duration were associated with increased vascular width (p = <0.05 In the predictive study, at baseline, reduced small-child vascular width (OR = 0.73 (95% CI 0.58-0.92)), was predictive of future progression to PDR. Conclusions: The study findings suggest that RVG alterations can act as novel markers indicative of progression of DR severity and establishment of PDR. RVG may also have a potential predictive role in determining the risk of future retinopathy progression. Diabetic retinopathy; Retinal vascular geometry; Retinal vascular analysis; Retinal bifurcations - Background Diabetic retinopathy (DR) is a leading cause of visual loss in many developed countries [1,2]. Given that eyes with severe non proliferative diabetic retinopathy (NPDR) have a 52% risk of developing PDR within one year and 60% risk of developing high risk PDR within 5 years [3] there is a need in clinical practice for effective risk stratification. Structural and functional changes in the retinal vasculature have been shown in several studies to be closely related to diabetes and DR [4,5]. Populationbased cohort studies have described various vascular * Correspondence: 1Sunderland Eye Infirmary Queen Alexandra Road, Sunderland SR2 9HP, UK Full list of author information is available at the end of the article calibre changes occurring with the development of diabetes [6], or in diabetics with no retinopathy [7], as well as with the development of early retinopathy in type 1 and type 2 diabetics [1,2,8,9], and indeed the progression of DR [7,10-13]. Some of the retinal vascular changes were also found to be significantly associated with progression risk to PDR while controlling for other baseline risk factors [14]. Changes in the retinal vascular calibre represent only a sole parameter of the retinal vascular network and do not convey information regarding the complexity of the retinal vascular branching pattern. According to Murray [15], the optimal vascular architecture achieves the most efficient blood flow transport with minimum energy allowing for maximum vascular diffusion into the surrounding tissues. Alterations in the geometry of the retinal vascular network may thus reflect a state of vascular dysfunction, and might potentially predict disease development [16-19]. Several features representative of the retinal microvasculature geometry have been evaluated in association with various systemic conditions [16-18,20-22]. Being non-dimensional, these features are less likely to be affected by variations in digital image resolution, ocular magnification or differences in refractive errors. In diabetes, recent studies have explored the role of changes in retinal vascular geometry (RVG) as risk markers for incident DR in young type 1 diabetics. Sasongko et al. [23] demonstrated that certain demographic factors such as age and sex as well as diabetes-related factors such as type, and the duration of diabetes were associated with subtle alterations in the RVG parameters including the branching angles. Moreover, other RVG features predicted incident retinopathy after a median of 3.8 years follow up [24]. In this study, we hypothesise that RVG changes are associated with increased severity of DR, and that, alterations in RVG can act as a novel marker in identifying networks at risk of future progression to proliferative retinopathy. We measure RVG at vascular bifurcations and analyse the relationship of these to progression of DR. Methods Study population The study was designed in two parts. Firstly a comparative observational study to evaluate the associations of RVG changes in age-matched cohorts of diabetic subjects with different grades of DR. Within this study, the relationships of certain demographic and clinical risk factors with RVG changes in the diabetic population were analyzed. Secondly, a pilot study of longitudinal data to assess the predictive role of RVG changes for subsequent development of proliferative retinopathy, using retrospective data for patients newly presenting with PDR. The studies were carried in accordance with the declaration of Helsinki (1989) of the World Medical Association. The protocols were approved by the local Sunderland Research Committee (SLREC 1129) and all recruited subjects gave written informed consent. Study participants were recruited from patients attending the DR clinic at Sunderland Eye Infirmary and the local diabetic retinopathy screening programme centre. Type I or II diabetic patients aged 25 65 years, with different grades of DR and no previous pan retinal laser photocoagulation treatment were invited to participate in the study. The retinopathy grade was classified into; clinically non-detectable, minimal non-proliferative, severe non-proliferative and proliferative diabetic retinopathy. Exclusion criteria included refractive error beyond +/ 3 dioptres, concurrent ocular pathology including cataract, glaucoma, corneal opacities or any other retinal and optic nerve head pathology. Patients with uncontrolled systemic hypertension with documented blood pressure measurements of more than 140/90 repeatedly recorded within the previous 12 months were also excluded. For the predictive study, patients who were referred with PDR from the diabetic retinopathy screening service to the hospital between January 2008 January 2009 were identified. Amongst these cases, patients aged 65 years or less who presented initially to the local diabetic retinopathy screening service with no clinically detectable retinopathy and with at least preceding 6-years screening follow-up period were (...truncated)