SAT0641 RETINAL VASCULOMETRICS AS A POTENTIAL BIOMARKER FOR RISK OF POLYMYALGIA RHEUMATICA. (June 2019)

• Record Type: Journal Article
• Authors: Yates, Max; Peto, Tunde; Rudnicka, Alicja; Welikala, Roshan; Macgregor, Alex; Khawaja, Anthony; Watts, Richard; Broadway, David; Hayat, Shabina; Luben, Robert; Barman, Sarah; Owen, Christopher; Khaw, Kay-Tee; Wareham, Nick; Foster, Paul
• Is Part Of: Annals of the rheumatic diseases. Volume 78(2019)Supplement 2
• Page Start: 1418
• Page End: 1418
• Publication Date: 2019-06
• Subjects: Rheumatism -- Periodicals; 616.723005
• DOI: 10.1136/annrheumdis-2019-eular.3332
• Languages: English
• ISSNs: 0003-4967

Abstract

Abstract : Background: Vascular risk factors are implicated in the aetiology of polymyalgia rheumatica (PMR) but lack sensitivity as a screening tool to assess risk of future disease or disease progression. Retinal vascular measures from digital images of the retinal fundus provide a direct objective measure of vascular health and have potential value in estimating future disease risk. Objectives: The aim of this study was to determine if quantitative changes on retinal vasculometric imaging could be detected prior to the onset of PMR among a large prospective population-based cohort with a view to assessing their clinical utility as a biomarker of disease risk. Methods: Known risk factors for vascular disease were obtained at enrolment to the EPIC-Norfolk cohort in 1993, with data available on BMI, blood pressure, smoking status and blood drawn for cholesterol and high-sensitivity C-reactive protein. Eye screening took place between 2004 to 2011. Cases of PMR were identified retrospectively among 8112 volunteers by: (i) free text questionnaire responses at baseline, 18 months, 3, 10 and 13 years; (ii) linkage to hospital electronic discharge summaries containing ICD codes (iii) linkage to key word searches ( polymyalgia or rheumatica) of out-patient clinic letters. Case definition conformed to validated methods. Measures of retinal vessel width were generated by an automated software programme (QUARTZ), using Gaussian modelling to determine vessel widths and an ensemble classifier of bagged decision trees to classify vessels into either venules or arterioles with greater than 80% probability. Stepwise Cox modelling was used to investigate vascular risk factors, including vasculometric characteristics, and subsequent development of PMR, within a multivariate adjusted model. Results: We identified 71 cases of PMR (median age at diagnosis: 70.5 years, minimum 50.5 years, maximum 87.0 years; mean ESR at diagnosis: 52 mm/hr; 67.6% female) amongst 5130 participants who attended ocular screening and had high quality retinal images acceptable for analysis. There were 50 incident diagnosis of PMR with vascular risk data recorded at enrolment into EPIC-Norfolk prior to their diagnosis of PMR. Results from the univariate Cox modelling revealed positive associations for age, female sex, low-density lipoprotein (LDL) and retinal venular width. There were no associations for arteriolar measures. After adjustment a positive association remained for age and venular dilatation: the hazard ratio for age was 1.14 (95% CI 1.09, 1.18) p<0.0001 and for venule width dichotomised at 90 microns was 2.88 (1.47, 5.66) p=0.002. Conclusion: Retinal venular dilatation is a marker of vascular health and within this data an independent risk factor for subsequent development of PMR. Retinal fundus images are easily captured in the clinical setting. While validation of these findings in other datasets is warranted, they provide a potential biomarker for assessment of PMR risk. Disclosure of Interests: None declared