Abstract
Background: Increasing evidence suggests that characteristic changes in retinal vessel morphometry are biomarkers for vascular health. Polymyalgia Rheumatica (PMR) has been associated with increased risk of vascular disease in some epidemiological studies, but to date there are limited data on objective changes in blood vessels in patients with this condition. This study examines the retinal vessel morphology of individuals with PMR in a population-based cohort.
Methods: The study was conducted in participants diagnosed with PMR identified from the EPIC-Norfolk cohort, a population-based study of 25,639 men and women aged 40-79 years, recruited between 1993-1997. Diagnoses of PMR were established through electronic linkage to laboratory, hospital care episodes and questionnaire responses, supplemented by chart reviews during follow-up to 2015. In 2006-2011 digital retinal photography was undertaken and analysed using a validated automated system, which measures vessel width in arterioles and venules respectively. Gaussian modelling was used to determine vessel widths. An ensemble classifier of bagged decision trees was used to classify vessels into probabilities of being either venules or arterioles. Associations were analysed using multi-level linear regression, adjusted for gender and age, allowing for within person clustering.
Results: In total, 10,494 good quality images were obtained from 5959 participants were analysed. Median arteriolar and venular widths were 68.5µm (5th to 95th centiles 49.4, 91.0µm), and 99.1µm (5th to 95th centiles 83.1, 116.4µm) respectively. There were 298 incident diagnoses of PMR (72.5% female) during the follow-up period. The median age at diagnosis was 75.6 years of which 140 had good quality retinal imaging. Those with PMR diagnoses showed wider venules compared to those without disease (mean difference 3.1µm. 95% CI 1.1,5.1µm); there were no appreciable differences in arteriolar diameter (mean difference 0.3µm, 95% CI -2.1,2.8 µm).
Conclusion: These results provide the first objective evidence of a biomarker for vascular health in patients with PMR, a disease in which there is an absence of adequate clinical process and outcome measures. Although limited by the cross-sectional design, the findings indicate that retinal morphometric analysis could prove a useful monitoring tool for vascular complications in patients with rheumatic disease.
Methods: The study was conducted in participants diagnosed with PMR identified from the EPIC-Norfolk cohort, a population-based study of 25,639 men and women aged 40-79 years, recruited between 1993-1997. Diagnoses of PMR were established through electronic linkage to laboratory, hospital care episodes and questionnaire responses, supplemented by chart reviews during follow-up to 2015. In 2006-2011 digital retinal photography was undertaken and analysed using a validated automated system, which measures vessel width in arterioles and venules respectively. Gaussian modelling was used to determine vessel widths. An ensemble classifier of bagged decision trees was used to classify vessels into probabilities of being either venules or arterioles. Associations were analysed using multi-level linear regression, adjusted for gender and age, allowing for within person clustering.
Results: In total, 10,494 good quality images were obtained from 5959 participants were analysed. Median arteriolar and venular widths were 68.5µm (5th to 95th centiles 49.4, 91.0µm), and 99.1µm (5th to 95th centiles 83.1, 116.4µm) respectively. There were 298 incident diagnoses of PMR (72.5% female) during the follow-up period. The median age at diagnosis was 75.6 years of which 140 had good quality retinal imaging. Those with PMR diagnoses showed wider venules compared to those without disease (mean difference 3.1µm. 95% CI 1.1,5.1µm); there were no appreciable differences in arteriolar diameter (mean difference 0.3µm, 95% CI -2.1,2.8 µm).
Conclusion: These results provide the first objective evidence of a biomarker for vascular health in patients with PMR, a disease in which there is an absence of adequate clinical process and outcome measures. Although limited by the cross-sectional design, the findings indicate that retinal morphometric analysis could prove a useful monitoring tool for vascular complications in patients with rheumatic disease.
Original language | English |
---|---|
Number of pages | 2 |
DOIs | |
Publication status | Published - Oct 2016 |
Event | American College of Rheumatology - USA, Washington DC, United States Duration: 12 Nov 2016 → 16 Nov 2016 |
Conference
Conference | American College of Rheumatology |
---|---|
Country/Territory | United States |
City | Washington DC |
Period | 12/11/16 → 16/11/16 |