Objective: To assess the reliability and validity of the tool Automatic image analyzer to assess retinal vessel caliber (ALTAIR) to analyze the vascularization of the retina and cardiovascular risk prediction.
Design and method: Cross-sectional of tools validation study. We included 250 subjects in total, aged 62 +/- 9 years, 51% males. We have made a validation of reliability analyzing the intraclass correlation (ICC) intra observer, inter observer and inter device (compared with AVindex calculator) to the thickness, area and length of arteries and veins of the retina in 3 concentric circles from the disc in 120 retinographies. The concurrent validity was performed with 250 subjects and 497 retinographies, analyzing the relationship with age, blood pressure, parameters of vascular structure and function, renal function and cardiovascular risk estimated with scales.
Results: In the sample, the 32% are obese, 68% hypertensive and 17% diabetic. The interobserver ICC for thickness, area and length of veins and arteries ranged from 0.809 to length of arteries to 0.916 to veins area. The intra observer ICC for intra thickness, area and length of veins and arteries ranged from 0.640 for the length of the veins and 0.906 for the area of the arteries and the inter device ICC was for arteriovenous ratio (AVR) 0.887, thickness of arteries 0.590 and veins thickness 0.677. We found a moderate correlation of the age (r between 0.30 and 0.50, p < 0.001) with the retinal vascular parameters analyzed.
In multiple linear regression analysis after adjusting for age and sex, the association of AVR and arterial thickness with the diastolic blood pressure and albumin creatinine ratio and arterial area and length with systolic arterial blood pressure and carotid intima-media thickness remains. Also the thickness area and length of the vessels show an association with cardiovascular risk estimated SCORE scale.
Conclusions: The ALTAIR tool shows a good reliability in the concordance inter observers, intra observer and inter device measurements and a concordant validity to show an association with vascular parameters, target organ damage and cardiovascular risk.