Advanced vascular exams improve the accuracy of conventional parameters in distinguishing normal tension from primary open angle glaucoma

• Published in: Acta ophthalmologica (Oxford, England) England), 2017-09, Vol.95 (S259), p.n/a
• Main Authors: Barbosa Breda, J., Van Keer, K., Abegão Pinto, L., Nassiri, V., Willekens, K., Vandewalle, E., Rocha Sousa, A., Stalmans, I.
• Format: Article
• Language: English
• Subjects: Acuity; Blood pressure; Cornea; Doppler effect; Eye; Glaucoma; Headache; Migraine; Oximetry; Oxygen content; Prediction models; Pulse amplitude; Raynaud disease; Retina; Saturation; Tension; Veins & arteries; Velocity; Visual acuity; Visual field; Visual fields
• ISSN: 1755-375X; 1755-3768
• DOI: 10.1111/j.1755-3768.2017.01123

Purpose The role vascular factors play in glaucoma is still not fully understood. We aim to understand if advanced vascular exams can improve our ability to distinguish primary open angle glaucoma (POAG) from normal tension glaucoma (NTG) patients. Methods We applied multivariate logistic regression on the Leuven Eye Study data (182 NTG and 202 POAG patients), and created three prediction models. One with conventional parameters, one with parameters from advanced vascular exams [colour Doppler imaging (CDI), retinal oximetry, ocular pulse amplitude and choroidal thickness] and one where both kinds of parameters were allowed in (mixed). The area under the curve (AUC) was calculated and compared between them. Intra‐ocular pressure (IOP) and IOP‐related variables were excluded. Results The conventional model included Raynaud phenomenon, mean arterial pressure (MAP), central corneal thickness, migraine, visual field mean deviation (VF MD), cup‐disc (C/D) ratio, visual acuity and mean retinal nerve fiber layer (mRNFL) thickness (in decreasing order of importance). The advanced vascular model included parameters from CDI and retinal oximetry. The mixed model included Raynaud phenomenon, MAP, arterio‐venous retinal oxygen saturation difference, acceleration (acc) of central retinal artery (CRA), end‐diastolic velocity of CRA, resistive index of ophthalmic artery, VF MD, C/D ratio, mRNFL thickness, acc of temporal posterior ciliary artery and migraine (in decreasing order of importance). The AUCs were 0.687, 0.677 and 0.743, respectively. The mixed model was able to significantly increase the prediction ability of the conventional model (p = 0.049). Conclusions Adding advanced vascular exams increased the accuracy of the conventional model to distinguish between groups. These results show how important vascular factors can be in the pathogenesis of glaucoma.