Abstract

P 119

Moorfields Regression Analysis (MRA) vs glaucoma probability score in patients with open-angle-glaucoma, ocular hypertension and normal subjects

Charlotte Knipper, Norbert Pfeiffer, Esther M. Hoffmann
Universitäts-Augenklinik Mainz

Objective
To compare the Moorfields Regression Analysis (MRA) with the Glaucoma Probability Score (GPS) with regard to their diagnostic accuracy, their agreement and the effect of different ophthalmologic parameters.
Methods
This prospective study included one eye of 50 patients with open-angle glaucoma, of 20 patients with ocular hypertension and of 30 normal subjects. All participants underwent a complete ophthalmic examination including medical history. The diagnostic accuracy of MRA and GPS was evaluated based on the clinical examination of the optic nerve head. Agreement between GPS and MRA and the effect of clinical parameters on MRA and GPS were analyzed.
Results
GPS is more sensitive but less specific than MRA. The highest sensitivity was 0.84 (global, GPS) resp. 0.54 (temp.- inf., nasal-inf. sector, MRA). The best relation between sensitivity and specifity was obtained when regarding a borderline classification as a negative result for GPS and as a positive result for MRA. Neither MRA nor GPS were able to differentiate between ocular hypertensive and normal eyes (p-value > 0.05). Complete or partial agreement between MRA and GPS was obtained in ~90% of cases. With the GPS, spherical equivalent (OR:1.714 – 2.446) and optic disc size (OR:1.175 – 1.254) had an effect on the classification. Severity of disease (global, temp.-sup., temp.-inf., nasal, nasal-inf.) and optic disc size (global, nasal, nasal-inf., final classification) were found to have an effect on some sectors of the MRA.
Conclusions
Both GPS and MRA are able to differentiate between normal and glaucomatous eyes, whereby GPS is more sensitive. Yet the agreement between GPS and MRA is high. When using the HRT the clinician should take optic disc size and spherical equivalent into account. The results of the HRT should always be evaluated in relation to other clinical data.

Copyright: 2009 Porstmann Kongresse GmbH. All rights reserved. Impressum, Disclaimer.