Abstract

SO.14.03

Agreement among three clinical bleb-grading systems

Daniel Herzog, Joanna Wasielica-Poslednik, Christina Müller, Norbert Pfeiffer, Esther M. Hoffmann
Universitäts-Augenklinik Mainz

Objective
Using a bleb-grading system clinically facilitates long-term follow-up of patients with previous glaucoma surgery. Clinical evaluation of these patients can be challenging for untrained ophthalmologists.
In this study the newly developed MaBAGS (Mainz Bleb Appearance Grading System), a classification system for filtering blebs according to clinical morphologic parameters, is compared with well-known classification systems (MBGS/Moorfields Bleb Grading System, IBAGS/Indiana Bleb Appearance Grading Scale) in reference to usability and reliability.
Methods
Forty-two glaucoma filtering blebs of 31 patients with a wide range of appearances were included. Two experienced and one inexperienced observer graded blebs using MaBAGS, MBGS and IBAGS during slit-lamp examination. Analysis was performed to determine consistency between the observers using intraclass correlation coefficients (ICC).
Results
In all grading systems, consistency was considerably better between the two experienced observers compared to the inexperienced grader.
Using MaBAGS excellent levels of agreement were found for vascularity indices (0.82 and 0.82) and good agreement for transparency (0.69). Parameters for area and height yielded moderate agreement (0.47 and 0.59), while indices for mobility and microcysts failed to show satisfying levels of agreement (<0.20). Using MBGS resulted in excellent and good interobserver consistency for parameters regarding subconjunctival blood (0.93), central and peripheral vascularity (0.84 and 0.74). Indices for height and non-bleb vascularity reached moderate levels of agreement (0.53 and 0.60). Levels of agreement for area-parameters were low (<0.40). With IBAGS, good levels of agreement were found for height and vascularity (0.61 and 0.73) and moderate for extent (0.44).
Conclusions
MaBAGS shows good reproducibility. Using more clinical parameters improves precision of the description of a highly variable clinical finding. The experience of the observer influences grading quality and hence interobserver consistency considerably.

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