Quantitative evaluation of the optic nerve head in early glaucoma

Abstract
AIMS Progressive loss of neuroretinal rim tissue is known to occur early in glaucoma and measurement of the neuroretinal rim area is possible by magnification corrected analysis of optic disc photographs (planimetry). This study was performed to determine whether the facility to distinguish between glaucomatous and normal optic discs could be improved upon by: (a) taking into account the known relation between optic disc size and neuroretinal rim area, and (b) measuring rim area in a number of segments, in order to detect focal changes. METHODS Planimetric examination of the optic disc photographs of 88 control subjects and 51 patients with early visual field defects was performed. In the control group, multiple linear regression analysis was performed between neuroretinal rim area and optic disc area, age, sex, eye side, refraction, and keratometry. This was repeated for the whole disc and for each of twelve 30 degree segments. Normal ranges were defined by the 98% prediction intervals of the regression analysis and the sensitivity and specificity for correct identification of optic discs in the two groups determined. RESULTS Multiple linear regression demonstrated significant associations between the neuroretinal rim area and optic disc area and age in normal subjects. Sensitivity and specificity for glaucoma diagnosis, using the cut off derived from the 98% prediction intervals, was 37.7% and 98.9% respectively when total neuroretinal rim area alone was considered, and 88.7% and 94.3% respectively when the 30 degree segments were included. The most frequent pattern of neuroretinal rim loss was diffuse, followed by thinning in more than one sector and then by thinning in the inferotemporal sector alone. CONCLUSIONS This method of optic disc analysis enables the examiner to identify glaucomatous optic discs at the stage of early perimetric loss with a high degree of precision. Optic disc photography is simple, and fundus cameras are widely available. This method for glaucoma case identification may therefore be suitable for the primary care setting as well as hospital practice.