Abstract
Purpose: :
Recent studies of patients with ocular hypertension and glaucoma have found that progression occurs as frequently with perimetric measures as with structural measures, when specificity is equated across tests (Artes & Chauhan, 2005, Prog Ret Eye Res 24:333; Strouthidis et al., 2006 IOVS 47:2904). Because these studies assessed progression, it may be difficult for clinicians to apply these conclusions to test results acquired from patients at a single visit. To address this need, we compared the ability of perimetric and imaging tests to detect abnormality in individual optic nerve sectors of patients with established glaucoma.
Methods: :
We tested one eye each of 112 subjects enrolled in a prospective longitudinal study: 62 patients with established glaucoma aged 63 +/- 8 yrs, and 50 control subjects free of eye disease aged 61 +/- 10 yrs. All patients had both optic nerve and visual field damage typical of glaucoma in the tested eye, but few eyes had damage in both inferior temporal (IT) and superior temporal (ST) optic nerve sectors. We analyzed data from the most recent visit on which both optic nerve tomography (HRT-III) and automated perimetry (24-2 SITA Standard) met reliability criteria (rejected if SD on HRT > 30 microns, fixation losses or false positive errors > 25% on perimetry). The optic nerve map of Garway-Heath et al, 2002 (IOVS 43:2213) was used to compute arithmetic means across perimetric locations corresponding to IT and ST optic nerve sectors. Data from control eyes were used to set specificity for each test and sector at 90%, 94% or 98%, and each of the 124 sectors from glaucomatous eyes was scored as either within or outside normal limits.
Results: :
When specificity was set to 90%, perimetry identified 76 sectors in diseased eyes while rim area identified 75 sectors (chi square = 0.02, p > 0.89). Specificity of 94% yielded 65 sectors by perimetry and 56 sectors by rim area (chi square = 1.31, p > 0.25), and specificity of 98% yielded 47 sectors by perimetry and 40 sectors by rim area (chi square = 0.87, p > 0.35).
Conclusions: :
Our results demonstrate that perimetric measures can have similar ability to detect glaucomatous defects as do as imaging measures, at all levels of specificity.
Keywords: perimetry • optic nerve • visual fields