Structure and function evaluation (SAFE): I. Criteria for glaucomatous visual field loss using standard automated perimetry (SAP) and short wavelength automated perimetry (SWAP)

PURPOSE: To develop criteria for detecting glaucomatous visual field loss for standard automated perimetry (SAP) and short wavelength automated perimetry (SWAP). DESIGN: Longitudinal observational study. METHODS: Three populations were evaluated: (1) 348 normal subjects (348 eyes) were tested to develop nor, mative databases and statistical analysis packages for SAP and SWAP. (2) An independent group of 47 normal subjects (94 eyes) with 4 years of longitudinal follow-up was evaluated to determine specificity of different criteria. (3) A group of 298 patients (479 eyes) with elevated intraocular pressure and normal baseline SAP visual fields were evaluated to determine the sensitivity of different criteria for detecting early glaucomatous visual field loss. RESULTS: Six criteria demonstrated high specificity for correctly identifying eyes with normal visual fields (98%-100%) for both SAP and SWAP: (1) a pattern standard deviation (PSD) worse than the normal 1% level, (2) a glaucoma hemifield test (GHT) "outside normal limits," (3) one hemifield cluster worse than the normal 1% level, (4) two hemifield clusters worse than the normal 5% level, (5) four abnormal (P < .05) locations, (6) five abnormal locations (P < .05) on the pattern deviation probability plot. For all criteria, confirmation on a second visual field was required for high specificity. The GHT "outside normal limits," two hemi. field clusters worse than the normal 5% level and four abnormal (P < .05) test locations on the pattern deviation probability plot provided the highest percentages of conversion from a normal to a glaucomatous visual field. CONCLUSIONS: Criteria based on the GHT, GHT hemifield clusters, and the pattern deviation probability plot provide high sensitivity and specificity for detecting early glaucomatous visual field changes. (C) 2002 by Elsevier Science Inc. All rights reserved.