Automatic Perimetry (COMPETER)

The ability of fully automatic computerized perimetry to detect early glaucomatous field damage was compared with that of careful static and kinetic manual perimetry in a clinical study on 104 patients, 51 of whom had early glaucomatous field defects, 20 of whom were glaucoma suspects with no field defects, and 33 of whom were normal. The automatic perimeter used was the COMPETER automatic perimeter employing actual threshold measurements. The interpretation of the automatic fields followed a set of predetermined criteria. Fifty-one eyes had defects in the manual charts, 48 (94%) of which were detected by automatic perimetry using the central test point pattern of the perimeter. Two (4%) fields thought to be normal after manual perimetry were correctly found to be abnormal by the automatic perimeter, which yielded four (8%) false-positives in the normal fields. By using a different set of criteria for the interpretation of the automatic fields, the sensitivity could be increased to 98% of these early defects, but at the cost of 22% false-positives.