No abstract

Of 115 patients with chronic open angle glaucoma or suspected glaucoma 115 eyes were subjected to a visual field examination with the manual Goldmann perimeter and the automatic Octopus perimeter. In 84.4 +/- 8.9% the automaton detected more visual field loss, including 27.8 +/- 10.9% of the eyes where no visual field loss was found by the manual perimetry. Since some eyes with normal fields on manual perimetry and abnormal fields detected by automatic perimetry remained abnormal on retesting, it is assumed that scotomas found by automatic perimetry and not shown by manual perimetry constitute false-negative manual fields rather than false-positive automatic fields.

The analytical programme Delta was used to determine longterm fluctuation and accuracy of measurement of the programme 31 of Octopus when used on glaucoma patients. Programme 31 examines the 30 degrees field. The test locations are arranged in a square grid with 6 degrees resolution. The programme Delta determines and compares 1) the disturbed area in %; 2) the total loss, the total sensitivity being around 2000 dB; 3) the loss in dB per mean number of disturbed points. Thirty-two eyes of 22 patients with established glaucomatous field defects were examined twice within two to six days and two months later again twice. The size of the disturbed area served for classification of our sample into three groups: 1st group: disturbed area 1-33%; 2nd group: disturbed area 34-66%; 3rd group: disturbed area 67-100%. Long-term fluctuations and accuracy of measurement could be determined as respectively follows: 1) Disturbed area between 0.7 +/- 8% in group 3 and 1.7 +/- 13% in group 2. 2) The total loss increases proportionately to the disturbed area and was 4.9 +/- 29.2 dB in group 1 and 31.8 +/- 82.4 dB in group 3. 3) The total loss per mean number of disturbed points was 0.5 +/- 2 dB in group 1 and 0.3 +/- 1.2 dB in group 2. This signifies that if the learning effect is over, changes of more than 2 dB, especially if several adjacent points are affected, are a significant loss. The learning effect, as determined in an earlier study, may go up as high as 2 dB per point.

No abstract

No abstract