Influence of fixation accuracy on threshold variability in patients with open angle glaucoma

To evaluate the contribution that fixation errors make to the overall variability of perimetric responses in patients with glaucoma.

Frequency of seeing curves were established, with and without fixation error correction, at two locations in each of 14 patients with glaucoma and good visual acuity. One location corresponded to a relatively normal region of the visual field, whereas the second corresponded to a region in which there was a sensitivity deficit. All patients had an acuity of better than 20/63 (except one whose acuity was 20/100). The locations of the first and fourth Purkinje images of a collimated infrared source were used to give a measure of eye position, during each stimulus presentation (accuracy 10 minutes of arc).

Considerable variation was found in patient fixation accuracy. In the worst case, fixation was within 30 minutes of the target in only 7% of presentations whereas in the best, it was within this range in more than 60%. No relationship was found between accuracy of fixation and extent of loss. The gradient of the frequency of seeing curve was found to be shallow at regions of reduced sensitivity, a finding that supports the recognized relationship between variability and sensitivity deficit. A recalculation of the frequency of seeing curves, using only those responses in which the patient's fixation was within a specified range ( < 60 minutes of arc), did not show a meaningful reduction in variability at either location.

It is concluded that fixation errors, though contributing to variability, are not the major cause of the increased variability seen at locations with reduced sensitivity.