Optimizing distribution and number of test locations in perimetry

Abstract 

• Background: Short programs with few test locations have been proposed for glaucoma screening. Program G1x of the Octopus 1-2-3 automated perimeter is divided into four stages and permits the examination to be stopped after 16, 32, 45, or all 59 test locations. We investigated whether such short programs provide information comparable with that supplied by standard programs in glaucoma. • Materials and methods: In 99 visual fields of 81 glaucomatous and 18 glaucoma-suspect right eyes, mean defect (MD) and loss variance (LV) of the entire visual field were compared with MD and LV of the 4 stages of Program G1x. • Results: MD of the entire visual field averaged 0.40 dB, with averages of –0.17 dB, 0.34 dB, 0.47 dB, and 1.04 dB for stages 1, 2, 3, and 4, respectively. LV of the entire visual field averaged 13.0 dB² and was similar for all four stages. • Conclusions: The results show that the 16 test locations of stage 1 of Program G1x underestimate the visual-field damage present in the entire field. We recommend examination of at least 32 test locations, i.e., two stages of Program G1x. However, a prospective study is required to evaluate the sensitivity and specificity of short programs for glaucoma screening. The selection of test locations for the stages might be improved.