Purchase this article with an account.
William H. Swanson, Victor E. Malinovsky, Dawn M. Meyer, Julie K. Torbit, Bradley M. Sutton; Optimizing Contrast Sensitivity Perimetry (CSP) For Assessing Glaucomatous Damage. Invest. Ophthalmol. Vis. Sci. 2011;52(14):5500.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Hot, Dul & Swanson (2008 IOVS 49:3049-3057) demonstrated that contrast sensitivity perimetry (CSP) can reduce test-retest variability in glaucomatous defects. However, their 6-reversal staircase algorithm allowed testing of only 26 locations in 8 minutes, and their stimuli were too large for testing macular locations. We developed smaller CSP stimuli appropriate for testing macular locations, and alternate algorithms to increase the number of locations without increasing test duration.
One eye each for 54 patients with glaucoma and 40 age-similar controls was tested with alternate CSP methods on our customized testing station. Stimuli were Gabor patterns in both sine and cosine phase (peak spatial frequencies 0.4 to 1.1 cycle per deg), and two-dimensional Gaussians; standard deviations ranged from 0.25° to 1.7°. We used both staircase and ZEST algorithms. Bland-Altman analysis was used to assess test-retest variability and agreement on defect depth across the different tests. These results were compared to results for the CSP method of Hot et al. ("Hot-CSP").
A two-reversal staircase allowed 46 locations to be tested in an average of 8.2 ± 1.1 minutes. However, test-retest variability had an SD of 0.21 log unit, almost three times greater than the 0.08 log unit SD for Hot-CSP (F = 6.9, p <0.001) The staircase was also used to compare depth of defect for three stimuli: a 1.0 c/deg Gabor, and Gaussians with SDs of 0.25° and 0.5°. For all three comparisons of stimuli the mean difference in defect depth was less than ±0.01 log unit (t = 0.06, p= 0.50) and 95% confidence limits for agreement were ± 0.6 log unit. A ZEST algorithm allowed us to increase the number of locations to 55 with a test duration of 6.3 ± 0.4 minutes, for an 0.5 c/deg Gabor cosine and for Gaussians with SDs of 0.25° and 0.5°. The mean difference in depth of defect across stimuli was ±0.02 log unit and the confidence limits for agreement were ± 0.5 log unit. The standard deviation for repeated testing with the ZEST algorithm was 0.14 log unit, about half that for the two-reversal staircase.
Use of a ZEST algorithm and low-spatial-frequency stimuli allowed us to expand CSP to 55 locations and use stimuli appropriate for macular testing, while retaining the low test-retest variability of Hot-CSP.
This PDF is available to Subscribers Only