PURPOSE: To design a contrast sensitivity perimetry test with the potential to improve clinical management of glaucoma by decreasing test-retest variability in defective areas while maintaining good sensitivity to glaucomatous loss.
METHODS: Twenty patients with glaucoma, ten age-similar control subjects and ten young control subjects were recruited. One eye was tested per subject. All subjects were tested with contrast sensitivity perimetry (CSP), and conventional automated perimetry (CAP). Stimuli for CSP were Gabor patches with a peak spatial frequency of 0.375 cpd and a two-dimensional spatial Gaussian envelope with the most of the energy concentrated within a central circular region 4 degrees in diameter.
For CSP, stimuli were presented at 26 locations over the central visual field, excluding the central 5 degrees and with the emphasis on the nasal hemifield. Neuroretinal rim area of the patients was measured using retinal tomography (HRT II). Young subjects were tested on CSP using 4-reversal and 8-reversal staircases to estimate variability and test duration. Bland-Altman analysis of agreement was used to assess test-retest variability, to compare depth of defect for the two perimetric tests, and to investigate the relation between contrast sensitivity and neuroretinal rim area.
RESULTS: With 4-reversal staircases, for both Size III stimuli and Gabor stimuli, variability increased as sensitivity decreased (r2 > 5%, p<0.001), but at a shallower rate for the Gabor stimuli (Z > 2.9, p<0.005). With 6- and 8- reversal staircases, the correlation between sensitivity and variability was not significant (r2 = 0.05%, p>0.5) and the slope of the regression line was shallower than for the 4-reversal staircases (Z>2.4, p< 0.01). Depth of defect was on average similar for the two devices, but for some patients the size III stimuli tended to yield deeper defects than Gabors in regions of lower sensitivity. The relation between rim area and perimetric sensitivity was more consistent for contrast sensitivity perimetry than for conventional perimetry (Z=9.3, p < 0.0005). For control subjects there was a significantly shallower decline in sensitivity with eccentricity using Gabor stimuli than with conventional size III stimuli (Z=3.78, p<0.0005), and overall test-retest variability was similar for staircases with 4 to 8 reversals.
CONCLUSION: Contrast sensitivity perimetry demonstrated reasonably low test-retest variability and good ability to detect defect. It also showed a structure-function relation that was independent of the degree of glaucomatous loss. Further research on contrast sensitivity perimetry is needed to evaluate its utility in monitoring progression of glaucoma and effects of treatment.