Abstract
Purpose:
To study the effect of perimetric eccentricity on the test-re-test characteristics in participants with early to moderate glaucoma.
Methods:
The sample consisted of 161 participants with early (n=135) to moderate (n=26) glaucoma. Each participant was experienced at perimetry and underwent 2 standard automated perimetry tests (SAP) within 8 weeks, using the Heidelberg Edge Perimeter (HEP; ASTAStd, size III, 24-2). The visual field was divided into 3 zones (central 10°, 10-16°, and 16-25°) with each zone additionally divided into upper and lower hemifields (6 zones in total). All blindspot locations were removed and 0dBs were additionally removed in consideration of the floor effect. Threshold values were analyzed as a function of concentric zone, hemifield and visit. Test/ re-test characteristics (taken as visit 1 - visit 2) for each zone and hemifield were analyzed using scatter plots and Bland and Altman (B-A) plots. Effects of visual field location, hemifield and visit were analyzed with repeated measures ANOVA (p<0.05).
Results:
There was a significant effect by zone (p<0.001), sensitivity reducing with eccentricity, but not between visits (p=0.434). There was also a significant difference between hemifields, with the lower hemifields showing higher sensitivity (p<0.001), but no interaction between hemifield and visit (p=0.639). The mean of differences (MoD) between visits ranged between -0.02 and -0.15dB, showing a small, but non-significant learning effect. The B-A limits of agreement (1.96SD of the MoD) were greatest in the most peripheral zone (C10°: ±5.15; 10-16°: ±4.68; 16-25°: ±5.80), although there was no significant interaction for Zone x Visit. The upper hemifield showed the largest limits of agreement (e.g. U16-25°: ±6.02 vs. L16-25°: ±5.45). There was a significant interaction for Zone x Visit x Hemifield (p=0.002). Additionally, when looking at defect depth per zone, there were more abnormal locations at all defect depths in the 16-25° zone. However, when corrected for the number of locations per zone there was surprisingly an equal chance of detecting defect within each zone.
Conclusions:
As expected, sensitivity reduced with eccentricity and the upper hemifield gave the lowest sensitivity and worst repeatability. The likelihood of finding defect in early glaucoma was equal across zones when corrected for the number of locations tested.
Keywords: 758 visual fields •
642 perimetry •
638 pathology: human