Purpose
While some studies find a clear preponderance of glaucomatous defects in the superior visual field (VF), others find little or no difference between hemifields. The purpose of this study was to test the hypothesis that the ratio of superior to inferior VF defects depends upon the area of the VF tested and the definition of an abnormal hemifield.
Methods
In a prospective study, reliable 24-2 and 10-2 VFs (HFA II, CZM, Inc.), as well as frequency domain optical coherence tomography (fdOCT) disc and macular cube scans (3D-OCT 2000, Topcon, Inc.), were obtained from 109 eyes of 109 glaucoma patients and suspects. All eyes had a mean deviation (MD) better than -6 dB on 24-2 VF and glaucomatous optic neuropathy on fundus exam. Two criteria were used to classify a hemifield as abnormal: 1. cluster criteria (CC) and 2. deep defect (DD). The CC was 3 contiguous points at 5,5,1% or 5,2,2% or worse on either total or pattern deviation plots. For the DD, at least one point had to be more than 8 dB below machine norms. To essentially eliminate false positives, a second analysis required a hemifield to be abnormal on both the VF and on probability maps generated from the OCT scans.
Results
For the CC, the ratio of superior to inferior VF defects was nearly the same for the 24-2 (1.16) and 10-2 (1.11) VFs (Table 1). However, for the DD, the ratio for the 24-2 (1.74, p=0.07, 2-tail sign test), and especially the ratio for the 10-2 (4.2, p<0.003), were greater than 1.0 (Table 1). While there were fewer abnormal hemifields when both VF and OCT were required to be abnormal, the general pattern of results was the same (Table 2).
Conclusions
In patients with early damage, when an abnormal hemifield is defined based upon a cluster criterion, there are only about 10 to 30% more abnormal superior VFs as compared to inferior VFs. However, deep defects are more common in the superior VF, especially on 10-2 tests, where they are about 4 times more likely. By adding an assumption that diffuse damage is common and equally likely to occur in both hemifields, a recent model [1] can explain the results. 1. Hood, Raza et al, PRER, 2013.