Abstract
Purpose: :
To determine if global and local rates of visual field (VF) change are equivalent in Standard Automated Perimetry (SAP) and Frequency Doubling Perimetry (Matrix).
Methods: :
Longitudinal VF data from patients with high-risk ocular hypertension and early glaucoma in the Portland Progression Project were searched. Eyes with > 5 SAP and Matrix VFs within 60 days of each other were selected. For 98% of examinations, SAP and Matrix were performed on the same day. The most recent 6 VFs of each type were then selected for analysis. We assessed the global rate of VF change using linear regression of mean deviation over time and extracting its rate of change (MDR). Since Matrix uses 20 dB/log unit of contrast change and SAP uses 10 dB/log we expect a 2:1 ratio between Matrix MDR and SAP MDR if the tests were changing at the same rate. We quantified the local rate of VF change by performing pointwise linear regression (PLR) and then applied 4 change criteria to the PLR slopes; 1) -1dB/Yr for SAP and -2dB/Yr for Matrix at p<0.01, 2) -1dB/Yr and -2dB/Yr at p<0.02, 3) -0.5dB/Yr and -1dB/Yr at p<0.01 and 4) -0.5dB/Yr and -1dB/Yr at p<0.02.
Results: :
Data from 229 eyes (115 participants) were available after applying selection criteria. Mean age at the 1st visit was 61.6+/-9.7 years (mean+/-SD) and visual acuity averaged 0.0 logMAR (20/20). At the 1st visit within a sequence the mean SAP MD and PSD were -2.8+/-2.6 dB and 2.3+/-2.0 dB respectively. The mean SAP and Matrix MDRs were -0.20+/-0.31 and -0.10+/-0.45 dB/Yr. Given the 2:1 ratio in the definitions used for 1 dB, this suggests the global rate of SAP change was approximately four times more rapid (p<0.01, generalized estimating equation) than the global rate of Matrix change. In the PLR analyses, the mean number of significantly progressing locations within the VF for SAP and Matrix respectively for the 4 criteria described above were 1) 0.31 & 0.32, 2) 0.51 & 0.51, 3) 0.79 & 0.63 and 4) 1.42 & 1.15. There was a trend for more PLR progression in SAP than Matrix when the more liberal PLR criteria were applied but the difference between tests was not significant (p>0.09 for all 4 comparisons).
Conclusions: :
The global rate of VF change in patients with high-risk ocular hypertension and early glaucoma, as estimated by regressing MD over time, was significantly more rapid for SAP than Matrix. Local VF change, estimated via PLR, was similar for the two tests. Less experience with Matrix combined with a prolonged learning effect might account for this seeming inconsistency.
Keywords: visual fields • perimetry