RESUMO
PURPOSE: To compare the effect of uncorrected refractive error on threshold estimation using frequency doubling perimetry (FDP) full-threshold N-30 with emmetropia among normal subjects. METHODS: One thousand two hundred ninety-nine subjects were enrolled from the Chennai Glaucoma Study, a population-based glaucoma prevalence study. Subjects underwent a comprehensive eye examination including the FDP full-threshold N-30 test. Normal subjects (with spherical equivalent refractive errors within +/-7 D) with no other ocular pathology were stratified into six groups based on the degree of myopia and hyperopia and compared with age-matched emmetropic controls. A subset of 22 subjects with higher refractive errors was assessed for within-subject effect on FDP parameters. The following FDP parameters were compared: mean deviation, pattern standard deviation, central threshold (CT), mean sensitivity (MS), paracentral points (PA), and peripheral threshold. RESULTS: The one-way analysis of variance between all refractive error groups and emmetropes showed no statistically significant difference for the mean deviation (p = 0.1002) and pattern standard deviation (p = 0.4789). FDP parameters did not show a statistically significant difference for between and within-group comparisons. The variability of FDP sensitivity (derived from the 95% confidence interval range) as a proportion of the threshold range of the instrument was 31, 41, 46, and 41% for CT, MS, PA, and peripheral sensitivity, respectively, without spectacle correction and 29, 34, 36, and 35% for CT, MS, PA, and peripheral sensitivity, respectively, with correction. The mean CT, PA, and peripheral sensitivity show a decreasing trend from central to periphery in all the refractive error groups and a similar trend was noted in the emmetropic controls. CONCLUSION: Between-subject (uncorrected ammetropes and age-matched emmetropes) and within-subject comparisons showed no statistically significant differences in any of the FDP parameters or in the contrast sensitivity estimates between the central and peripheral test locations. This finding is likely due to the high within-subject variability of FDP.
