Functional and morphologic benefits in early detection of neovascular age-related macular degeneration using the preferential hyperacuity perimeter.

PURPOSE To estimate the usefulness of preferential hyperacuity perimetry (PHP) in detecting conversion of early to late age-related macular degeneration in the Carotenoids and co-antioxidants in patients with Age-Related Maculopathy, a multicenter randomized controlled clinical trial. METHODS This was a nested case control study within the Carotenoids and co-antioxidants in patients with Age-Related Maculopathy (CARMA) clinical trial and included all participants enrolled in a single center (n = 200). Data are from participants who progressed to neovascular age-related macular degeneration (nvAMD) during time on study, Group 1 (n = 10) before the use of PHP and Group 2 (n = 10) during use of PHP. We also randomly selected 21 other participants (Group 3) who did not progress to nvAMD during time on study as a control group. Change in best-corrected visual acuity and contrast sensitivity and size of neovascular lesion at detection of conversion to nvAMD in Groups 1 and 2. RESULTS At detection of nvAMD, mean best-corrected visual acuity in Group 1 was 57.5 letters versus 67.4 in Group 2. In Group 1, the change in best-corrected visual acuity from baseline to detection of nvAMD was twice that of Group 2 (21.6 ± 9.0 versus 11.9 ± 10.7) with a mean difference of 9.7 letters (95% confidence interval, 0.41 to 19.0, P = 0.04, independent-samples t-test). The size of the neovascular lesion at detection was 3.06 mm in Group 1 versus 0.89 mm in Group 2 (P = 0.02). Two thirds of the participants in Group 2 were asymptomatic at detection of nvAMD compared with one fifth in Group 1. Preferential hyperacuity perimetry distortion maps were abnormal in 9 of 10 eyes in Group 2, which were confirmed by optical coherence tomography. Of the 21 eyes in Group 3, PHP maps were normal in 18 and abnormal in 3. CONCLUSION Preferential hyperacuity perimetry detected abnormalities in central visual function with high reliability. Eyes with nvAMD lesions detected by PHP had smaller lesions and better function when compared with the group before the introduction of PHP. The false-negative rate was <10% on PHP. The PHP distortion map was helpful in alerting clinicians to the presence of subclinical nvAMD.