Visual field examinations using different strategies in Asian patients taking hydroxychloroquine

Abstract

In this study, we investigated the patterns of visual field (VF) defects and the diagnostic abilities of VF tests using different strategies in Asian patients with hydroxychloroquine retinopathy. Patients screened for hydroxychloroquine retinopathy using optical coherence tomography, fundus autofluorescence, VF, and/or multifocal electroretinography were included. The VF was performed using the Humphrey 30-2 and/or 10-2 strategy, and 2,107 eyes of 1,078 patients with reliable results, including 136 eyes of 68 patients with hydroxychloroquine retinopathy, were analyzed. The characteristics of VF findings were evaluated and the sensitivity and specificity were compared between the 30-2 and 10-2 tests in subgroups of retinopathy severity and pattern. The most common VF defect pattern was partial- or full-ring scotoma in both the 10-2 and 30-2 tests. Among the eyes with hydroxychloroquine retinopathy that underwent both tests, 14.2% showed a disparity between the two tests, almost all at the early stage. In overall and early pericentral retinopathy, the sensitivity of the 30-2 test was significantly higher than that of the 10-2 test (95.7% vs. 77.1% and 90.6% vs. 53.1%, respectively; P < 0.05). However, the specificity of the 10-2 test was significantly higher than that of the 30-2 test (89.6% vs. 84.8%, P < 0.001). Therefore, the pattern of retinopathy should be carefully considered when choosing a VF strategy for better detection of hydroxychloroquine retinopathy.

Figure 1

Representative examples of scotoma patterns noted in 10-2 and 30-2 Humphrey visual field (VF) tests noted in 4 patients with hydroxychloroquine retinopathy: patchy scotoma, partial- or full-ring scotoma, ring scotoma, central scotoma, and whole-field defect. In each case, results from fundus autofluorescence (FAF, top left), optical coherence tomography (OCT, bottom left), grayscale map (top right, with text of VF protocol), and pattern deviation map (bottom right) are presented. All patients showed abnormalities, including photoreceptor loss on OCT, and hyper- or hypo-autofluorescence on FAF in the parafoveal or pericentral areas. Yellow arrowheads indicate the areas of retinal damage. N = nasal; T = temporal; S = superior; I = inferior.

Figure 2

Pointwise frequency map of scotoma for pattern deviation maps of 10-2 and 30-2 tests in early and overall eyes with hydroxychloroquine retinopathy. Probabilities (the proportion of eyes showing abnormality at each test location in the 10-2 and 30-2 tests among those tested with each of the VF strategies) are indicated as numbers within each test point and also represented by using grayscale in each group of parafoveal and pericentral retinopathy among overall and early cases. The numbers in the lower right corner indicate the sample sizes (n) for the frequency maps.

Figure 3

Examples of visual field (VF) results showing disparity in abnormality between 10-2 and 30-2 tests in eyes with (A) parafoveal and (B) pericentral retinopathy, and (C) Venn diagram showing disparity between the tests. (A) In the eye with parafoveal retinopathy, the 30-2 test shows an ambiguous finding, whereas the 10-2 test shows a temporal patchy scotoma. (B) In the eye with pericentral retinopathy, the 30-2 test shows a partial-ring scotoma, while the 10-2 test plots are essentially normal. The yellow arrowheads indicate abnormal findings on fundus autofluorescence (FAF, top left) and optical coherence tomography (OCT, bottom left) imaging. (C) Percentage of eyes classified as having an abnormal VF with the 30-2 (white) and 10-2 (black) tests for each retinopathy pattern among overall and early cases. Gray indicates those with the same results (overlap) between the 10-2 and 30-2 tests.