Relative sensitivity and specificity of 10-2 visual fields, multifocal electroretinography, and spectral domain optical coherence tomography in detecting hydroxychloroquine and chloroquine retinopathy

Abstract

Background: The purpose of this study was to determine the relative sensitivity and specificity of 10-2 visual fields (10-2 VFs), multifocal electroretinography (mfERG), and spectral domain optical coherence tomography (SD-OCT) in detecting hydroxychloroquine retinopathy.

Methods: A total of 121 patients taking hydroxychloroquine (n=119) or chloroquine (n=2) with 10-2 VF, mfERG, and SD-OCT tests were retrospectively reviewed. Rates of test abnormality were determined.

Results: Retinopathy was present in 14 and absent in 107. Eleven of 14 (78.6%) patients with retinopathy were overdosed. Twelve (85.7%) had cumulative dosing greater than 1,000 g. The sensitivities of 10-2 VF, mfERG, and SD-OCT in detecting retinopathy were 85.7%, 92.9%, and 78.6%, respectively. The specificities of 10-2 VF, mfERG, and SD-OCT in detecting retinopathy were 92.5%, 86.9%, and 98.1%, respectively. Positive predictive values of 10-2 VF, mfERG, and SD-OCT in detecting retinopathy were less than 30% for all estimates of hydroxychloroquine retinopathy prevalence. Negative predictive values were >99% for all tests.

Conclusion: Based on published estimates of hydroxychloroquine retinopathy prevalence, all three tests are most reliable when negative, allowing confident exclusion of retinopathy in patients taking ≤6.5 mg/kg/day. Each test is less useful in allowing a confident diagnosis of retinopathy when positive, especially in patients taking ≤6.5 mg/kg/day.

Keywords: chloroquine; hydroxychloroquine; ideal body weight; multifocal electroretinography; retinopathy; spectral domain optical coherence tomography; toxicity.

Figure 1

Hydroxychloroquine toxicity detected by three ancillary tests. Notes: This 72-year-old woman had taken hydroxychloroquine 400 mg/day for rheumatoid arthritis for over 25 years. She was 5 feet 4 inches tall and weighed 120 pounds (less than her ideal body weight of 140 pounds). Her adjusted daily dose was 7.33 mg/kg/day. Her cumulative dose was 3,650 g. All images are of the right eye (left eye similar). (A) The multifocal electroretinogram was abnormal with flattened waveforms throughout, but especially centrally (red circled area). The averaged N₁-P₁ waveforms are small in amplitude, which has led to erroneous automatic cursor placement (black arrows) and an artifactual R₁/R₂ ratio of 3.1. The N₁-P₁ amplitudes for rings R₂ and R₃ are below the lower limit of normal. (B) The color fundus photograph of the right eye shows a bull’s eye lesion in the macula. The mid-phase fluorescein angiogram shows window hyperfluorescence corresponding to the bull’s eye lesion. (C) The spectral domain optical coherence tomography line scan shows loss of the ellipsoid zone line (ends at the dashed white arrow) and the retinal pigment epithelial line (ends at the dotted white arrow). An unrelated epiretinal membrane is present (solid white arrow). (D) 10-2 visual field shows a central scotoma that has a slightly less impaired foveal threshold. Abbreviations: 3D, three dimensional; RMS, root mean square; PSD, pattern standard deviation; MD, mean deviation; FN, false negatives; FP, false positives; T, temporal; N, nasal; RE, right eye.

Figure 2

Hydroxychloroquine retinopathy detected by multifocal electroretinography, but not by 10-2 visual field testing or spectral domain optical coherence tomography. Notes: This 71-year-old woman with dermatologic lupus and arthritis had taken 400 mg/kg/day of hydroxychloroquine for 37 years. She was 5 feet tall with an ideal body weight of 123 pounds. Her adjusted daily dose was 7.15 mg/kg/day. Her cumulative dose was 2,920 g. Funduscopy was normal. All images are of the left eye (right eye similar). (A) The multifocal electroretinogram showed N₁-P₁ amplitudes that were below the lower limit of normal for the central hexagon (R₁) and rings R₂ and R₃ (dashed circled area). The waveforms in the individual hexagons are flat (black, solid circled area). The averaged waveforms are so small that the machine-placed cursors are erroneously positioned (black arrows). (B) The 10-2 visual field from December 4, 2009 and December 16, 2011 are normal. There are some elevated thresholds at isolated points in the field, but they are not reproducible. (C) Spectral domain optical coherence tomography shows an intact ellipsoid zone line throughout the scan. Abbreviations: 3D, three dimensional; RMS, root mean square; PSD, pattern standard deviation; MD, mean deviation; FN, false negatives; FP, false positives; T, temporal; N, nasal; LE, left eye.

Figure 3

Hydroxychloroquine toxicity detected by 10-2 visual field and spectral domain optical coherence tomography testing, but not by multifocal electroretinography. Notes: This 67-year-old woman took hydroxychloroquine 400 mg/day for systemic lupus erythematosus for 9 years. She was 5 feet 2 inches tall with ideal body weight of 131 pounds. Her adjusted daily dose was 6.72 mg/kg/day. Her cumulative dose was 1,223 g. Funduscopy was normal. All images are of the left eye (right eye similar). (A) The multifocal electroretinography of the left eye was normal. The waveforms in the individual hexagons, the averaged waveforms for the rings, the averaged N₁-P₁ amplitudes for the central hexagon (R₁), and the rings (R₂–R₅) were all normal. The R₁/R₂ ratio was normal. (B) The 10-2 visual fields show a serial increase in the threshold for paracentral locations, especially superiorly (compare the solid black arrow to the dashed black arrow). (C) The spectral domain optical coherence tomography shows paracentral focal loss of the ellipsoid zone line (white arrows). Abbreviations: 3D, three dimensional; RMS, root mean square; FP, false positives; T, temporal; N, nasal; LE, left eye.