Visual field examinations using different strategies in Asian patients taking hydroxychloroquine
We retrospectively reviewed the medical records of a cohort of 1221 consecutive Asian patients with a history of hydroxychloroquine use who underwent ophthalmic examinations for hydroxychloroquine retinopathy at Hanyang University Hospital between January 2016 and December 2020. A total of 301 eyes of 161 patients were excluded due to unavailable or unreliable VF testing. In addition, patients with combined macular diseases (eg, epiretinal membrane, age-related macular degeneration, central serous chorioretinopathy, macular edema associated with diabetic retinopathy or retinal vein occlusion), glaucomatous optic neuropathy, other abnormalities or optic disc neuropathies (eg, optic neuropathy ischemia), or neurological/systemic diseases affecting the VF (eg, pituitary tumor) were excluded. Finally, 2107 eyes of 1078 patients were included in the study (Supplementary Fig. S1). This study adhered to the principles of the Declaration of Helsinki and was approved by the Institutional Review Board (IRB) of Hanyang University Hospital. The IRB waived the need for informed consent due to the retrospective nature of the study.
All included patients underwent comprehensive ophthalmic examinations, including slit lamp examination, measurement of best-corrected visual acuity and refractive error (KW-1500; Kowa, Tokyo, Japan), non-contact tonometry (KT-500 automated tonometer; Kowa), fundus examination by indirect ophthalmoscopy, OCT, FAF and SAP. Swept-source OCT (DRI-Triton; Topcon Inc., Tokyo, Japan) was performed after pupil dilation. A wide-field 3D macular volume scan that generated a data cube through a 9×12 mm2 after acquiring a series of 256 B-scans, each consisting of 512 A-scans, and line scans (either 12 mm radial scans or 5-line vertical and horizontal raster scans) were used for imaging OCT. Blue-light FAF, obtained using an F-10 confocal scanning laser ophthalmoscope (Nidek, Gamagori, Japan) and/or ultra-widefield FAF (Optos PLC, Dunfermline, UK), was also been performed as an additional objective screening test to identify photoreceptor/retinal pigment epithelium (RPE) damage. In patients requiring additional objective evidence for the diagnosis of hydroxychloroquine retinopathy, mfERG (Diagnosys LLC, Lowell, MA, USA) was performed according to the guidelines of the International Society for Clinical Electrophysiology of Vision (ISCEV)ten.
All patients underwent a standardized VF test using the Swedish Interactive Threshold Algorithm 10-2 and/or 30-2 strategy on HFA II or III (Carl Zeiss Meditec, Dublin, CA, USA). Only reliable VFs defined as 11. All VF tests were further examined for the presence of artifacts associated with fatigue, inattention, inappropriate fixation, eyelid or lens edge effect, and any artifacts showing such artifacts was excluded from further analyses. In the following comparative analysis of diagnostic performance between the 10-2 and 30-2 tests for the detection of hydroxychloroquine retinopathy, only patients who underwent both tests within one month were included. Both VF strategies were performed randomly for patients who underwent both tests on the same day.
Diagnosis and classification of hydroxychloroquine retinopathy
For the assessment of the diagnostic capabilities of VF testing, we used the modified diagnostic criterion (gold standard) for the diagnosis of hydroxychloroquine retinopathy which excludes VF from the most recent AAO guidelines.3: at least one objective test anomaly (FAF or mfERG) confirming the OCT anomaly. As OCT is currently the preferred primary test for the diagnosis of hydroxychloroquine retinopathy, the OCT abnormality has been used as the primary requirement for diagnosis; also notably, “OCT abnormality plus additional structural or functional evidence” is consistent with the most recent guidelines of the Royal College of Ophthalmology12. Specifically, all patients diagnosed with hydroxychloroquine retinopathy have characteristic parafoveal/pericentral photoreceptor abnormalities and/or RPE abnormalities on SD-OCT images.3.13. Hyper or hypo-autofluorescence on FAF or decreased amplitude on mfERG in areas corresponding to outer retinal defects on OCT was used as additional objective evidence for the diagnosis of retinopathy in hydroxychloroquine3.14.
The severity and pattern of retinopathy was further determined in eyes with hydroxychloroquine retinopathy. The severity of hydroxychloroquine retinopathy was graded, as in previous reports, as follows: early (localized and patchy photoreceptor defects without involvement of RPE on OCT and/or FAF), moderate (lesion photoreceptors with [> 180°] or complete ring on imaging), or severe (RPE damage combined [thinning/attenuation of RPE line on OCT or hypo-autofluorescence on FAF])3.14. Depending on the type of retinopathy, eyes with hydroxychloroquine retinopathy were classified as parafoveal (disturbance of photoreceptors/RPE within 2-6° of the fovea), pericentral (outer retinal damage >6° from the fovea ) or mixed (both models)3.
Evaluation of VF test results
We assessed the characteristics of abnormal VF test results and the diagnostic capabilities of VF using different strategies. For interpretation of VF results, abnormal VF were defined as follows: a total of ≥ 3 points with PP15.16. In patients with abnormal VFs, the pattern of the scotoma was determined based on its shape. Specifically, uneven scotoma (consecutive scotoma points with extent less than 90°; Fig. 1A), partial annular scotoma (arc-shaped scotoma with extent between one and three quadrants; Fig. 1B), scotoma with complete ring (ring-shaped scotoma involving all quadrants; Fig. 1C,D), central scotoma (round scotoma involving all four focal points but not involving the entire field; Fig. 1C) and field defects integer (loss of integer field; Fig. 1D). The location of the scotoma was further analyzed using a point frequency map17which indicates the proportion of eyes with an anomaly (P
Descriptive statistics were obtained for demographic data, details of hydroxychloroquine dosage, clinical characteristics and VF results of the included patients. Continuous values are presented as mean ± standard deviation. Clinical characteristics and details of hydroxychloroquine use were compared using Fisher’s exact test (for dichotomous variables), student you-test (for normally distributed continuous variables) or Mann–Whitney test (for continuous variables violating the normality assumption assessed by the Shapiro–Wilk test). Sensitivity (the number of diseased eyes showing a VF abnormality on the VF test [correctly classified] divided by the number of eyes with retinopathy examined with the same strategy) and the specificity (the number of normal eyes with normal VF results using the protocol divided by all normal eyes receiving the same strategy) of the strategies 10 -2 and 30-2 of the Humphrey VF test were calculated and compared using the McNemar eye test with the results of both tests. Statistical analyzes were performed using SPSS software (version 23.0; IBM Corp., Armonk, NY, USA). Statistical significance was set at P