Changes in visual function associated with hydroxychloroquine retinal toxicity

Abstract

AbstractIntroduction: Non‐infectious inflammatory diseases, such as systemic lupus erythematosus, rheumatoid arthritis, dermatomyositis, antiphospholipid antibody syndrome and others cause significant morbidity and are treated with immunosuppression. Hydroxychloroquine (HCQ) sulfate is used as a second line immunosuppressant and has many advantages over other agents. These include increased life expectancy, reduced risk of neoplasia, reduced risk of diabetes and hypercholesterolaemia. HCQ is safe in pregnancy. After more than 5 years of treatment, HCQ can cause a degenerative maculopathy characterized by loss of parafoveal photoreceptors and RPE cells, leading to a ring‐like visual field defect. Current detection of early retinal toxicity relies on central visual field testing, optical coherence tomography (OCT) scan and fundus autofluorescence. The reason for the parafoveal predilection for retinopathy is not understood. HCQ is concentrated in melanin‐containing cells, and there is increased melanin concentration in the macular RPE cells. A hypothesis is that HCQ concentrated in the macular RPE cells leads to toxicity, with macular pigment protecting the fovea but leaving a ring of potentially vulnerable tissue in the parafoveal region. We were interested to investigate parafoveal function in patients with mild HCQ maculopathy and report the results of detailed macular function assessment of red/green (RG) and yellow / blue (YB) colour vision and rod‐ and cone‐mediated rapid flicker sensitivity.Methods: Two patients with confirmed early hydroxychloroquine maculopathy participated in the study. RG and YB colour thresholds were measured with the Colour Assessment and Diagnosis (CAD) test. Cone and rod‐mediated rapid flicker sensitivity was also measured in both central and paracentral vision using the Flicker‐Plus test. Both patients had taken HCQ for more than 15 years and had paracentral scotomas on 10–2 field test and loss of the parafoveal outer nuclear layer and ellipsoid layer, confirming the presence of maculopathy. To test the hypothesis of parafoveal toxicity risk, 20 short wavelength (BAF) and infrared wavelength (IRAF) autofluorescence images were reviewed. The BAF images are surrogate markers of the macular pigment and the IRAF the melanin concentration. Normalized maps of the difference between BAF and IRAF were calculated to investigate the possibility of a ring of vulnerable tissue.Results: Both patients had normal RG, but abnormal YB colour thresholds. One subject had normal cone flicker thresholds, but the rod flicker thresholds were raised in all quadrants in both eyes. The second subject had significantly elevated rod thresholds and elevated cone flicker thresholds in both eyes. Of note is that the degree of retinopathy in the second patient was greater than in the first. The areas of greatest threshold elevation correlated with the areas of greatest field loss and anatomical disruption on OCT.Conclusion: These new findings show that hydroxychloroquine retinopathy affects preferentially both YB colour vision and rod‐mediated flicker sensitivity. The cone‐mediated flicker sensitivity is also affected in more advanced disease. A model for the anatomical location of the toxicity will be presented, based on the hypothesis outlined above.