LAUSR

Editor, W e read with interest the article in issue 93, November 2015, by Debellemani ere et al. (2015), who described the use of an adaptive optics (AO) camera (RTX1, Imagine Eyes, Orsay, France) and observed cone loss with increasing hydroxychloroquine (HCQ) cumulative doses in patients with no clinical evidence of maculopathy. Adaptive optics (AO) imaging can be used to compensate for optic aberrations on imaging the eye and allows for direct visualization of the photoreceptor mosaic. Previous study of patients with confirmed HCQ-induced maculopathy revealed a correlation between AO features and spectral domain-optical coherence tomography (SD-OCT) findings (Jacob et al. 2015); also, irregularities in the cone photoreceptor density and mosaic were seen in patients with normal findings on SDOCT alone (Stepien et al. 2009). In experimental studies of HCQ toxicity, photoreceptor alterations are observed before choroid and pigment epithelium deterioration (Rosenthal et al. 1978). Between November 2014 and November 2015, we conducted an observational prospective study in the ophthalmology department of Montpellier University Hospital, France, to evaluate the association of risk factors for HCQ toxicity (ideal body weight daily dose, cumulative dose, treatment duration) and cone density. Following the American Academy of Ophthalmology revised guidelines for HCQ toxicity screening, we measured bestcorrected visual acuity and used fundoscopy combined with Humphrey white visual field 10-2 perimetry testing and SD-OCT (Marmor et al. 2011). Cone metrics were obtained by the AO camera in the Debellemani ere study (RTX1, Imagine Eyes). The region of interest corresponded to a 100 9 100 lm square placed superiorly and inferiorly at 2° of eccentricity from the fovea, where retinal HCQ toxic effects would appear first before appearing in other retinal areas (Marmor et al. 2011). We included 38 participants [35 females; mean (SD) age 45.74 (12.73) years] undergoing screening for HCQinduced maculopathy; 23 had systemic lupus erythematosus, six GougerotSjogren disease, three mixed connective tissue disease (SHARP syndrome), one polyarteritis nodosa, two rheumatoid polyarthritis, one limited scleroderma (CREST syndrome) and one sarcoidosis. The pathology for one patient was missing. For all participants, bestcorrected visual acuity evaluated by the Monoyer and Parinaud scale was 20/20. The Humphrey 10-2 visual field and macular thickness OCT were all within normal values. Mean (SD) foveal threshold was 34.8 (5.2) dB. None showed retinal HCQ toxic effects on screening tests. The median (interquartile range) treatment duration was 10.02 (4.0–17.0) years. Age was negatively correlated with cone density at 2° from the fovea in the superior and inferior quadrants (r = 0.47, p < 0.03 and r = 0.37, p = 0.04). Ageand sex-adjusted mean (SE) cone density in the inferior quadrant was significantly lower with ideal body weight daily dose ≥6.5 than those who took <6.5 mg/kg/day [22 339.23 (704.01) versus 20 480.93 (655.38), p = 0.035); results were not significant in the superior quadrant (p = 0.475) (Fig. 1A). Mean (SD) cone density in the inferior quadrant was lower with a cumulative dose ≥1000 than those who took <1000 g [20593.85 (603.41) versus 22365.04 (777.92), p < 0.008]; results were not significant in the superior quadrant (p = 0.739) (Fig. 1B). Mean cone density did not differ by treatment duration ≥10 versus <10 years in the superior or inferior quadrant (p = 0.269 and p = 0.405, respectively). Overall, cone density did not differ between the superior and inferior quadrant (p = 0.545). Risk of HCQ toxic effects was previously found increased with ideal body weight daily dose and cumulative dose >6.5 mg/kg/day and 1000 g, respectively (Ingster-Moati et al. 2004; Wolfe & Marmor 2010). In our study, the inferior cone density was lower with these risk factors. In patients