Vincenzoparisi.it

Graefes Arch Clin Exp Ophthalmol (2011) 249:715–721DOI 10.1007/s00417-010-1612-6 Retinal functional changes measured by frequency-doublingtechnology in patients treated with hydroxychloroquine Lucia Tanga & Marco Centofanti & Francesco Oddone & Mariacristina Parravano &Vincenzo Parisi & Lucia Ziccardi & Barbara Kroegler & Roberto Perricone &Gianluca Manni Received: 1 September 2010 / Revised: 20 December 2010 / Accepted: 29 December 2010 / Published online: 21 January 2011 frequency-doubling technology (FDT) in patients treated Background Antimalarial drugs such as chloroquine (CQ) and hydroxychloroquine (HCQ) are mainly used in the Methods Forty-eight eyes of 48 subjects treated with treatment of rheumatologic diseases, and their use may be hydroxychloroquine (HCQ), with no signs of retinal associated with irreversible retinal toxicity. Previous studies toxicity, and 36 eyes of 36 age and sex-matched healthy indicate early paracentral visual field loss (Humphrey 10-2) subjects were enrolled in this cross-sectional, prospective, in patients taking HCQ". These paracentral defects appear observational, case control study. Functional testing includ- before changes in other clinical parameters as visual acuity ed frequency-doubling Humphrey-matrix perimetry (FDP), and fundoscopy. The mechanism of CQ toxicity remains white-on-white Humphrey visual field perimetry (HFA), unclear. It was reported that toxic doses of CQ administered using the 24-2 and 10-2 threshold programs, multifocal for as long as 4.5 years to Rhesus monkeys caused an initial electroretinogram (mfERG, Veris 4.9) and low contrast dramatic effect on ganglion cells, followed later by photo- receptors and RPE degeneration. The purpose of this study Results FDP mean deviation (MD) was found to be is to explore early retinal functional changes measured by significantly reduced in HCQ-treated patients compared tocontrols both in the 24-2 (−1.38±2.41 dB vs 0.21±1.83dB,p<0.01) and in the 10-2 program (−0.97 ±2.88 dB vs 0.15± The authors have no financial relationships.
1.72dB, p<0.01). FDP pattern standard deviation (PSD) The authors have the full controls off all primary data and are agree to was found to be significantly worse in HCQ-treated patients allow Graefe’s Archives for clinical and Experimental Ophthamology compared to controls both in the 24-2 (2.70±0.65 dB vs 2.41±0.31 dB, p<0.01 and in the 10-2 program (2.86± L. Tanga (*) M. Centofanti F. Oddone M. Parravano 0.48 dB vs 2.48 ±0.39 dB, p<0.01). HFA PSD and CS was also significantly reduced in HCQ patients, while response Fondazione G.B. Bietti for Study and Research in Ophthalmology, amplitude densities (RAD) were similar between patients and controls. A statistically significant difference in the ratio of the 5°–10° RAD and the 0°–2.5° RAD (0.31±0.08 vs 0.36±0.07 respectively, p<0.05) was found between M. Centofanti G. ManniUOSD Glaucoma, University of Tor Vergata, Conclusion Frequency doubling perimetry could be useful to detect early retinal impairment in patients treated with B. Kroegler R. PerriconeU.O.C. Reumatologia, Policlinico Tor Vergata, Keywords Frequency doubling perimetry . Retinal Graefes Arch Clin Exp Ophthalmol (2011) 249:715–721 Chloroquine (CQ) and hydroxychloroquine (HCQ) are Patients treated with HCQ for rheumatological diseases currently used in the treatment of malaria and in treatment were included in this cross-sectional case-control study.
of rheumatologic diseases such as rheumatoid arthritis, Healthy subjects were included in the control group. The systemic lupus erythematosus, sarcoidosis, dermatomyositis study was carried out in accordance with the Declaration of Helsinki, was approved by the ethics committee of the Retinal toxicity represents the major and potentially institution and written informed consent was obtained from most serious irreversible side-effect of this type of treatment, occurring in approximately 10–20% of patients The inclusion criteria for patients included: treatment who received CQ and in 3% of patients who received HCQ with HCQ for at least 3 months, age >18 years and Hystological analysis of human and animal retinas with Control subjects needed to meet the following inclusion CQ toxicity has shown multilamellar structures through the criteria: no history of treatment with HCQ, age >18 years retina, followed by loss of retinal ganglion cells (RGCs), photoreceptors and retinal pigment epilthelium (RPE) The exclusion criteria were the same for both groups: atrophy. The accumulation of these multilamellar struc- spherical refractive error >±6 diopters, astigmatism >±3 tures was thought to result either from inhibition of diopters, any active or past retinal pathologies (including lysosomal phospholipases or from inhibition of protein diabetic retinopathy or age-related macular degeneration), diagnosis of glaucoma or ocular hypertension (intraocular Funduscopically, the early stages of retinal toxicity can pressure >22 mmHg), opacities of optic media that could be characterized as a reversible loss of foveal reflex, bias functional and structural retinal testing, history of followed by a development of abnormalities of the retinal pigmented epithelium (RPE) associated with paracentral All subjects underwent a complete ophthalmologic scotomas, and, in the advanced stage, by the classical bull’s examination including: best-corrected visual acuity (BCVA) eye maculopathy, associated with a central scotoma and with Early Treatment Diabetic Retinopathy Study (ETDRS) charts at 4 meters, contrast sensitivity (CS) with Multi- Anatomically, a retinal nerve fiber layer (RNFL) loss has Contrast Sloan Letter Flip Book at 2 meters testing over a been shown in patients under long-term CQ treatment range of contrasts to 100%, 25%, 10%, 5%, 2,5%, 1,25%.
Goldmann applanation tonometry, slit-lamp examination and Nevertheless, the identification of early functional retinal indirect ophthalmoscopy were also performed. Functional impairment of CQ/HCQ retinal toxicity appears to be testing included frequency-doubling Humphrey-matrix peri- difficult before the development of anatomic changes at metry (FDP, Carl Zeiss Meditec), white-on-white Humphrey fundoscopy and symptomatic loss of vision.
visual field perimetry (HFA, Carl Zeiss Meditec, Dublin, Examination tests that are commonly used in evalua- CA, USA), and multifocal electroretinogram (mfERG, Veris tion of CQ/HCQ retinal toxicity include Amsler grid, VA 4.9). The primary objective was to explore by FDP a testing, ophthalmoscopy and visual field testing [, possible retinal sensitivity reduction in HCQ-treated patients. Secondary objectives included the evaluation of Standard achromatic perimetry is considered more additional functional retinal changes at HFA and mfERG, sensitive in the early detection of functional changes than and the influence of the duration of the therapy on the VA, color test and retinal changes at fundoscopy , Functional retinal changes, such as contrast sensitivity ] and multifocal electroretinographic (mfERG) response abnormalities [have been detected at early stages inHCQ-treated patients, even in the preclinical phases.
All enrolled subjects performed two FDP tests within Recently, frequency-doubling perimetry (FDP) has been 2 weeks using the 24-2 and 10-2 threshold program, in introduced as a visual field testing technology that allows order to assess the test–retest variability and to rule out a selective stimulation of the low redundant magnocellular relevant learning curve ]. In addition, subjects per- sub-population of ganglion cells, allowing early detection formed two white-on-white 24-2 and 10-2 SITA standard HFA tests. Reliability criteria were defined as fixation The aim of this study is to explore early retinal errors <15%, false-positives and false-negatives <15%. If functional changes measured by frequency-doubling tech- the second test was unreliable, a third exam was performed nology (FDT) in patients treated with HCQ.
and considered for statistical analysis. In the case of a third Graefes Arch Clin Exp Ophthalmol (2011) 249:715–721 unreliable test, the subject was excluded from the study.
a dose of HCQ >6.5 mg/kg//day with those that received a Mean deviation (MD) and pattern standard deviation (PSD) dose <6.5 mg/kg/day in our sample population.
values were considered for the analysis.
With 84 patients enrolled in the study, the power to Contrast sensitivity was performed with Multi-Contrast detect a 1.5dB difference of FDP mean deviation between groups was 80%, with a standard deviation of 2.4 and a These charts are retroilluminated, presented in a dark type II error set at 0.05. The analysis was performed using room, and allow testing over a range of contrasts (1.25%, JMP 7.0 (SAS Institute Inc. Cary, NC, USA) 2.5%, 5%, 10%, 25%, and 100%). They use a logarithmicprogression of letter size (0.1 per row), a constant numberof letters per row, and letters of equal legibility, making task difficulty constant regardless of the level of acuity tested.
Testing was conducted monocularly in the study eye at 2 m Forty-eight patients treated with HCQ (51.40 ±11.57 years, under the same mesopic luminance level for all subjects.
range 27–69, M/F 12/36) and 36 age- and sex- matched Patients were tested with the best optical correction, and healthy control subjects (47.7±10.50 years, range 27–69, results are reported as the logarithm of the logMAR.
M/F 7/29) were included in the analysis.
VERIS Clinic™ 4.9 (Electro-Diagnostic Imaging, San Most of the included patients suffered from rheumatoid Mateo, CA, USA) was used for mfERG assessment, using a arthritis (91.7 %), and only four patients (8.3 %) from previously published method []. In all patient and control eyes, mfERG was recorded in the presence of Patients were treated with HCQ at a dosage of 400 mg/ day for an median time of 36 months (range 3.5 months to The average response amplitude densities (RAD in nanovolt/degree between the first negative peak, N1, Age, BCVA and IOP values were similar between and the first positive peak, P1 obtained from five concentric anular retinal regions (rings) centered on the fovea were FDP MD was found to be significantly reduced in HCQ- analysed. The N1-P1 RADs derived from 0 to 2.5 degrees treated patients compared to controls, both in the 24-2 (ring 1, R1), from 2.5 to 5 degrees (ring 2, R2), from 5 to (−1.38±2.41 dB vs 0.21±1.83 dB, p<0.01) and in the 10-2 10 degrees (ring 3, R3), from 10 to 15 degrees (ring 4, R4) program (-0.97 ±2.88 dB vs 0.15±1.72 dB, p<0.02). The and from 15 to 20 degrees (ring 5, R5) were also relationship between 24-2 HFA MD and FDP MD is investigated. Furthermore, the ratio between R3/R1 and R2/R1 was calculated and compared between groups.
FDP pattern standard deviation (PSD) was found to be A signal-to-noise ratio of ≥3 was accepted for mfERG significantly worse in HCQ-treated patients than in controls both in the 24-2 (2.70 ±0.65 dB vs 2.41±0.31 dB, p<0.01)and in the 10-2 program (2.86±0.48 dB vs 2.48 ±0.39 dB, HFA PSD was found to be significantly worse in Demographic and descriptive data were expressed as HCQ-treated patients compared to controls in 24-2 program (1.54 ± 0.34 dB vs 1.40 ±0.33 dB, p < 0.05) Fig .
Normal distribution of data was assessed by the HFA MD 24-2, HFA MD 10-2 and HFA PSD 10-2 were Shapiro–Wilk test. The right eye was arbitrarily chosen similar between patients and controls.
for statistical analysis. Frequencies of categorical variables The CS was found to be significantly reduced in HCQ- were compared between groups by Chi-squared and Fisher’s treated patients compared to controls at the range of exact test as appropriate. Comparisons of continuous variablesbetween groups were performed by independent samples t-test or Mann–Whitney U test as appropriate.
The relationship between functional parameters and duration of HCQ treatment was explored by linearregression analysis.
It has been reported in the literature that HCQ retinal toxicity is correlated with the dose of the drug per body weight, and the American Academy of Opthalmology considers 6.5 mg/kg/day to be the highest safe dose of treatment, at least in the first 5 years []. We then compared retinal sensitivity between patients that received Graefes Arch Clin Exp Ophthalmol (2011) 249:715–721 FDP and HFA parameters were found to be significantly different between each HCQ group and controls. Moreover,FDP and HFA parameters were found to be worse inpatients treated for a longer time than in patients treated forashorter time (<36 months).(Tables , ) Furthermore, FDP MD was found to be statistically significantly different between patients treated with a dosageof HCQ >6.5 mg/kg/day compared to patients treated withlower dosage (−2.43±2.13 vs −0.75±1.95 dB respectively,p<0.05). No significant differences were detected in FDPPSD (2.9±0.65 vs 2.71±0.71 dB respectively p=0.13) or HFAglobal indices (MD: −1.61±1.2 vs 1.52±1.16 dB , p=0.73;PSD: 1.77±0.49 vs 1.47±0.25 dB respectively, p=0.08).
This study supports previous knowledge of an early retinalfunctional impairment in HCQ-treated patients without any Fig. 1 Scatterplot between frequency-doubling perimetry (FDP) mean deviation (MD) and Humphrey field analyzer (HFA) MD. HCQ-treated patients are indicated by circles and control subjects by crosses We found that HCQ-treated patients, without any clinically detectable RPE abnormalities, showed a de- contrast of 10% (0.09±0.07 vs 0.04±0.10; p<0.02), 5% creased threshold of retinal sensitivity as measured by (0.17±0.10 vs 0.12±0.98; p<0.05), 2.5% (0.31±0.12 vs FDP and HFA. FDP mean deviation and pattern standard deviation have been found to be significantly worse in the In the mfERG no statistically significant difference was treated patients than in control subjects. Furthermore, we found at any ring between the two groups in the response found a reduction of contrast sensitivity, of HFA pattern amplitude density (RAD): R1 RAD N1-P1 (HCQ patients standard deviation and of mfERG R3/R1 ring ratio.
84.77±26.60 vs controls 79.31±23.81, p=0.50), R2 RAD Retinal toxicity is probably due to the inhibition of N1-P1 (HCQ patients 42.45±11.10 vs controls 42.14±12.98, lysosomal phospholipases and/or of protein synthesis with p=0.93), R3 RAD N1-P1 (HCQ patients 25.79±7.15 vs multilamellar structures accumulation through the retina.
controls 26.73±8.07, p=0.58), R4 RAD N1-P1 (HCQ Histologically, this accumulation is followed by retinal patients 17.78±5.90 vs controls 18.68±5.83, p=0.50), R5 ganglion cell loss, photoreceptor loss and RPE atrophy RAD N1-P1 (HCQ patients 13.85±4.02 vs controls 13.93±4.23, p=0.70). Furthermore, the mean amplitude ratio of R3/R1 and R2/R1 between the groups has been explored, and astatistically significant difference between HCQ-treatedpatients and controls was found for the R/3/R1 ratio (0.31±0.08 vs 0.36±0.07 respectively, p<0.05), but not for the R2/R1 ratio (0.51±0.09 vs 0.54±0.09 respectively, p=0.22).
No significant linear relationship was found between any functional parameters and the duration of therapy. The Rof FDP 24-2 MD and PSD versus treatment durationwas 0.05 (p=0.12) and 0.06 (p=0.09) respectively.
Differences between functional parameters obtained with FDP, HFA and mfERG were evaluated by dividing thepatients according to the duration of the HCQ therapy.
Patients were divided into three groups: patients takingHCQ for less than 36 months (26 patients, group 1), patientstaking HCQ for more than 36 months (22 patients, group 2)and healthy control subjects (36 patients, group 3). The valueof 36 months was chosen as the cut-off time because it was the Fig. 2 Relationship between frequency-doubling perimetry (FDP) median value of treatment duration in our sample.
mean deviation (MD) and treatment duration (months) Graefes Arch Clin Exp Ophthalmol (2011) 249:715–721 Table 2 Comparison between controls and HCQ-treated patients taking HCQ for more than 36 months (22 patients); group 3: healthy based on duration of therapy in Humphrey-matrix; group 1: patients taking HCQ for less than 36 months (26 patients); group 2: patients Rosenthal et al. reported that toxic doses of CQ adminis- temporal characteristics of Humphrey-matrix 10-2 test do tered to Rhesus monkeys caused an initial dramatic effect not allow to the stimulus to appear to be frequency- on RGCs, followed by photoreceptor loss and RPE doubled, so the threshold determination is a flicker degeneration []. Hallberg et al. found that morphological sensitivity response, which corresponds to a contrast and biochemical signs of phospholipidosis were evident sensitivity test. The changes of Humphrey-matrix 10-2 found in this study are consistent with the results of Bishara Bonanomi et al. showed that patients under long-term et al., who showed that contrast sensitivity is reduced in the CQ treatment could present a significantly reduced RNFL early phases of retinal toxicity ]. Also the low-contrast thickness compared to healthy subjects by means of sensitivity, evaluated with Multi-Contrast Sloan Letter Flip scanning laser polarimetry GDx Nerve Fibre Analyser Book at 2 meters, was found to be reduced early in HCQ- (software v.2.0.01), and that the RNFL loss was correlated treated patients in the present study.
The Humphrey-matrix parameters showed abnormalities Clinically advanced retinal toxicity could be easily even in the presence of normal mfERG;, indeed, no evaluated with different tests as VA, standard visual field statistically significant differences have been found be- testing and ophthalmoscopy, but those tests do not seem to tween HCQ-treated patients and controls in the RAD in the detect early stages For this reason, this study explored mfERG in all rings. In the mfERG, the bioelectrical signal the possible detection of early functional retinal impairment is derived from cones and bipolar cells, with smaller using frequency-doubling technology.
contribution from other retinal neurons. Since RGCs could The finding of a generalized reduction of retinal be affected first, even in the absence of a defect in sensitivity to frequency-doubled stimuli in HCQ-treated photoreceptor or bipolar cell function, we evaluated the patients might be related to a higher susceptibility of the abnormalities in Humphrey-matrix results that selectively magnocellular RGCs to HCQ; this might be responsible for study the MGCs. Early HCQ-induced changes could be the early selective loss of this cell component. Nonetheless, detected early by detecting changes in the M-y ganglion another possible explanation is that the MGCs have similar cell population, which represents a low-redundancy system, susceptibility to HCQ compared to other RGCs, but are and therefore could be followed with frequency-doubling simply less redundant than the whole RGC population, with no overlap between receptive fields, allowing the earlier It has been described in the literature that the first detection of MGC loss if selectively stimulated with changes in mfERG occur in the paracentral regions, where the ratio of the rings may be affected [Although the In this study, patients were tested both 24-2 and 10-2 mean ring amplitudes appeared to be similar in this study programs with both HFA and FDP. The spatial and between HCQ and controls, a statistically significant Table 3 Comparison between controls and HCQ-treated patients based on duration of therapy in HFA. Group 1: patients taking HCQ for lessthan 36 months (26 patients); group 2: patients taking HCQ for more than 36 months (22 patients); group 3: healthy control subjects (36 pts) Graefes Arch Clin Exp Ophthalmol (2011) 249:715–721 difference was found in the ratio between the outer ring R3 The presence of functional retinal changes in late-stage and the inner ring R1, supporting the finding of subtle early HCQ retinopathy could be related to the long duration of the HCQ therapy, but also to the long duration of the It should be also highlighted that more sophisticated, disease. In fact, a limitation of this study is that control non-conventional methods of stimulation with mfERG have subjects are healthy, rather than patients with rheumatic been described as detecting pre-clinical retinal function changes in HCQ-treated patients, emphasizing second-order In conclusion, functional retinal testing by frequency- adaptational effects such as modulated multifocal flashes doubling perimetry could be useful for the identification of with interleaved global flashes, as previously described in early retinal impairment in HCQ-treated patients without any sign of clinically detectable retinal abnormalities, No statistically significant relationship between func- although broader diagnostic studies are required to accu- tional data versus treatment duration was found, as explored rately assess its sensitivity and specificity in clinical by linear regression analysis in the present study. Never- settings, for screening purposes and follow-up.
theless, a difference in functional parameters between long-term treated and short-term treated patients was found,indicating that there might be a threshold effect of treatment duration on visual function change, rather than a linearcontinuous effect causing a continuous decay over time.
1. Rynes RI (1997) Antimalarial drugs in the treatment of rheumato- Nevertheless, despite the fact that the spectrum of treatment logical diseases. Br J Rheumatol 36:799–805 duration was broad in our sample population (3.5 to 2. Mavrikakis I, Sfikakis PP, Mavrikakis E, Rougas K, Nikolaou A, 156 months), most of the patients were treated between Kostopoulos C, Mavrikakis (2003) The incidence of irreversible 3.5 and 50 months (32/48) and only 16/48 for a longer retinal toxicity in patient treated with hydroxychloroquine: areappraisal. Ophthalmology 110:1321–1326 time, indicating that the lack of relationship found between 3. Levy GD, Munz SJ, Paschal J, Cohen HB, Pince KJ, Peterson T functional data and HCQ treatment duration should be (1997) Incidence of hydroxychloroquine retinopathy in 1,207 patients in a large multicenter outpatient practice. Arthritis Rheum Furthermore, the patients_ stratification according to the 4. Tzekov R (2005) Ocular toxicity due to chloroquine and duration of the therapy indicated that patients that assumed hydroxychloroquine: electrophysiological and visual function therapy for a short period of time showed only minimal retinal functional changes, compared to larger changes 5. Marmor MF, Carr RE, Easterbrook M, Farjo AA, Mieler WF, observed in patients taking therapy over a longer period of Academy A, American Academy of Ophthalmology (2002)Recommendations on screening for chloroquine and hydroxy- time. Moreover, previously published studies reported that chloroquine retinopathy: a report by the American Academy of retinal toxicity related to HCQ treatement is related to the Ophthalmology. Ophthalmology 109:1377–1382 daily dosage per kg of body weight, and the American 6. Hallberg A, Naeser P, Andersson A (1990) Effects of long-term Academy of Opthalmology considers 6.5 mg/kg/day to be chloroquine exposure on the phospholipids metabolism in retinaand pigment epithelium of the mouse. Acta Ophthalmol (Copenh) the highest safe dose of treatment, at least in the first 5 years []. In our study, the prescribed dose is constant 7. Rosenthal AR, Kolb H, Bergsma D, Huxsoll D, Hopkins JL (400 mg/day), and we assume that patients are fully (1978) Choroquine retinopathy in the rhesus monkey. Invest compliant, but this could be a source of error in the 8. Easterbrook M (1999) Detection and prevention of maculop- evaluation of the relationship between retinal toxicity and athy associated with antimalarial agents. Int Ophthalmol Clin The results of our study indicate that retinal sensitivity as 9. Carr RE, Gouras P, Gunkel RD (1966) Chloroquine retinopathy.
expressed by FDP MD in patients treated with a daily Early detection by retinal threshold test. Arch Ophthalmol75:171–178 dosage >6.5 mg/kg is reduced compared to patients treated 10. Pasadhika S, Fishman GA (2010) Effects of chronic exposure to with lower daily dosages, while no differences were hydroxychloroquine or chloroquine on inner retinal structures.
detected for other retinal function measures.
The Humphrey-matrix changes are present in the early 11. Bonanomi MT, Dantas NC, Medeiros FA (2006) Retinal nerve fibre layer thickness measurements in patient using chloroquine.
phases, and they could be considered as good candidates for an early detection of the HCQ retinopathy and for its 12. Browning DJ (2002) Hydroxychloroquine retinopathy: screening for drug toxicity. Am J Ophthalmol 133:649–656 The differences in FDP values found in this study are 13. Elder M, Rahman AM, McLay J (2006) Early Paracentral Visual Field Loss in Patients Taking Hydroxychloroquine. Arch Oph- moderate. This is a limitation for their use in clinical practice, even if these results are consistent with pathoge- 14. Bishara SA, Matamoros N (1989) Evaluation of several tests in netic mechanism of HCQ retinal toxicity.
screening for chloroquine maculopathy. Eye 3:777–782 Graefes Arch Clin Exp Ophthalmol (2011) 249:715–721 15. Maturi RK, Yu M, Weleber RG (2004) Multifocal electroretino- macular drusen in age-related macular degeneration. Prognosis graphic evaluation of long-term hydroxychloroquine users. Arch and risk factors. Ophthalmology 1101:1522–1528 20. Parisi V, Perillo L, Tedeschi M, Scassa C, Gallinaro G, 16. Lai TY, Chan WM, Li H et al (2005) Multifocal electroretino- Capaldo N, Varano M (2007) Macular function in eyes with graphic changes in patients receiving hydroxychloroquine therapy.
early age-related macular degeneration with or without contra- lateral late age-related macular degeneration. Retina 27:879– 17. Johnson CA, Cioffi GA, Van Buskirk EM (1999) Frequency- doubling technology perimetry using 24-2 stimulus presentation 21. Lyons JS, Severns ML (2007) Detection of early hydroxy- chloroquine retinal toxicity enhanced by ring ratio analysis of 18. Centofanti M, Fogagnolo P, Oddone F, Orzalesi N, Vetrugno M, multifocal electroretinography. Am J Ophthalmol 143(5):801– Manni G, Rossetti L (2008) Learning effect of humphrey matrix frequency doubling technology perimetry in patients with ocular 22. Penrose PJ, Tzekov RT, Sutter EE, Fu AD, Allen AW Jr, Fung WE, Oxford KW (2003) Multifocal electroretinography evalua- 19. Holz FG, Wolfensberger TJ, Piguet B, Gross-Jendroska M, Wells tion for early detection of retinal dysfunction in patients taking JA, Minassian DC, Chisholm IH, Bird AC (1994) Bilateral hydroxychloroquine. Retina 23(4):503–512

Source: http://www.vincenzoparisi.it/pdf/Tanga_Retinal%20functional%20changes%20measured%20by%20frequency-doubling%20technology%20in%20patients%20treated%20with%20hydroxychloroquine_Graefes_2011.pdf