SAccepted fo

From theCenter (RaB.J.K., T.TUniversity,OphthalmolNew YorkRotterdam, N

InquiriesRadboudumcThomas.The

964

The Effect of Light Deprivation in PatientsWith Stargardt Disease

MICHEL M. TEUSSINK, MICHELE D. LEE, R. THEODORE SMITH, RAMON A.C. VAN HUET,CAROLINE C. KLAVER, B. JEROEN KLEVERING, THOMAS THEELEN, AND CAREL B. HOYNG

� PURPOSE: To investigate whether long-term protectionfrom light exposure affects the rate of disease progressionin patients with autosomal recessive Stargardt disease(STGD1), measured using fundus autofluorescenceimaging.� DESIGN: Longitudinal, retrospective, interventionalcase series.� METHODS: Five patients with Stargardt diseaseprotected 1 eye from light exposure by applying a blackcontact lens during waking hours for ‡12 months. Dis-ease progression was followed by performing autofluores-cence imaging at semi-regular intervals. Longitudinalchanges in autofluorescence were studied by evaluatingareas of decreased autofluorescence and areas of increasedautofluorescence as a measure of retinal pigment epithe-lium damage and lipofuscin accumulation, respectively.� RESULTS: We observed less progression of decreasedautofluorescence in 4 out of 5 light-protected eyes relativeto their respective nonprotected eyes. The progression ofincreased autofluorescence, on the other hand, was highlyvariable and did not respond consistently to treatment.� CONCLUSIONS: Areas of decreased autofluorescencemay serve as a useful biomarker for measuring the pro-gression of Stargardt disease. The reduced progressionof decreased autofluorescence in the light-protected eyessuggests that light deprivation might be beneficial inpatients with Stargardt disease. (Am J Ophthalmol2015;159(5):964–972. � 2015 by Elsevier Inc. Allrights reserved.)

AUTOSOMAL RECESSIVE STARGARDT DISEASE

(STGD1) is the most common inherited juvenilemacular degeneration.1 Most patients develop

bilateral loss of vision in childhood or early adulthood.This subtype of Stargardt disease is caused by mutations

upplemental Material available at AJO.com.r publication Feb 3, 2015.Department of Ophthalmology, Radboud University Medicaldboudumc), Nijmegen, Netherlands (M.M.T., R.A.C.H.,., C.B.H.); Department of Ophthalmology, ColumbiaNew York, New York (M.D.L.); Department of

ogy, New York University School of Medicine, New York,(R.T.S.); and Erasmus University Medical Center,etherlands (C.C.K.).to Thomas Theelen, Department of Ophthalmology,, PO Box 9101, 6500 HB Nijmegen, Netherlands; e-mail:elen@radboudumc.nl

� 2015 BY ELSEVIER INC.

in the ABCA4 gene, which encodes a retina-specifictransporter protein (ABCR) in the rims of rod and conephotoreceptor outer segment discs.2–4 Retinaldegeneration in ABCA4-linked Stargardt disease isbelieved to result from the toxic effects of lipofuscin thataccumulates in the retinal pigment epithelium (RPE) andthe subsequent degeneration of photoreceptors.5

Light can induce photochemical injury at the ocularfundus. Depending on the level and duration of the irradi-ance, the primary site of damage can be either the photore-ceptors or the RPE.6 In ABCA4-linked retinopathies,products generated by the visual cycle accumulate andcontribute to retinal damage via both direct toxic effectsand increased photosensitivity. A major fluorophore of lipo-fuscin, bis-retinoid N-retinylidene-N-retinyl-ethanolamine(A2E), accumulates with other, currently unidentified lipo-fuscin constituents within the RPE.7–9 Thus, an excessiveaccumulation of A2E has been observed in both Abca4-/-

mice and patients with Stargardt disease.5,10 Lipofuscins(and A2E in particular) are potent photosensitizers11–14

that can induce oxidative damage, thereby acceleratinglight-induced retinal damage and RPE atrophy.14–16 Thisoxidative damage may affect the rate of disease progressionin patients with Stargardt disease.The total quantity of A2E oxiranes in Abca4-/- mice in-

creases in response to light exposure.17 In addition, a recentreview of light-induced and inherited retinal degenerationssuggested that light exposure might modify the diseasecourse of genetically well-defined retinal dystrophies,including Stargardt disease.18 An early study used func-tional and ophthalmoscopic examinations to test the effectof 5 years of unilateral light deprivation in a patient withautosomal recessive retinitis pigmentosa and in a patientwith autosomal dominant retinitis pigmentosa; the authorobserved symmetrical disease progression.19 However, inthe absence of a genetic diagnosis, and without clearinsight into the pathogenesis of these 2 RP patients, noconclusions can be drawn with respect to patients withStargardt disease.Because the aforementioned results suggest that patients

with Stargardt disease might be more sensitive to light, weconsidered using light protection as a means to slow diseaseprogression in patients with Stargardt disease. Therefore,we retrospectively examined the effects of using a singleblack contact lens in an attempt to spare disease progres-sion in 1 eye in patients with Stargardt disease. Wefollowed disease progression in 5 patients using the fundus

0002-9394/$36.00http://dx.doi.org/10.1016/j.ajo.2015.02.004

ALL RIGHTS RESERVED.

autofluorescence imaging data, as this test provides a sensi-tive, noninvasive measure of RPE abnormalities and lipo-fuscin accumulation in patients with Stargardt disease.20–22

METHODS

THIS STUDYWASA LONGITUDINAL, RETROSPECTIVE, INTER-

ventional case series of patients diagnosed with Stargardtdisease. The intervention was applied in the context ofclinical care at the request of the patients (or legal guard-ian) and not in the context of a prospective clinical trial.Therefore, nonstandard analyses (eg, routine imaging)were performed to ensure internal quality control and todetect any adverse events in a timely manner. All patientshad been diagnosed previously as having Stargardt diseasewith at least 1ABCA4mutation, and all patients presentedwith the typical clinical symptoms associated with thisretinal dystrophy.

All genetic analyses were performed by the Departmentof Human Genetics at Radboudumc (Nijmegen, theNetherlands). Patients were screened for known ABCA4mutations using the arrayed primer extension microarray(Asper Biotech, Tartu, Estonia), and exon duplicationsand/or deletions were detected using multiplex ligation-dependent probe amplification (P151 and P152;MRC-Holland, Amsterdam, the Netherlands). If nomutations—or only a single heterozygous mutation—were identified, the exons and intron-exon boundarieswere sequenced using the Sanger method to screen for mu-tations in the other allele. All identified mutations wereconfirmed using Sanger sequencing.

The study was performed in accordance with the tenetsestablished by the Declaration of Helsinki, and informedconsent was obtained from all patients (or from the legalguardians of underage patients) for a retrospective analysisof therapy-related data. Patients were previously advisedregarding the potential benefits of wearing sunglasses,avoiding direct light exposure, and limiting their dietaryintake of vitamin A.23 Complete protection from lightexposure was suggested as a treatment option to patientswho had a pressing request for any potentially effectivetreatment, given the current absence of suitable treatmentsfor Stargardt disease. The institutional review board(Commissie Mensgebonden Onderzoek, Region Arnhem-Nijmegen) retrospectively approved this study (2013/062) after the last patient’s final visit and during the study.

Patients were enrolled in the study and followed fromJanuary 2006 through December 2009 at the clinical prac-tice of the Department of Ophthalmology, Radboudumc,Nijmegen. After eliminating any medical concernsregarding the use of contact lenses, we suggested that thepatients protect their intuitively best eye (determined atthe time of enrollment) during waking hours for at least 1year using a black contact lens. The age of disease onset

VOL. 159, NO. 5 LIGHT PROTECTION IN S

was defined as the age at which visual loss was first recorded.Best-corrected visual acuity (BCVA) was measured using aSnellen chart or the finger-counting method. To providethe most complete light protection, a customized, black,soft contact lens designed to cover the entire cornea(Ercon, Assen, the Netherlands) was applied; this lensblocked >90% of light in the visible spectrum.Patients underwent a routine ophthalmologic examina-

tion, including BCVA, slit-lamp examination, binocularfunduscopy, and confocal fundus autofluorescence imagingof both eyes with a Spectralis device (Heidelberg Engineer-ing, Heidelberg, Germany). After pharmacologic pupildilation to >_6 mm with phenylephrine 2.5% and tropica-mide 0.5% eye drops, we recorded fundus autofluorescenceimages centered on the fovea and including part of theoptic disc; images were obtained using an excitation wave-length of 488 nm and a barrier filter (>_500 nm) in high-resolution mode (30-degree field of view, 1536 pixels 31536 pixels). Baseline and follow-up images of both eyeswere captured with equal system sensitivity settings. Clin-ical phenotyping was performed in accordance with Fish-man and associates (1999)24 by an experienced clinician(author R.H.) who was masked with respect to the study re-sults. Fundus autofluorescence was used to phenotype thedisease stages as described by Cideciyan and associates.25

Fundus autofluorescence images were acquired using astandard procedure as follows. First, focusing was performedusing the near-infrared reflection mode of the Spectralis.To account for chromatic aberrations, slight refocusing inthe autofluorescence mode was performed, and 9 separateimages were acquired after sensitivity adjustment, therebycovering the entire macular area and including part ofthe optic disc. To increase the signal-to-noise ratio, all im-ages were aligned and a mean image was calculated usingthe internal software (without normalization by histogramstretch). All other image processing and analysis wasperformed using scripts written in MatLab (R2006; Math-Works, Natick, Massachusetts, USA). Initial and finalfundus autofluorescence images were automatically regis-tered as reported previously.20 The alignment of each im-age pair was manually confirmed; if necessary, registrationwas adjusted iteratively until a satisfactory result was ob-tained. Pixel size was normalized to the fovea-disc margindistance.26

Increased autofluorescence and/or decreased autofluores-cence relative to the image background was quantified byan observer (author M.L.; this author was masked withrespect to the treated eye) based on non-normalizedmean images as published previously.20,27,28 In brief, eachfundus autofluorescence image was leveled by subtractinga 12-zone quadratic polynomial mathematical model ofthe image’s background autofluorescence, which was calcu-lated for each image. The threshold for increased ordecreased autofluorescence was set at 1.5 3 s above orbelow, respectively, the mean pixel intensity of the result-ing image. Consistent with previous reports,20 we found

965TARGARDT DISEASE

that this threshold provided the best detection of visuallyevident autofluorescence abnormalities. In some cases,the precise location of the fovea could not be determinedowing to pathologic changes; in such cases, determiningwhere to place the center of the quadratic polynomialmodel was questionable. Therefore, we adjusted themodel’s position slightly until the selection of any visuallyevident autofluorescence abnormalities was optimized. Forsome images, multiple repositioning of the model wasrequired in order to capture all autofluorescence abnormal-ities; in these cases, we combined the resulting pixelsshowing increased and/or decreased autofluorescence,respectively. Pixels that were identified incorrectly in ves-sels and/or bifurcations were removed manually.

The total number of pixels showing increased and/ordecreased autofluorescence was measured in each image,and the annual rate of change in decreased autofluores-cence (change in decreased AF) in serial image pairs ofthe right eye (OD) was calculated as follows:

% Change in decreased AF ðODÞ=yr ¼�Decreased AFpost ðODÞ � Decreased AFpreðODÞDecreased AFpreðODÞ þ Decreased AFpreðOSÞ

�:100%:½Time between imagesðyrÞ��1

:

The annual rate of change in the left eye (OS) was calcu-lated as follows:

%Change in decreased AFðOSÞ=yr ¼�Decreased AFpostðOSÞ � Decreased AFpreðOSÞDecreased AFpreðODÞ þ Decreased AFpreðOSÞ

�:100%:½Time between imagesðyrÞ��1

:

Dividing by the total initial number of pixels showingdecreased autofluorescence in both eyes minimized poten-tial artificial increases in progression that might have arisenwhen the initial quantity in 1 eye was low. To evaluate theprogression of decreased autofluorescence in the light-protected eye, we normalized its progression to the control(nonprotected) eye as follows:

Progression of decreased AF ðnormalizedÞ ¼�1�

�%Change in decreased AFuntreated=yr�%Change in decreased AFtreated=yr

j% Change in decreased AFuntreated=yrj��

:100%:

All calculations were performed for annual changes inincreased autofluorescence. Statistical analyses were notperformed owing to the relatively small sample size and

966 AMERICAN JOURNAL OF

the absence of empirical data regarding the progression ofaltered autofluorescence under conditions of light depriva-tion.

RESULTS

THE PATIENTS’ DEMOGRAPHICS AND CLINICAL DATA ARE

summarized in Table 1. Five white patients (3 male, 2female) with Stargardt disease (mean age at inclusionwas 22.6 years; range: 10–46 years) enrolled andcompleted the treatment with a mean follow-up periodof 17.8 months (range: 11–26 months). We identifieddisease-related mutations in both ABCA4 alleles of4 patients; only 1 ABCA4 mutation was identified inPatient 5.Treatment compliance was assumed based on informa-

tion provided by the patients or their legal guardians, and

they reported that they complied. The course of BCVAwas similar in both eyes of each patient and was stable in

all but 1 patient (Patient 1) throughout the study period.In Patient 2, whose BCVAwas stable throughout the study,bilateral nuclear cataract was identified at the final follow-up visit. No geographic atrophy was observed in any patientthroughout the study period. None of the study participantsdeveloped any contact lens–related adverse events or signsof deprivation amblyopia.

Figure 1 shows the characteristic fundus autofluores-cence appearance in 1 patient during the study period.All 5 patients presented with a complex textural fundus

MAY 2015OPHTHALMOLOGY

TABLE 1. Demographics, Genetics, and Clinical Appearance of 5 Patients With Stargardt Disease

Patient Sex

Age at

Inclusion

(y)

Age at

Onset

(y)

Disease

Durationa

(mo)

Follow-up

Period (mo) Eye

Initial

BCVA

Final

BCVA

Mutation

Allele 1bMutation

Allele 2b Ophthalmoscopy

Fluorescence

Angiography Electroretinography

Stargardt

Phenotypec

1 F 20 19 41 28 OD

OS*

20/40

20/40

20/100

20/100

c.768G>T c.872C>T Yellowish white flecks in

parafoveal area, central

parafoveal pigment

alterations, bullseye-like

appearance

No dark choroid Within normal limits PF

2 F 46 46 32 27 OD

OS*

20/16

20/12.5

20/16

20/20

c.2588G>C

c.656G>Cd

c.2828G>A Yellowish flecks

throughout the posterior

pole extending anterior

to the vascular arcades

and peripapillary region

No dark choroid, but cSLO

recording, which may

mask the dark choroid

sign

Within normal limits FF

3 M 10 10 22 13 OD*

OS

20/400

20/400

20/400

20/400

c.3335C>A c.5461 -10T>C Yellowish white flecks in

para- and perifoveal

area, central RPE

alterations

No dark choroid OD photopic: moderately

reduced, OS

pathologically reduced

(lowered amplitude)

PF

4 M 13 11 53 29 OD

OS*

20/125

20/125

20/160

20/125

c.2588G>C c.4539 þ1G>T Central RPE alterations

and widespread

atrophic lesions

throughout the posterior

pole extending anterior

to the vascular arcades

Diffuse hyperfluorescent

lesions

Photopic moderately

reduced

DA

5 M 24 12.3 284 12 OD

OS*

20/400

CF

CF

CF

c.IVS39> -10T>C NF Central RPE alterations

with intraretinal

pigmentations; small

atrophic lesions

(presumably resorbed

yellow flecks) scattered

throughout the posterior

pole extending anterior

to the vascular arcades

Not performed OD and OS photopic

moderately reduced, OD

scotopic moderately

reduced, OS scotopic

pathologically reduced

DA

Mean 23 18 86 22

SD 14 17 111 9

BCVA ¼ best-corrected Snellen visual acuity; CF ¼ finger counting; cSLO ¼ confocal scanning laser ophthalmoscopy; DA ¼ diffuse atrophic type; FF ¼ fundus flavimaculatus type; NF ¼ not

found; OD ¼ right eye; OS ¼ left eye; PF ¼ perifoveal type; RPE ¼ retinal pigment epithelium.

*indicates treated/light-protected eyes.aTime between age at onset of the disease and the final image.bMutations in the ABCA4 gene.cAs defined by Fishman et al (1999).24

dAllele unknown.

VOL.

159,N

O.5

967

LIG

HTPROTEC

TIO

NIN

STA

RGARDTD

ISEASE

FIGURE 1. Example of autofluorescence in a patient with Stargardt disease. Fundus autofluorescence was measured and analyzed in apatient with Stargardt disease (Patient 2). The left eye (OS) was protected from light exposure using a black contact lens for24.4 months (see Methods); the right eye (OD) was not light protected. (Top row) Original, nonsegmented autofluorescence images.(Middle row) Increased autofluorescence (pink) and decreased autofluorescence (purple) segmentation. (Bottom row) Longitudinalchange analysis of pixels showing decreased autofluorescence; baseline autofluorescence is shown in orange and autofluorescence atfollow-up is shown in green. The grids demarcate the retinal areas analyzed; the outer radius is 500 pixels adjusted to the fovea-discmargin distance.26

autofluorescence appearance (ie, lesions of increased ordecreased fundus autofluorescence). Central diffuse hypo-fluorescent areas were present in the initial images of 4 pa-tients; however, these areas were not sharply demarcatedand had stronger signals than would be expected in areasof geographic atrophy in Stargardt disease.20

During the study period, pixels showing increased and/ordecreased autofluorescence emerged, disappeared, and/orchanged (ie, from an increase to a decrease or vice versa).We observed that the temporal sequence of autofluores-cence changes was heterogeneous, both between patientsand between a given patient’s eyes (eg, some hypoauto-fluorescent areas expanded, but also regressed partiallyinto the background autofluorescence). This heterogeneitywas observed in visually evident lesions as well as in lesion-free areas. Progression analysis was performed in accor-dance with Cideciyan and associates25 and revealed thatall 5 patients progressed during the study period; 1, 2, and2 patients progressed to stage 3, stage 4, and stage 5, respec-tively.

The total number of pixels showing decreased autofluor-escence generally increased in all but 1 eye. However, in 4of the 5 patients, the normalized annual progression rate in

968 AMERICAN JOURNAL OF

the light-protected eye was less severe compared to therespective untreated eye (Figure 2). In 1 patient, weobserved the opposite pattern; in Patient 3, the progressionof decreased autofluorescence was slightly higher in thelight-protected eye than in the control eye. On the otherhand, the progression of increased autofluorescence wasmore variable than the progression of decreased autofluores-cence; increased autofluorescence developed symmetrically(ie, with similar progression in both the light-protected anduntreated eye) in all patients. The results of our autofluor-escence analysis are summarized in Table 2.

DISCUSSION

HERE, WE REPORT THE EFFECT OF USING A BLACK CONTACT

lens to protect the eye from light exposure in patients withStargardt disease. We observed less progression ofdecreased autofluorescence in 4 out of 5 light-protectedeyes compared to the patients’ unprotected eyes. Incontrast, we found no effect of light protection on the pro-gression of increased autofluorescence. Despite an

MAY 2015OPHTHALMOLOGY

FIGURE 2. Summary of the effect of light exposure on the pro-gression of decreased autofluorescence in 5 patients with Star-gardt disease. Normalized progression of decreasedautofluorescence (AF) measured in the light-protected eyes(black bars) and the untreated eyes (gray bars) of 5 patients.The progression of decreased autofluorescence was reduced inPatients 1, 2, 4, and 5; in contrast, Patient 3 had more progres-sion in the light-protected eye compared to the untreated eye.The value below each patient indicates the treatment duration(in months).

extensive literature search (for details see SupplementalMaterial, available at AJO.com), we were unable to findany published reports regarding the effect of light protec-tion in patients with Stargardt disease.

A recent study using the Abca4-/- mouse model foundthat the accumulation of lipofuscin in the retina makesthe retina more vulnerable to the effects of light withrespect to disease progression.29 How increased levels oflipofuscin lead to photoreceptor degeneration and visualloss in patients with Stargardt disease is poorly understood.Photo-oxidative processes appear to accelerate cell damagein retinas that have increased lipofuscin levels followinglight exposure.29,30 The primary site of this light-induceddamage in patients with Stargardt disease could be photo-receptor outer segments, the RPE, or both. For instance,prolonged light exposure may cause primary photoreceptorcell damage by allowing photopigment regeneration andsubsequent repeated bleaching, which may promote theformation of bisretinoids in photoreceptor cells. Owing toouter segment shedding, however, bisretinoids in photore-ceptor cells are kept to a minimum and instead accumulatein the RPE. Although some studies of light damage have re-ported the greatest damage in the RPE,31–36 others havereported damage in both the RPE and photoreceptors.29

Patients with Stargardt disease can develop increased ordecreased autofluorescence compared to background auto-fluorescence. In Stargardt disease, pixels showing increasedand/or decreased autofluorescence tend to be unstable andcan vary with respect to the temporal sequence of autofluor-escence changes.20Whether a given fundus area with back-ground, increased, or decreased autofluorescence willdevelop a focal increase and/or decrease in autofluorescence

VOL. 159, NO. 5 LIGHT PROTECTION IN S

cannot be predicted at baseline; such a change depends onthe stage of the disease in that specific area.37 In our study,we quantified the increased and/or decreased autofluores-cence relative to background autofluorescence. Becausewe did not measure autofluorescence quantitatively,38 wecannot conclude whether background autofluorescence—and by extension, lipofuscin—is equal to, higher than, orlower than normal in a given patient. Therefore, in contrastto well-defined areas of geographic atrophy, pixels showingdecreased autofluorescence may not necessarily reflect totalRPE atrophy in that region. Pixels showing decreased auto-fluorescence may therefore represent decreased, normal, oreven increased autofluorescence compared to age-matchedhealthy subjects. Nevertheless, pixels showing decreasedautofluorescence (which was referred to in 2009 by Smithand associates20 as ‘‘focally decreased autofluorescence’’)suggest RPE cell damage.20 Because recent quantitativemeasurements showed that background autofluorescenceis nearly always increased in Stargardt disease (comparedto age-matched controls),39 pixels showing increased auto-fluorescence should indicate increased autofluorescencevalues in general. Pixels showing increased autofluores-cencemay therefore be interpreted as an increase in lipofus-cin, even when precise quantification is not available.We observed that a pixel showing increased autofluores-

cence could change to decreased autofluorescence; how-ever, it could also remain increased or even return tobackground autofluorescence levels. On the other hand,pixels showing decreased autofluorescence could emergefrom background autofluorescence or increased autofluores-cence. As discussed above, a pixel showing increased auto-fluorescence suggests a focal area of increased levels offluorophores such as lipofuscin and lipofuscin bisretinoids.Thus, our data suggest that light protection does not affectthe progression of focal lipofuscin accumulation.Interestingly, the 3 youngest patients in our cohort (Pa-

tients 1, 3, and 4) had a decrease in the number of pixelsshowing increased autofluorescence even in the light-protected eye. It should be noted that Patients 1, 3, 4,and 5 were classified as having early-onset Stargardt diseaseand were included in a recent study conducted by ourdepartment.40 This decrease—in combination with an in-crease in the number of pixels showing decreased autofluor-escence in Patients 1, 3, and 4—is likely associated with atrend of decreasing overall autofluorescence.20 This trendmay be more pronounced when the disease progressesrapidly, as occurs in early-onset Stargardt disease.40 Pa-tients 3 and 5, however, had an extremely large centraldiffuse hypoautofluorescent lesion and a large number ofpixels showing decreased autofluorescence. This findingsuggests more destructive disease, as suggested by Lamber-tus and associates.40

On the other hand, decreased autofluorescence appearedto be associated with light exposure in our cohort; in otherwords, light deprivation reduced the progression ofdecreased autofluorescence in 4 of 5 patients. Irradiance

969TARGARDT DISEASE

TABLE 2. Course of Fundus Autofluorescence in 5 Patients With Stargardt Disease

Patient Eye Treatment Period (mo) Initial FIAF (Pixels) Final FIAF (Pixels) FIAF Change %/Year Initial FDAF (Pixels) Final FDAF (Pixels) FDAF Change %/Year

1 OD 1664 326 �16.6 88 199 22.6

OS* 25.9 1897 431 �18.2 129 129 0.0

2 OD 3894 6160 11.6 4231 4342 0.6

OS* 24.4 4650 5745 5.6 4117 4214 0.5

3 OD* 10.9 1464 750 �15.4 12172 14395 6.7

OS 2720 1859 �18.6 17665 19394 5.2

4 OD 2946 1302 �16.9 4583 4987 1.9

OS* 16.0 1127 329 �8.2 4098 4153 0.3

5 OD 1613 2663 27.0 29546 35152 10.4

OS* 12.0 2284 3636 34.7 24363 27187 5.2

Mean 17.8 2426 2320 �1.5 10099 11415 5.4

SD 7.0 1136 2200 20.1 10432 12174 7.0

FDAF ¼ focal areas of decreased autofluorescence; FIAF ¼ focal areas of increased autofluorescence; OD ¼ right eye; OS ¼ left eye.

Images were registered and adjusted for pixel size (ie, the fovea-disc distance had identical pixel numbers in follow-up series). Total numbers

of pixels showing increased and/or decreased autofluorescence in initial and final imagesweremeasured, and yearly changes in the numbers of

pixels showing increased and/or decreased autofluorescence were calculated as a percentage of the total number in the initial images of both

eyes (as % FIAF/year and % FDAF/year, respectively).

*indicates the study eyes.

with visible light of sufficient intensity photobleaches lipo-fuscin autofluorescence in RPE cells, and primary RPEdamage with no anatomic changes in the retinal photore-ceptors can occur following exposure to visible light.30,41

In Abca4-/- mice, light damage in the RPE can also becaused by toxicity attributable to increased bisretinoidsand their oxidation products.29 Therefore, it is likely thatexcessive light exposure is harmful to patients with Star-gardt disease. Other lines of evidence suggest that blue lightmay be primarily responsible for this damage.42–45

Therefore, eyeglasses or contact lenses that filter outshort-wavelength light might be a viable alternative to to-tal light deprivation.

The advantage of our study is that we compared eachlight-protected eye to the untreated eye in the same patient,providing each treated eye with its own internal control. Bycomparing the light-protected eye to the untreated eye inthe same patient, both genetic and environmental effects(aside from the use of the contact lens) were minimized.

On the other hand, this study also has several potentiallimitations. First, we studied a relatively small number of pa-tients with a range of ages and in various stages of Stargardtdisease. Therefore, our cohort included eyes in different dis-ease stages and eyes with different degrees of lipofuscinaccumulation. However, we included only patients withrecent subjective and objective progression, strong motiva-tion to participate in this lengthy treatment, and no contra-indication to contact lens use. Moreover, our approach tolight-deprive the subjectively better eye may have influ-enced our results. However, we believe that any selectionbias was unlikely, as the disease was symmetrical at the startof treatment with respect to both functional and anatomic

970 AMERICAN JOURNAL OF

examination (shown using BCVA, funduscopy, and the num-ber of pixels showing increased autofluorescence anddecreased autofluorescence). In addition, randomization wasnot possible, as this was not a prospective clinical trial. Thetreating physician (author C.H.) felt that any treatment—ifeffective—should be applied to the subjectively better eyein an attempt to preserve function in that eye.Furthermore, because this study was not performed pro-

spectively, it relied on retrospective data analysis. Therefore,autofluorescence imaging was not performed using a stan-dardized protocol, nor were the images obtained at predeter-mined time points. Moreover, media opacity and pupildilation—2 potential confounders of autofluorescence—were not routinely measured before imaging. Takentogether, our study was subject to 2 specific forms of selectionbias: (1) the strong motivation of the patients to participatein this study, and (2) our retrospective quality selection ofthe imaging data.Consistent with previous studies, our study was designed

to observe the progression of increased autofluorescenceand/or decreased autofluorescence during a specific periodof time and to identify any difference in progression be-tween the untreated and treated eyes in the same patient.20

However, our study did not follow the continuous develop-ment of changes in autofluorescence over a prolonged in-terval (ie, until geographic atrophy and vision lossoccurred). A long-term prospective study including morefrequent and regular measurements using standardizedautofluorescence imaging over several years might helpaddress the important question of whether light protectioncan delay vision loss—rather than merely delayingfocal RPE damage—in patients with Stargardt disease.

MAY 2015OPHTHALMOLOGY

Moreover, adding quantitative autofluorescence imaging tofuture study protocols would be highly informative. Unfor-tunately, appropriate research tools for normalizing auto-fluorescence grayscale values to an internal reference38

were not available at the time of our study.In conclusion, our results suggest that decreased auto-

fluorescencemay serve as a biomarker for disease progression

VOL. 159, NO. 5 LIGHT PROTECTION IN S

in patients with Stargardt disease. Under light-deprivationconditions, we found reduced overall progression ofdecreased autofluorescence compared with the untreatedeye, suggesting a reduction in the rate of RPE damage inthe treated eye. Our data suggest that light deprivationmay confer a protective effect in patients with Stargardtdisease and therefore merits further investigation.

ALL AUTHORS HAVE COMPLETED AND SUBMITTED THE ICMJE FORM FOR DISCLOSURE OF POTENTIAL CONFLICTS OF INTERESTand the following were reported. C.K. has received honoraria for participation on advisory boards for Novartis (Basel, Switzerland) and Bayer (Leverkusen,Germany). T.T. has received honoraria for consultancy lectures and for preparing educational presentations from Heidelberg Engineering (Heidelberg,Germany). All other authors have no commercial financial support or financial conflict of interests to disclose. This research was supported by a grantfrom the Foundation Fighting Blindness, Columbia, Maryland (grant no. C-CL-0811-0549-RAD05), and by the Gelderse Blindenstichting, Velp, theNetherlands. These funding organizations had no role in the design or execution of this research. Contributions of authors: design of study (T.T.,C.H.); data collection (M.T.); management, analysis, and interpretation of the data (M.T., M.L., R.S., R.H., C.K., B.K., T.T.); provision of materials,patients, and resources (R.S., C.K., B.K., C.H.); preparation of the manuscript (M.T., M.L., T.T.); review of the manuscript (R.S., R.H., C.K., B.K.,C.H.); approval of the manuscript (M.T., M.L., R.S., R.H., C.K., B.K., T.T., C.H.); and responsibility for the integrity of the entire study and manuscript(M.T., M.L., R.S., R.H., R.H., C.K., B.K., T.T., C.H.).

The authors are grateful to Dr C.F. Boon, Leiden University Medical Center, Leiden, the Netherlands, for data acquisition and valuable intellectualdiscussions.

REFERENCES

1. Klevering BJ, Deutman AF, Maugeri A, Cremers FPM,Hoyng CB. The spectrum of retinal phenotypes caused bymu-tations in theABCA4 gene.Graefes Arch Clin ExpOphthalmol

2005;243(2):90–100.2. Allikmets R, Singh N, Sun H, et al. A photoreceptor cell-

specific ATP-binding transporter gene (ABCR) is mutatedin recessive Stargardt macular dystrophy. Nat Genet 1997;15(3):236–246.

3. Azarian SM, Travis GH. The photoreceptor rim protein is anABC transporter encoded by the gene for recessive Stargardt’sdisease (ABCR). FEBS Lett 1997;409(2):247–252.

4. Illing M, Molday LL, Molday RS. The 220-kDa rim protein ofretinal rod outer segments is a member of the ABC trans-porter superfamily. J Biol Chem 1997;272(15):10303–10310.

5. Weng J, Mata NL, Azarian SM, Tzekov RT, Birch DG,Travis GH. Insights into the function of Rim protein in pho-toreceptors and etiology of Stargardt’s disease from thephenotype in Abcr knockout mice. Cell 1999;98(1):13–23.

6. Organisciak DT, Vaughan DK. Retinal light damage: mecha-nisms and protection. Prog Retin Eye Res 2010;29(2):113–134.

7. Eldred GE, LaskyMR. Retinal age pigments generated by self-assembling lysosomotropic detergents. Nature 1993;361(6414):724–726.

8. Ablonczy Z, Higbee D, Anderson DM, et al. Lack of correla-tion between the spatial distribution of A2E and lipofuscinfluorescence in the human retinal pigment epithelium. InvestOphthalmol Vis Sci 2013;54(8):5535–5542.

9. Mata NL,Weng J, Travis GH. Biosynthesis of a major lipofus-cin fluorophore in mice and humans with ABCR-mediatedretinal and macular degeneration. Proc Natl Acad Sci U S A2000;97(13):7154–7159.

10. Mata NL, Tzekov RT, Liu X, Weng J, Birch DG, Travis GH.Delayed dark-adaptation and lipofuscin accumulation inAbcrþ/- mice: implications for involvement of ABCR inage-related macular degeneration. Invest Ophthalmol Vis Sci2001;42(8):1685–1690.

11. Gaillard ER, Atherton SJ, Eldred G, Dillon J. Photophysicalstudies on human retinal lipofuscin. Photochem Photobiol1995;61(5):448–453.

12. Rozanowska M, Jarvis-Evans J, Korytowski W, Boulton ME,Burke JM, Sarna T. Blue light-induced reactivity of retinalage pigment. In vitro generation of oxygen-reactive species.J Biol Chem 1995;270(32):18825–18830.

13. Rozanowska M, Wessels J, Boulton M, et al. Blue light-induced singlet oxygen generation by retinal lipofuscin innon-polar media. Free Radic Biol Med 1998;24(7-8):1107–1112.

14. Davies S, Elliott MH, Floor E, et al. Photocytotoxicity of lipo-fuscin in human retinal pigment epithelial cells. Free RadicBiol Med 2001;31(2):256–265.

15. Godley BF, Shamsi FA, Liang FQ, Jarrett SG, Davies S,Boulton M. Blue light induces mitochondrial DNA damageand free radical production in epithelial cells. J Biol Chem2005;280(22):21061–21066.

16. Shamsi FA, Boulton M. Inhibition of RPE lysosomal andantioxidant activity by the age pigment lipofuscin. InvestOphthalmol Vis Sci 2001;42(12):3041–3046.

17. Radu RA, Mata NL, Bagla A, Travis GH. Light exposurestimulates formation of A2E oxiranes in a mouse model ofStargardt’s macular degeneration. Proc Natl Acad Sci U S A

2004;101(16):5928–5933.18. Paskowitz DM, LaVail MM, Duncan JL. Light and inherited

retinal degeneration. Br J Ophthalmol 2006;90(8):1060–1066.19. Berson EL. Light deprivation and retinitis pigmentosa. Vision

Res 1980;20(12):1179–1184.20. Smith RT, Gomes NL, Barile G, Busuioc M, Lee N, Laine A.

Lipofuscin and autofluorescence metrics in progressiveSTGD. Invest Ophthalmol Vis Sci 2009;50(8):3907–3914.

21. Smith RT. Fundus autofluorescence patterns in Stargardt dis-ease over time. Arch Ophthalmol 2012;130(10):1354; authorreply 1354–1355.

22. McBain VA, Lois N. Progression of retinal pigment epithelialatrophy in Stargardt disease. Am J Ophthalmol 2012;154(1):146–154.

971TARGARDT DISEASE

23. Allikmets R. Stargardt disease. In: Tombran-Tink J,Barnstable CJ, eds. Retinal Degenerations. Totowa, New Jer-sey: Humana Press; 2007:105–118.

24. Fishman GA, Stone EM, Grover S, Derlacki DJ, Haines HL,Hockey RR. Variation of clinical expression in patients withStargardt dystrophy and sequence variations in the ABCRgene. Arch Ophthalmol 1999;117(4):504–510.

25. Cideciyan AV, Aleman TS, Swider M, et al. Mutations inABCA4 result in accumulation of lipofuscin before slowingof the retinoid cycle: a reappraisal of the human diseasesequence. Hum Mol Genet 2004;13(5):525–534.

26. Smith RT, Nagasaki T, Sparrow JR, Barbazetto I,Klaver CC, Chan JK. A method of drusen measurementbased on the geometry of fundus reflectance. Biomed Eng On-line 2003;2:10.

27. Smith RT, Koniarek JP, Chan J, Nagasaki T, Sparrow JR,Langton K. Autofluorescence characteristics of normalfoveas and reconstruction of foveal autofluorescence fromlimited data subsets. Invest Ophthalmol Vis Sci 2005;46(8):2940–2946.

28. Smith RT, Chan JK, Busuoic M, Sivagnanavel V, Bird AC,Chong NV. Autofluorescence characteristics of early, atro-phic, and high-risk fellow eyes in age-related maculardegeneration. Invest Ophthalmol Vis Sci 2006;47(12):5495–5504.

29. Wu L, Ueda K, Nagasaki T, Sparrow JR. Light damage inAbca4 and Rpe65rd12 mice. Invest Ophthalmol Vis Sci 2014;55(3):1910–1918.

30. Hunter JJ, Morgan JI, Merigan WH, Sliney DH, Sparrow JR,Williams DR. The susceptibility of the retina to photochem-ical damage from visible light. Prog Retin Eye Res 2012;31(1):28–42.

31. Ham WT Jr, Mueller HA, Sliney DH. Retinal sensitivity todamage from short wavelength light. Nature 1976;260(5547):153–155.

32. Ham WT Jr, Mueller HA, Ruffolo JJ Jr, et al. Basic mecha-nisms underlying the production of photochemical lesionsin the mammalian retina. Curr Eye Res 1984;3(1):165–174.

33. HamWT Jr, Allen RG, Feeney-Burns L. The involvement ofthe retinal pigment epithelium. In: Waxler M, Hitchins VM,eds. CRC Optical Radiation and Visual Health. Boca Raton,Florida: CRC Press; 1986:43–67.

972 AMERICAN JOURNAL OF

34. Friedman E, Kuwabara T. The retinal pigment epithelium.IV. The damaging effects of radiant energy. Arch Ophthalmol1968;80(2):265–279.

35. Busch EM, Gorgels TG, Roberts JE, van Norren D. The ef-fects of two stereoisomers of N-acetylcysteine on photochem-ical damage by UVA and blue light in rat retina. PhotochemPhotobiol 1999;70(3):353–358.

36. Ham WT Jr, Ruffolo JJ Jr, Mueller HA, Clarke AM,Moon ME. Histologic analysis of photochemical lesions pro-duced in rhesus retina by short-wave-length light. Invest

Ophthalmol Vis Sci 1978;17(10):1029–1035.37. Charbel Issa P, BarnardAR, SinghMS, et al. Fundus autofluor-

escence in the Abca4-/- mouse model of Stargardt disease–Correlation with accumulation of A2E, retinal function, andhistology. Invest Ophthalmol Vis Sci 2013;54(8):5602–5612.

38. Delori F, Greenberg JP, Woods RL, et al. Quantitative mea-surements of autofluorescence with the scanning laserophthalmoscope. Invest Ophthalmol Vis Sci 2011;52(13):9379–9390.

39. Burke TR, Duncker T, Woods RL, et al. Quantitative fundusautofluorescence in recessive Stargardt disease. InvestOphthalmol Vis Sci 2014;55(5):2841–2852.

40. Lambertus S, van Huet RAC, Bax NM, et al. Early-onsetStargardt disease: phenotypic and genotypic characteristics.Ophthalmology 2015;122(2):335–344.

41. Morgan JI, Hunter JJ, Masella B, et al. Light-induced retinalchanges observed with high-resolution autofluorescence im-aging of the retinal pigment epithelium. Invest OphthalmolVis Sci 2008;49(8):3715–3729.

42. Zhou J, Sparrow JR. Light filtering in a retinal pigmentepithelial cell culture model. Optom Vis Sci 2011;88(6):759–765.

43. Sparrow JR, Nakanishi K, Parish CA. The lipofuscin fluoro-phore A2E mediates blue light-induced damage to retinalpigmented epithelial cells. Invest Ophthalmol Vis Sci 2000;41(7):1981–1989.

44. Sparrow JR, Cai BL. Blue light-induced apoptosis of A2E-containing RPE: involvement of caspase-3 and protectionby bcl-2. Invest Ophthalmol Vis Sci 2001;42(6):1356–1362.

45. Schutt F, Davies S, Kopitz J, Holz FG, BoultonME. Photodam-age to humanRPE cells by A2-E, a retinoid component of lipo-fuscin. Invest Ophthalmol Vis Sci 2000;41(8):2303–2308.

MAY 2015OPHTHALMOLOGY

SUPPLEMENTAL MATERIAL

AN EXTENSIVE LITERATURE SEARCH WAS PERFORMED USING THE

following search string in the PubMed database (availableat http://www.ncbi.nlm.nih.gov/). The search wasperformed on November 2, 2014.

(irradiation[tiab] OR irradiations[tiab] OR photoradia-tion[tiab] OR photoradiations[tiab] OR radiations[tiab]OR illumination[tiab] OR light[tiab] OR photo[tiab] OR(intraocular[tiab] AND light[tiab])) AND

((treat[tiab] OR treatment[tiab] OR treated[tiab]) OR(protect[tiab] OR protection[tiab] OR protected[tiab])OR (filter[tiab] OR filtering[tiab] OR filtered[tiab]) OR(shield[tiab] OR shielding[tiab] OR shielded[tiab]OR block[tiab] OR blocking[tiab] OR blocked[tiab]) OR(avoid[tiab] OR avoidance[tiab] OR avoided[tiab]) OR(deprived[tiab] OR deprivation[tiab] OR deprive[tiab]))AND

(fundi[tiab] OR fundus[tiab] OR ‘‘retina’’[MeSHTerms] OR ‘‘retina’’[tiab] OR retinal[tiab] OR retinas[tiab] OR ‘‘macula’’[tiab] OR maculas[tiab] OR ‘‘macu-lar’’[tiab] OR ‘‘fovea’’[tiab] OR ‘‘foveal’’[tiab] OR‘‘foveas’’[tiab] OR photoreceptor[tiab] OR photorecep-tors[tiab] OR ‘‘cone’’[tiab] OR cones[tiab] OR rod[tiab]

VOL. 159, NO. 5 LIGHT PROTECTION IN S

OR rods[tiab]OR amacrine[tiab] OR (horizontal[tiab]AND (cell[tiab] OR cells[tiab])) OR (optic[tiab] AND(disk[tiab] OR disc[tiab])) OR choroid[tiab] OR choroid[mesh]) AND(‘‘Stargardt Macular Degeneration’’ [Supplementary

Concept] OR ‘‘Stargardt disease 1’’ [Supplementary Concept]OR ‘‘Stargardt disease 3’’ [Supplementary Concept] OR Star-gardt[tiab] OR Fundus Flavimaculatus [tiab] OR STGD[tiab]OR STGD1[tiab] OR STGD3 [tiab] OR STGD4 [tiab]) OR(abca4[tiab] OR ((ABC1[tiab] OR ABC[tiab]) AND (trans-porter[tiab] OR transporters[tiab]) AND (photoreceptor[tiab] OR photoreceptor[tiab])) OR ((ATP-Binding CassetteTransporters [Mesh]) AND (photoreceptor cells [Mesh] ORphotoreceptor [tiab] OR retina [Mesh] OR retina [tiab]) OR(rim[tiab] AND (protein[tiab] OR proteins [tiab]) OR Abcr[tiab]))) AND(patient[tiab] OR patients[tiab] OR trial[tiab] OR (clin-

ical[tiab] AND trial[tiab]))NOT (tuberculosis[tiab] OR tuberculous[tiab] OR tuber-

cular[tiab] OR bacterial[tiab] OR bacterium[tiab] ORbacillary[tiab] OR tumor[tiab] OR tumors[tiab] OR rifam-picin[tiab] or (regional[tiab] AND myocardial[tiab] ANDperfusion[tiab]))

972.e1TARGARDT DISEASE

Biosketch

Michel M. Teussink, MSc, received his Masters degree in Medical Biology from the Radboud University Nijmegen with a

specialty in biomedical imaging. He is experienced in advanced microscopical techniques and analyses, and is currently

completing his PhD project on functional optical coherence tomography and imaging biomarkers of retinal diseases. He

has co-authored 3 publications in peer reviewed journals. Mr. Teussink is PhD student at the Department of

Ophthalmology, Radboudumc, Nijmegen, The Netherlands.

972.e2 MAY 2015AMERICAN JOURNAL OF OPHTHALMOLOGY