Hyprereflective foci in Diabetic Macular edema

oaMm

s

rwcahlt

tEV©

7

Association Between Hyperreflective Foci in the OuterRetina, Status of Photoreceptor Layer, and Visual Acuity

in Diabetic Macular Edema

AKIHITO UJI, TOMOAKI MURAKAMI, KAZUAKI NISHIJIMA, TADAMICHI AKAGI, TAKAHIRO HORII,

t(tttstl

● PURPOSE: To determine if hyperreflective foci in theuter retina are associated with photoreceptor integritynd the logarithm of minimal angle of resolution (log-AR) visual acuity (VA) in patients with diabeticacular edema (DME).

● DESIGN: Retrospective, observational, cross-sectionaltudy.

● METHODS: Patients (n � 76; 108 eyes) with clinicallyelevant macular edema and no serous retinal detachmentere analyzed retrospectively. Spectral-domain optical

oherence tomography (SD-OCT) images were obtained forll patients. We investigated the relationship between theyperreflective foci in the outer retinal layers of the externalimiting membrane (ELM) at the fovea and the photorecep-or integrity and VA.

● RESULTS: SD-OCT showed that 58 eyes (53.7%) hadhyperreflective foci in the outer retinal layers, and 107eyes (99.1%) had hyperreflective foci in the inner retinallayers. The logMAR VA was significantly (P < .0001)worse in eyes with hyperreflective foci in the outerretinal layers than in eyes without them (0.463 � 0.382vs 0.127 � 0.206, respectively). Disruption of the ELMline on OCT was significantly (P < .0001, for bothcomparisons) associated with both hyperreflective foci inthe outer retinal layers and poor logMAR VA. Disrup-tion of the junction of the inner and outer segment line(IS/OS) also was related to hyperreflective foci in theouter retinal layers and poor logMAR VA (P < .0001for both comparisons).● CONCLUSIONS: The presence of hyperreflective foci inhe outer retina is closely associated with a disruptedLM and IS/OS line on SD-OCT images and decreasedA in DME. (Am J Ophthalmol 2012;153:710–717.2012 by Elsevier Inc. All rights reserved.)

Accepted for publication Aug 26, 2011.From the Department of Ophthalmology and Visual Sciences, Kyoto

University Graduate School of Medicine, Kyoto, Japan.Inquiries to Akihito Uji, Department of Ophthalmology and Visual

Sciences, Kyoto University Graduate School of Medicine, 54 ShogoinKawahara-cho, Sakyo-ku, Kyoto 606-8507, Japan; e-mail: akihito1@

© 2012 BY ELSEVIER INC. A10

D IABETIC MACULAR EDEMA (DME) IS A LEADING

cause of visual impairment in patients with dia-betes. The Wisconsin Epidemiologic Study of

Diabetic Retinopathy reported that the incidence of DMEover a 10-year period was 20.1% in a younger-onset group,25.4% in an older-onset group taking insulin, and 13.9%in an older-onset group not taking insulin.1 Since thenumber of patients with diabetic retinopathy (DR) andvision-threatening DR is expected to increase,2 furtherresearch is needed into development of treatments andevaluation techniques for DME.

Optical coherence tomography (OCT) has been usedwidely to assess the pathologic structures in the macula andobjectively measure the retinal thickness.3–6 To date, amodest correlation between OCT-measured central retinalthickness and visual acuity (VA) has been reported tosome extent in DME.5–11 Despite the clinical relevance ofthis finding, clinicians frequently have observed eyes withpoor visual outcomes regardless of complete resolution ofthe macular edema, and the VA is highly variable inassociation with a given degree of retinal edema.12,13

These clinical findings suggested that a pathogenesis otherthan macular thickening may be at play and should beinvestigated.

In addition to quantification, later generations of OCTalso enable clinicians to appreciate the individual retinallayers qualitatively. In particular, spectral-domain (SD)-OCT clearly delineates the external limiting membrane(ELM), which corresponds to the adherens junctionsbetween the Müller cells and photoreceptor cells and alsoprevents passage of macromolecules.14 The junction be-ween the inner and outer segments of the photoreceptorsIS/OS) on OCT images might represent the photorecep-or function, which is light perception. These layers onhe OCT images reportedly provide important informa-ion about pathologies of the photoreceptors. Severaltudies have reported a correlation between damage tohe foveal photoreceptors and the VA in retinal vascu-ar diseases.12,13,15–18 Collectively, those studies have

suggested both the clinical relevance of the photorecep-tor damage and a novel pathogenesis in DME thatinvolves photoreceptor dysfunction and macular thick-

LL RIGHTS RESERVED. 0002-9394/$36.00doi:10.1016/j.ajo.2011.08.041

shftaresd

3aat1cmsStmih

lhrTos(t5w

ac

iwascI

ra

ldvfmlpodflttwmsoraaEBofcg1opw

cpltco

Recently, Bolz and associates used SD-OCT to showthat hyperreflective foci are associated with DME and thatthese structures are scattered throughout all retinal lay-ers.19 The authors postulated that these foci might repre-ent the initial stages in the development of intraretinalard exudates and that the foci might be subclinical

eatures of lipoprotein extravasation after breakdown ofhe inner blood-retina barrier. Considering this, Ota andssociates, who have studied hyperreflective dots in serousetinal detachments (SRDs), speculated that fluid andxudates, such as lipids and proteins, that flow into theubretinal space through the outer retinal cleavage may beeposited as subfoveal hard exudates.20–22

In the current study, we characterized the sparse hyper-reflective foci, subclinical findings that are invisible duringclinical ophthalmoscopic examinations, and investigatedtheir association with the photoreceptor status and visualfunction in DME without SRDs.

PATIENTS AND METHODS

WE RETROSPECTIVELY EXAMINED 108 EYES OF 76 PATIENTS

(mean age � standard deviation, 65.0 � 9.1 years; range,5-82 years; 5 patients with type 1 diabetes mellitus [DM]nd 71 with type 2 DM) who had visual disturbancesttributable to clinically relevant macular edema and werereated at Kyoto University Hospital between December2, 2007 and March 30, 2010. All patients had undergoneomprehensive ophthalmologic examinations, includingeasurement of the best-corrected visual acuity (BCVA),

lit-lamp biomicroscopy, color fundus photography, andD-OCT. We excluded eyes with SRD, hard exudates athe fovea, other retinal diseases (including uveitis, vitreo-acular traction syndrome, or macular pucker), or signif-

cant media opacities, such as cataract or vitreousemorrhage.

● OPTICAL COHERENCE TOMOGRAPHY: The morpho-ogic features of DME, including the ELM, IS/OS, andyperreflective foci, were assessed with SD-OCT (Spect-alis; Heidelberg Engineering, Heidelberg, Germany).wenty-degree radial scans centered on the fovea werebtained in a clockwise manner in all eyes, and 49 rastercans were used to evaluate the mean retinal thicknessfrom the innermost internal limiting membrane [ILM] tohe retinal pigment epithelium [RPE]) at the fovea (radius,00 �m). The central fovea was defined as the locationithout the inner retinal layers in the macular region.23

Subsequently, the morphologic features within a scanned1-mm area centered on the fovea were evaluated for eacheye. The status of the ELM and IS/OS was classifiedaccording to previously described methods with somemodifications.16,18,24 Briefly, eyes with an ELM line thatppeared to be complete at the fovea in all scans were

HYPERREFLECTIVE FOCI, PHOTORECEPTOR LAYER,VOL. 153, NO. 4

n 1 scan or more were classified as ELM (�); and eyesith an undetectable ELM line in the fovea were classifieds ELM (�). Each eye also was classified based on thetatus of the IS/OS line beneath the fovea using the sameriteria described for the ELM line, that is, IS/OS (�),S/OS (�), and IS/OS (�).

Since a number of studies have reported the photo-eceptor status as a novel parameter in macular edemassociated with retinal vascular diseases,12,13,15–18,25,26

we evaluated the presence of the hyperreflective foci inthe outer retinal layers. It was previously reported thatELM acts as a barrier against macromolecules,14 whiched us to hypothesize that an intact ELM might blockiffusion of proteins or lipids from the damaged retinalasculature, whereas they might otherwise migratereely within the inner retinal layers. In other words,igration of hyperreflective foci into the outer retinal

ayers of the ELM might indicate that the barrierroperties of the ELM are disrupted or that the inner oruter segments necessary for light perception have beenamaged. Thus, we evaluated and defined the hyperre-ective foci in the outer retinal layers from the ELM tohe RPE within each 1-mm scanned area centered onhe fovea (Figure 1). As discussed previously, the ELMas often interrupted or absent at the fovea. When 2 orore fragments of the ELM line were detected in the

canned area, we drew imaginary lines between the endsf the ELM lines (Figure 2). When the ELM lineeached the outer side without interruption and endedt the RPE-choriocapillaris complex, we considered therea where the ELM line was absent as being above theLM line and did not draw an additional line (Figure 2,ottom row). The sparse hyperreflective foci were roundr oval as reported previously,19 and the hyperreflectiveoci aggregates were often seen on the OCT images andorresponded to hard exudates in the fundus photo-raphs. We excluded eyes with aggregates within the-mm central scanned area (Figure 1), because the aimf this study was to evaluate the association between thehotoreceptor status and sparse hyperreflective foci,hich are considered as subclinical hard exudates.Two experienced masked examiners (Y.M., N.A.)

lassified the status of the ELM and the IS/OS and theresence of the hyperreflective foci, and a third higher-evel grader (T.H.) classified any images about whichhere was disagreement. The kappa coefficient wasomputed as a measure of reliability between thebservers.

● STATISTICAL ANALYSIS: All values are expressed asthe mean � standard deviation. All BCVA measure-ments were converted to the logarithm of the minimalangle of resolution (logMAR) equivalents before statis-tical analysis. Student t tests and �2 tests were used tocompare the 2 hyperreflective foci groups (foci absent in

AND VISUAL ACUITY IN DIABETIC MACULAR EDEMA 711

lg

ptwehfr(h

tlfsWrftellb

retinal layers group) regarding sex, age, duration ofdiabetes, HbA1C, total cholesterol, the presence ofhypertension, logMAR VA, type of DR, the hyperre-flective foci in the inner retinal layers of ELM line(Figure 1), and the integrity of the IS/OS and the ELMlines. The �2 test also was used to analyze the relation-ship between the integrity of the ELM line and theIS/OS line. Comparisons of the logMAR VA levels ofthe 3 groups that were classified based on the status ofthe IS/OS or ELM lines and comparison of the fovealthickness of the 3 ELM groups were carried out using1-way analysis of variance, with post hoc comparisonstested by Scheffe procedure. The relationship betweenthe logMAR VA and the foveal thickness was analyzedusing the Pearson correlation coefficient. A P value �.05 was considered statistically significant. All analyses,except for the kappa coefficient, were performed usingStatView version 5.0 (SAS Institute, Cary, North Car-olina, USA). Calculation of the kappa coefficient wasperformed using SPSS version 17 (GraphPad Software,La Jolla, California, USA).

RESULTS

● ASSESSMENT OF INTER-OBSERVER AGREEMENT: Thekappa coefficient was 0.897 (P � .001) for the ELM, 0.833

FIGURE 1. Localization of hyperreflective foci at the fovea.tomography (OCT) images in diabetic retinopathy (Top rightphotograph (Top left). OCT shows sparse hyperreflective foci (membrane (Middle right) or the outer retinal layers (Bottomcorresponding area in the color fundus photographs (Middle le

(P � .001) for the IS/OS, and 0.926 (P � .001) for (

AMERICAN JOURNAL OF712

ocation of the hyperreflective foci. These results indicatedood inter-observer agreement.

● HYPERREFLECTIVE FOCI IN THE OUTER RETINAL LAY-

ERS OF THE EXTERNAL LIMITING MEMBRANE: Recentublications have documented the presence of hyperreflec-ive foci in Spectralis OCT images of DR, but it is unclearhether the finding is related to visual function, whichncouraged us to investigate the clinical relevance ofyperreflective foci in DME. We observed hyperreflective

oci in the central 1-mm area of the fovea in the inneretinal layers in 107 of the 108 eyes (99.1%) with DMEFigure 1). However, only 58 of the 108 eyes (53.7%) hadyperreflective foci in the outer retinal layers.The logMAR VA was significantly (P � .0001) worse in

he group with hyperreflective foci in the outer retinalayers compared with the group without hyperreflectiveoci in the outer retinal layers (Table 1), although no otherystemic or ocular parameters differed between the groups.

e further investigated how hyperreflective foci in the outeretinal layers contributed to the pathogenesis of DME andound that 45 of 50 eyes (90%) without hyperreflective foci inhe outer retinal layers had an intact ELM, whereas 45 of 58yes (77.6%) with hyperreflective foci in the outer retinalayers had a disrupted ELM (P � .0001) (Table 2). TheogMAR VA in the ELM (�) group also was significantlyetter than that in either the ELM (�) or ELM (-) group

uent hyperreflective foci (black arrow) on optical coherenceich correspond to subfoveal hard exudates in a color fundusarrows) in the inner retinal layers inside the external limitingt) at the fovea, whereas hard exudates are not seen in the

d Bottom left).

Confl), whblack

righft an

P � .0001 for both comparisons). These results suggested

OPHTHALMOLOGY APRIL 2012

bplhrtEh3h

ctItitv

msTfmnhg

that the pathologic association between the hyperreflectivefoci in the outer retinal layers and disrupted ELM maycontribute to visual dysfunction in DME.

● ASSOCIATION BETWEEN HYPERREFLECTIVE FOCI

AND PHOTORECEPTOR STATUS: Despite the clinicalrelevance of ELM disruption in the OCT images, whichaffects visual function but not light perception or signaltransduction per se,14 the major function of the ELM seems toe that of a barrier. We also evaluated another marker ofhotoreceptor integrity, the IS/OS, which directly representsight perception. We found that eyes with an intact IS/OSad significantly better VA than eyes with either an inter-upted or absent IS/OS (P � .0021 and P � .0001, respec-ively). All 27 eyes with an intact IS/OS also had an intactLM, whereas 50 of 81 eyes (61.7%) with a disrupted IS/OSad either an interrupted or absent ELM (P � .0001) (Table). In addition, 96.3% of eyes with an intact IS/OS had no

FIGURE 2. A disrupted external limiting membrane (ELM) anhyperreflective foci in the outer retinal layers in diabetic retinoan intact (Top row), discontinuous (Second row), or absent (ThirThe ELM line is intact and the hyperreflective foci are present beninner segment/outer segment (IS/OS) line (black arrowhead) is seearea between the black arrows. Imaginary lines (red dotted lines) cbeneath the lines (black arrowheads). (Second row) The fragmentdepicted within a 1-mm central scanned area. (Third row) The ELsank and merged with the retinal pigment epithelium (RPE) (blacfoci (black arrowheads) are seen in the areas where the ELM and

yperreflective foci in the outer retinal layers, especially l

HYPERREFLECTIVE FOCI, PHOTORECEPTOR LAYER,VOL. 153, NO. 4

ompared to the higher frequency of hyperreflective foci inhe outer retinal layers (71.3%) in eyes with a disruptedS/OS (P � .0001) (Table 3). These results suggested thathe pathologic association between the hyperreflective focin the outer retinal layers and a disrupted ELM may affecthe damage in the cone inner or outer segments and theisual function.

● MODEST ASSOCIATION BETWEEN FOVEAL THICKNESS

AND HYPERREFLECTIVE FOCI: The macular thicknessight represent the magnitude of the macular edema, the

everity of the damage in signal transduction, or both.hus, when we investigated whether the hyperreflective

oci in the outer retinal layers were associated with theacular thickening, we found that the mean foveal thick-ess was significantly (P � .0081) higher in the group withyperreflective foci in the outer retinal layers than in theroup with no hyperreflective foci in the outer retinal

e junction between the inner and outer segments in eyes withy. Eyes with hyperreflective foci in the outer retinal layers have) ELM on the optical coherence tomography images. (Top row)

the line (black arrow), whereas discontinuity of the photoreceptorecond row and Third row) The ELM line is not continuous in thect the terminals of the ELM and hyperreflective foci are observedoth the ELM line (black arrows) and the IS/OS line (asterisk) ared IS/OS lines are absent. (Bottom row) The ELM lines sometimesows). An irregular RPE line (red dotted line) and hyperreflectiveS are absent.

d thpathd roweathn. (Sonnes of bM ank arrIS/O

ayers (Table 4). In addition, the mean foveal thickness

AND VISUAL ACUITY IN DIABETIC MACULAR EDEMA 713

cbsopvmls

flt

inim

was correlated significantly (P � .0033) with the ELMstatus at the fovea (Table 4) and the mean foveal thicknesswas modestly correlated with the logMAR VA (R � 0.275;P � .0038).

DISCUSSION

A NUMBER OF STUDIES HAVE PROPOSED THAT DISRUPTION

of the photoreceptors in eyes with macular edema resultsfrom retinal vascular disease as a novel parameter thatpredicts visual function.12,13,15–18,25,26 However, it is un-lear what induces the photoreceptor damage after thereakdown of the blood-retinal barrier. The current studyhowed for the first time that hyperreflective foci in theuter retinal layers are associated with the disruptedhotoreceptor markers, ELM and IS/OS, and concomitantisual impairment, which suggested both novel pathologicechanisms and their clinical relevance. However, the

imitations of this study were its retrospective nature and

TABLE 1. Differences in Visual Acuity Between Eyes With or WExternal Lim

Characteristic

Ey

No. eyes/patients

Men/women

Age (years)

Diabetes duration (years)

HbA1C (%)

Total cholesterol (mg/dL)

Hypertension, no. (%)

Visual acuity (logMAR)

DR type, no. (%)

Mild NPDR

Moderate NPDR

Severe NPDR

PDR

No. eyes with hyperreflective foci in inner retinal layers (%)

DR � diabetic retinopathy; logMAR � logarithm of minimal ang

proliferative diabetic retinopathy.

TABLE 2. Association Between Hyperreflective Foci in thMembrane Seen on Spectral-Dom

ELM (�)

Hyperreflective foci in outer retinal layers

Absent (no. eyes) 45

Present (no. eyes) 13

Visual acuity (logMAR) 0.153 � 0.207

ELM � external limiting membrane; logMAR � logarithm of the m

mall sample size.

AMERICAN JOURNAL OF714

Bolz and associates previously suggested that hyperre-ective foci may represent subclinical features of lipopro-ein extravasation that act as precursors of hard exudates.19

Those authors pointed out that well-demarcated hyperre-flective foci were seen on Spectralis OCT scans in patientswith DME and that these foci were distributed throughoutall retinal layers. Our current study showed that hyperre-flective foci were present in the fovea in most eyes withDME, whereas only 58 eyes (53.7%) had hyperreflectivefoci in the outer retinal layers of the ELM. As mentionedpreviously, the ELM corresponds to the adherens junctionsbetween the Müller cells and photoreceptor cells and actsas a barrier against macromolecules. Thus, we hypothesizedthat the physiologic ELM restricts migration of lipopro-teins or proteins extravasated from the retinal vasculature,which is in the inner retinal layers. Disruption of the ELMmight permit these macromolecules to pass through thisbarrier and be deposited into the outer retinal layers.Another possible explanation is that hyperreflective fociresult from a neurodegenerative process; Baumüller and

ut Hyperreflective Foci in the Outer Retinal Layers outside theMembrane

h Hyperreflective Foci in

ter Retinal Layers

Eyes Without Hyperreflective Foci in

Outer Retinal Layers P Value

58/35 50/41 —

28/30 27/23 .5530

65.7 � 8.2 64.3 � 10.1 .1330

16.9 � 8.8 17.6 � 9.7 .0686

7.22 � 1.39 7.26 � 1.09 .8892

213 � 44 198 � 35 .1377

28 (48.3) 24 (48.0) .9772

.463 � 0.382 0.127 � 0.206 �.0001

0 (0.0) 1 (2.0) —

19 (32.8) 18 (36.0) —

20 (34.5) 18 (36.0) —

19 (32.8) 13 (26.0) .6428

58 (100.0) 49 (98.0) .2792

resolution; NPDR � nonproliferative diabetic retinopathy; PDR �

ter Retinal Layers and the Status of the External LimitingOptical Coherence Tomography

ELM (�) ELM (�) P Value

5 0 —

39 6 �.0001

0.416 � 0.370 1.008 � 0.229 �.0001

al angle of resolution.

ithoiting

es Wit

Ou

0

le of

e Ouain

associates previously documented outer retinal hyperreflec-

OPHTHALMOLOGY APRIL 2012

siBIoRrp

cssiepflpmw

issEMOcaj

plltrotHirtd

cocaTOawFbptaHs

tive spots in macular telangiectasia type 2.27 In the currenttudy, we found that the ELM lines merged with the RPEn about half of the eyes in the ELM (�) group (Figure 2,ottom row). In those cases, the absence of the ELM and

S/OS lines in those areas might represent cellular damager death, whereas the hyperreflective foci and irregularPE lines detected at a high frequency in those areas may

epresent degenerated retinal cells themselves or macro-hages attempting to engulf the cells.27,28

Increasing amounts of hard exudates are associated witha risk of visual impairment in patients with DR, anddeposition of submacular hard exudates commonly causes asevere decrease in VA.20,29,30 Intriguingly, Ota and asso-iates reported that hyperreflective dots in the subretinalpaces were significantly associated with deposition ofubfoveal hard exudates after resolution of SRD.20 Assum-ng that hyperreflective foci represent subclinical lipidxtravasation as a precursor of hard exudates and canermeate into the area under the ELM line, the hyperre-ective foci themselves are thought to disrupt the fovealhotoreceptor layer and the photoreceptor degenerationay already have begun by the time this process was under

TABLE 3. Damaged Junction Between the Inner and Outer SLimiting Membrane and Hyperrefl

IS/OS (�)

Visual acuity (logMAR) 0.060 � 0.182

ELM (�) 27

ELM (�) 0

ELM (�) 0

Hyperreflective foci in outer retinal layers

Absent (no. eyes) 26

Present (no. eyes) 1

ELM � external limiting membrane; IS/OS � the junction between

of minimal angle of resolution.

TABLE 4. Correlation Between Foveal Thickness and theExternal Limiting Membrane and Hyperreflective Foci in

the Outer Retina

Foveal Thickness (�m) P Value

Visual acuity (logMAR) 404 � 115 .0038; R � 0.275

ELM (�) 370 � 74 —

ELM (�) 441 � 132 —

ELM (�) 461 � 207 .0033

Hyperreflective foci in

outer retinal layers

Absent 373 � 78 —

Present 431 � 135 .0081

ELM � external limiting membrane; logMAR � logarithm of

minimal angle of resolution.

ay, when hard exudates are clinically invisible. t

HYPERREFLECTIVE FOCI, PHOTORECEPTOR LAYER,VOL. 153, NO. 4

The significant association between hyperreflective focin the outer retinal layers and a disrupted ELM suggestseveral possibilities: a cause-result relationship, a temporalequence, and an epiphenomenon. Considering that theLM corresponds to the adherens junctions between theüller cells and photoreceptors, disruption of the ELM onCT images might represent degeneration of the Müller

ells, photoreceptors, or both, or alternatively, disappear-nce or disorganization of the heterophilic adherensunctions.

As discussed previously, the breakdown of the barrierroperties of the ELM might permit macromolecules oripid-laden macrophages to migrate into the outer retinalayers, which agrees with the results of a previous reporthat documented outer retinal discontinuity in eyes withetinal vein occlusion.31 Further, extravasated lipoproteinsr proteins might contribute to degeneration of the pho-oreceptor cells or decreased junctional protein expression.yperreflective foci in the outer retinal layers beneath an

ntact ELM especially might represent degenerated photo-eceptor cells and ultimately result in a fragmentation ofhe ELM. A future longitudinal study should elucidate theetailed processes in this pathologic interaction.Damage to the IS/OS line and the ELM also was

orrelated with the presence of hyperreflective foci in theuter retinal layers. The outer segments of the photore-eptor cells are the major components of light perception,nd their metabolism is regulated in the inner segments.hus, structural disturbance of the IS/OS lines on theCT images might represent more directly the function-

lly impaired light perception, which might correspondell with the fragmented IS/OS lines on OCT images.urther, the extravasated macromolecules might lead tooth photoreceptor disorganization and deposition of hy-erreflective foci. These pathogeneses might begin afterhe ELM is disrupted, which agrees with the significantssociation between the status of the ELM and IS/OS lines.owever, the appearance of the IS/OS lines differed in

everal eyes (ie, merging of the ELM or IS/OS lines with

nts of the Photoreceptors in Eyes With a Disrupted ExternalFoci in the Outer Retinal Layers

IS/OS (�) IS/OS (�) P Value

0.299 � 0.294 0.733 � 0.385 �.0001

31 0 —

33 11 —

0 6 �.0001

23 1 —

41 16 �.0001

ner and outer segments of the photoreceptors; logMAR � logarithm

egmeective

the in

he RPE) and represented the absence of photoreceptor

AND VISUAL ACUITY IN DIABETIC MACULAR EDEMA 715

coh

rbcpwfcsv

hbitd

cells in that area (Figure 2, Bottom row). Hyperreflectivefoci might correspond to degenerated photoreceptors orthe macrophages engulfing them. Irregular RPE lines weresometimes seen on the OCT images in the current studyand may indicate macrophages on the RPE line, activatedor overphagocytosed RPE cells, or RPE metaplasia.22,30,32

Considering that there was a higher frequency of inter-rupted or absent IS/OS lines than of similar ELM lines, thepathologic interaction between degenerated photorecep-tors and RPE cells might result in the precedence ofdisrupted IS/OS (Figures 2 and 3).

Macular thickness is a well-accepted parameter asso-ciated with visual disturbance in DME.5–11 In theurrent study, maculae with hyperreflective foci in theuter retinal layers were thicker than those without

FIGURE 3. The integrity of the external limiting membrane (E(Top) In eyes without hyperreflective foci in the outer retisegment/outer segment (IS/OS) (black arrowhead) are frequentand the IS/OS (black arrowhead) status is sometimes seen on

yperreflective foci in the outer retinal layers. Thus, it is d

126(12):1740 –1747.

AMERICAN JOURNAL OF716

easonable that severe breakdown of the blood-retinalarrier might result in thickening of the retinal paren-hyma and extravasation of macromolecules or macro-hages. Interestingly, the ELM status also was associatedith the macular thickness, which led us to speculate

urther that degeneration or dysfunction of the Müllerells may cause the ELM lines to disappear, leading totructural fragility and a concomitant increase in theolume of the retinal parenchyma.The current study showed the clinical relevance of the

yperreflective foci, which are otherwise clinically invisi-le, in the outer retinal layers on the Spectralis OCTmages. We also identified a pathologic association be-ween hyperreflective foci in the outer retinal layers andisrupted ELM or IS/OS lines, suggesting photoreceptor

in eyes without hyperreflective foci in the outer retinal layers.ayers, the ELM (black arrow) and the photoreceptor innertact. (Bottom) A discrepancy between the ELM (black arrow)ptical coherence tomography images.

egeneration in DME.

THE AUTHORS INDICATE NO FUNDING SUPPORT OR FINANCIAL CONFLICTS OF INTEREST. INVOLVED IN DESIGN OF THEstudy (A.U., T.M., K.N., N.Y.); conduct of the study (A.U., T.M., K.N., T.A., T.H., N.A., Y.M., A.A.E., N.Y.); collection of data (A.U., T.M.,K.N., T.A., T.H., N.A., Y.M., A.A.E., N.Y.); interpretation of data (A.U., T.M., K.N., N.Y.); statistical analysis of data (A.U., T.M., K.N., N.Y.);and preparation, review, or approval of the manuscript (A.U, T.M., K.N, T.A., T.H., N.A., Y.M., A.A.E., N.Y.). All research and measurementsadhered to the tenets of the Declaration of Helsinki and were approved by the institutional review board at Kyoto University Graduate Schoolof Medicine. Each patient provided informed consent after a detailed explanation of the nature and possible consequences of the study procedures.

REFERENCES

1. Klein R, Klein BE, Moss SE, Cruickshanks KJ. The Wiscon-sin Epidemiologic Study of Diabetic Retinopathy. XV. Thelong-term incidence of macular edema. Ophthalmology1995;102(1):7–16.

2. Saaddine JB, Honeycutt AA, Narayan KM, Zhang X, KleinR, Boyle JP. Projection of diabetic retinopathy and othermajor eye diseases among people with diabetes mellitus:United States, 2005-2050. Arch Ophthalmol 2008;

3. Otani T, Kishi S, Maruyama Y. Patterns of diabetic macularedema with optical coherence tomography. Am J Ophthal-mol 1999;127(6):688–693.

4. Otani T, Kishi S. Tomographic findings of foveal hardexudates in diabetic macular edema. Am J Ophthalmol2001;131(1):50–54.

5. Hee MR, Puliafito CA, Wong C, et al. Quantitative assess-ment of macular edema with optical coherence tomography.Arch Ophthalmol 1995;113(8):1019–1029.

6. Browning DJ, Glassman AR, Aiello LP, et al. Relationship

LM)nal lly inthe o

between optical coherence tomography-measured central

OPHTHALMOLOGY APRIL 2012

2

2

2

2

2

2

2

2

2

2

3

3

3

retinal thickness and visual acuity in diabetic macularedema. Ophthalmology 2007;114(3):525–536.

7. Bandello F, Polito A, Del Borrello M, Zemella N, Isola M.“Light” versus “classic” laser treatment for clinically signifi-cant diabetic macular oedema. Br J Ophthalmol 2005;89(7):864–870.

8. Martidis A, Duker JS, Greenberg PB, et al. Intravitrealtriamcinolone for refractory diabetic macular edema. Oph-thalmology 2002;109(5):920–927.

9. Laursen ML, Moeller F, Sander B, Sjoelie AK. Subthresholdmicropulse diode laser treatment in diabetic macular oe-dema. Br J Ophthalmol 2004;88(9):1173–1179.

10. Ozdemir H, Karacorlu M, Karacorlu SA. Regression of serousmacular detachment after intravitreal triamcinolone ace-tonide in patients with diabetic macular edema. Am JOphthalmol 2005;140(2):251–255.

11. Massin P, Duguid G, Erginay A, Haouchine B, Gaudric A.Optical coherence tomography for evaluating diabetic mac-ular edema before and after vitrectomy. Am J Ophthalmol2003;135(2):169–177.

12. Alasil T, Keane PA, Updike JF, et al. Relationship betweenoptical coherence tomography retinal parameters and visualacuity in diabetic macular edema. Ophthalmology 2010;117(12):2379–2386.

13. Maheshwary AS, Oster SF, Yuson RM, Cheng L, Mojana F,Freeman WR. The association between percent disruption ofthe photoreceptor inner segment-outer segment junction andvisual acuity in diabetic macular edema. Am J Ophthalmol2010;150(1):63–67 e61.

14. Marmor MF. Mechanisms of fluid accumulation in retinaledema. Doc Ophthalmol 1999;97(3-4):239–249.

15. Forooghian F, Stetson PF, Meyer SA, et al. Relationshipbetween photoreceptor outer segment length and visualacuity in diabetic macular edema. Retina 2010;30(1):63–70.

16. Sakamoto A, Nishijima K, Kita M, Oh H, Tsujikawa A,Yoshimura N. Association between foveal photoreceptorstatus and visual acuity after resolution of diabetic macularedema by pars plana vitrectomy. Graefes Arch Clin ExpOphthalmol 2009;247(10):1325–1330.

17. Otani T, Yamaguchi Y, Kishi S. Correlation between visualacuity and foveal microstructural changes in diabetic macularedema. Retina 2010;30(5):774–780.

18. Kameda T, Nishijima K, Unoki N, et al. Geographic patternof central retinal sensitivity after intravitreal triamcinolonefor diabetic macular edema. Graefes Arch Clin Exp Oph-thalmol 2011;249(1):3–9.

19. Bolz M, Schmidt-Erfurth U, Deak G, Mylonas G, Kriech-baum K, Scholda C. Optical coherence tomographic hyper-reflective foci: a morphologic sign of lipid extravasation indiabetic macular edema. Ophthalmology 2009;116(5):914–

HYPERREFLECTIVE FOCI, PHOTORECEPTOR LAYER,VOL. 153, NO. 4

0. Ota M, Nishijima K, Sakamoto A, et al. Optical coherencetomographic evaluation of foveal hard exudates in patientswith diabetic maculopathy accompanying macular detach-ment. Ophthalmology 2010;117(10):1996–2002.

1. Fong DS, Segal PP, Myers F, Ferris FL, Hubbard LD, DavisMD. Subretinal fibrosis in diabetic macular edema. ETDRSreport 23. Early Treatment Diabetic Retinopathy StudyResearch Group. Arch Ophthalmol 1997;115(7):873–877.

2. Sigurdsson R, Begg IS. Organised macular plaques in exuda-tive diabetic maculopathy. Br J Ophthalmol 1980;64(6):392–397.

3. Murakami T, Nishijima K, Sakamoto A, Ota M, Horii T,Yoshimura N. Foveal cystoid spaces are associated withenlarged foveal avascular zone and microaneurysms in dia-betic macular edema. Ophthalmology 2011;118(2):359–367.

4. Yamaike N, Tsujikawa A, Ota M, et al. Three-dimensionalimaging of cystoid macular edema in retinal vein occlusion.Ophthalmology 2008;115(2):355–362 e352.

5. Ota M, Tsujikawa A, Murakami T, et al. Foveal photorecep-tor layer in eyes with persistent cystoid macular edemaassociated with branch retinal vein occlusion. Am J Oph-thalmol 2008;145(2):273–280.

6. Murakami T, Tsujikawa A, Ohta M, et al. Photoreceptorstatus after resolved macular edema in branch retinal veinocclusion treated with tissue plasminogen activator. Am JOphthalmol 2007;143(1):171–173.

7. Baumüller S, Charbel Issa P, Scholl HP, Schmitz-Valcken-berg S, Holz FG. Outer retinal hyperreflective spots onspectral-domain optical coherence tomography in maculartelangiectasia type 2. Ophthalmology 2010;117(11):2162–2168.

8. Schuman SG, Koreishi AF, Farsiu S, Jung SH, Izatt JA, TothCA. Photoreceptor layer thinning over drusen in eyes withage-related macular degeneration imaged in vivo with spec-tral-domain optical coherence tomography. Ophthalmology2009;116(3):488–496 e482.

9. Takaya K, Suzuki Y, Mizutani H, Sakuraba T, Nakazawa M.Long-term results of vitrectomy for removal of submacularhard exudates in patients with diabetic maculopathy. Retina2004;24(1):23–29.

0. Takagi H, Otani A, Kiryu J, Ogura Y. New surgical approachfor removing massive foveal hard exudates in diabetic mac-ular edema. Ophthalmology 1999;106(2):249–256; discus-sion 256-247.

1. Tsujikawa A, Sakamoto A, Ota M, et al. Serous retinaldetachment associated with retinal vein occlusion. Am JOphthalmol 2010;149(2):291–301 e295.

2. Murata T, Ishibashi T, Inomata H. Immunohistochemicaldetection of extravasated fibrinogen (fibrin) in human dia-betic retina. Graefes Arch Clin Exp Ophthalmol 1992;

AND VISUAL ACUITY IN DIABETIC MACULAR EDEMA 717

Biosketch

Akihito Uji, MD, PhD, is an Assistant Professor of Ophthalmology at Kyoto University, Kyoto, Japan. He received hisMD and PhD degrees from Okayama University Graduate School of Medicine, Okayama, Japan. His interests are diabeticretinopathy and image analysis studies.

HYPERREFLECTIVE FOCI, PHOTORECEPTOR LAYER, AND VISUAL ACUITY IN DIABETIC MACULAR EDEMAVOL. 153, NO. 4 717.e1