Reticular Pseudodrusen Are Subretinal Drusenoid Deposits

Reticular Pseudodrusen Are SubretinalDrusenoid Deposits

Sandrine A. Zweifel, MD,1,2 Richard F. Spaide, MD,2 Christine A. Curcio, PhD,3 Goldis Malek, PhD,4

Yutaka Imamura, MD1,2

Purpose: To characterize reticular pseudodrusen, a potential risk factor for late age-related maculardegeneration.

Design: Retrospective, observational case series.Participants: Fifty-eight eyes of 33 patients with pseudodrusen (20 female).Methods: Consecutive patients with reticular pseudodrusen, diagnosed by their typical appearance and

distribution using ophthalmoscopy, the blue channel of color fundus photographs, and near infrared images. Thepatients were imaged by spectral domain optical coherence tomography (SD OCT), and correlations were madebetween the near infrared images and the SD OCT images. The SD OCT findings in patients with pseudodrusenwere compared with previously reported histologic findings of subretinal drusenoid deposits. The histologicspecimens were reevaluated with the additional knowledge of the clinical information.

Main Outcome Measures: Spectral domain optical coherence tomography and histologic characteristics ofpseudodrusen.

Results: The mean age of the 33 patients was 81.7 years. The correlating SD OCT scans showed collections ofgranular hyperreflective material above the retinal pigment epithelium (RPE), in the subretinal space located primarilybetween the RPE and the boundary between the inner and outer segments of the photoreceptors (IS/OS boundary).In a more advanced stage, this material formed small mounds that broke through the IS/OS boundary. There were nocorrelates to the deposits seen under the RPE or in the choroid. These findings were similar in character to previouslyreported histologic characterization of subretinal drusenoid deposits, which had identified the presence of membra-nous debris, unesterified cholesterol, and complement within the deposits.

Conclusions: Pseudodrusen seen by clinical examination may be subretinal drusenoid deposits seen byhistologic examination. This unexpected location suggests that potential pathophysiologic mechanisms on both sidesof the RPE need to be taken into account in theories related to the development of age-related macular degeneration.

Financial Disclosure(s): Proprietary or commercial disclosure may be found after the references.Ophthalmology 2010;117:303–312 © 2010 by the American Academy of Ophthalmology.

Reticular pseudodrusen were first described by Mimoun etal1 in 1990 as a peculiar yellowish pattern in the fundus ofpatients with age-related macular degeneration (AMD).They named the structures “les pseudo-drusen visibles enlumière bleue” because of their enhanced visibility whenviewed using blue light. Arnold et al2 stated the reticularpseudodrusen were easily visualized with a scanning laserophthalmoscope, and seemed to be associated with lateAMD. They further postulated that the appearance arosefrom structures in the choroid. They performed histologicexamination in 1 eye, but only of the choroid, and did notfind any visible correlates to the pseudodrusen, although thepatient did have a markedly thin choroid.2 Descriptions andphotographs used to illustrate reticular pseudodrusen varysubstantially from publication to publication,2–9 withsome authorities classifying reticular pseudodrusen as beingno different from ordinary soft drusen material arranged ina reticular (derived from the Latin word “reticulum,” whichmeans net) pattern.8,9 This raises an important question

© 2010 by the American Academy of OphthalmologyPublished by Elsevier Inc.

The Heidelberg Spectralis (Heidelberg Engineering,Heidelberg, Germany) is capable of imaging the funduswith near infrared light using scanning laser ophthalmos-copy and can simultaneously perform spectral domainoptical coherence tomography (SD OCT) with point-to-point correlation between the 2 imaging modalities. Toinvestigate the characteristics of pseudodrusen, we per-formed a retrospective review of images from patientswith a reticular pseudodrusen appearance. Rudolf et al10

previously identified 3 eye bank eyes that appear to havedeposition consistent with what is described in the pa-tients in this series. The eyes were found to have sub-retinal deposition of drusenoid material that had manycharacteristics of soft drusen, which ordinarily are foundon the inner portion of Bruch’s membrane. At the time,this described material had no known clinical correlate.Later collaboration revealed that pseudodrusen seen bySD OCT corresponded to the material seen as subretinaldrusenoid deposits. This realization lead to additional

303ISSN 0161-6420/10/$–see front matterdoi:10.1016/j.ophtha.2009.07.014

Ophthalmology Volume 117, Number 2, February 2010

Materials and Methods

This was a retrospective study of consecutive patients with theclinical diagnosis of reticular pseudodrusen. These patients wereexamined in a private retinal referral practice in which nearly allexamined patients have OCT imaging, and all were seen in a2-month period starting at the beginning of January 2009. Thestudy had institutional review board approval through the WesternInstitutional Review Board and complied with the Health Insur-ance Portability and Accountability Act of 1996. Although thepresence of pseudodrusen can be readily suspected by simpleophthalmoscopy, the diagnosis of reticular pseudodrusen in thisstudy was based on the possible presence of 2 features: Thepseudodrusen were more easily identified in the blue channel ofthe color photograph or were evident in the near-infrared photo-graph taken with the scanning laser ophthalmoscope, the Heidel-berg Spectralis, or both. A reticular pattern, per se, was not acriterion used to establish the diagnosis. Patients who met thesecriteria independently of the underlying disease, such as neovas-cular AMD, dry AMD, branch retinal vein occlusion, and centralretinal vein occlusion, were included in this study.

The original description of pseudodrusen referred to their in-creased visibility in blue light. A functional equivalent is to look atthe blue channel of a color photograph, which by the additive colortheory contains the same information (for derivation see Appendix1, available at http://aaojournal.org). To examine the blue channelof the color photograph, high-resolution digital color fundus pho-tographs taken with a Topcon ImageNet camera (Topcon America,Paramus, NJ) were viewed in Topcon ImageNet (version 2.55,Topcon America) or Photoshop (Photoshop CS3, Adobe SystemInc., San Jose, CA). In the Topcon ImageNet program the com-mands Utilities�RGB channels were selected. The 3 principalchannels (red, green, and blue) comprising the color image werethen displayed along with the original color photograph. In Pho-

Figure 1. The top row shows schematic drawings with the stages of subreof the respective stage. The top grey line in the schematic representsinterdigitation of the photoreceptor outer segments and the apical procesdefined as diffuse deposition of granular hyperreflective material between t2 was considered to be mounds of accumulated material sufficient to alterside by side. In stage 3, the material adopted a conical appearance and brok

304

toshop, the “Channels Palette” was selected and the command“Split Channels” was used. The individual color channels werethen displayed individually.

The SD OCT scans of the eyes were obtained with the Heidel-berg Spectralis (version 1.6.1) as viewed with the contained Hei-delberg software (Spectralis Viewing Module 4.0.0.0; HeidelbergEngineering). The scanning protocol for the patients varied withthe underlying disease because the pseudodrusen were usually anincidental finding. The area of pathology in the posterior pole wasimaged using sections each comprising up to 100 averaged scans.The point-to point correlation feature of the Heidelberg Spectraliswas used to find corresponding pathology between the near-infraredimage and the SD OCT image. The pathology observed was scoredusing a defined grading system of 3 stages based on their cross-sectional appearance. Stage 1 was defined as diffuse deposition ofgranular hyperreflective material between the retinal pigment ep-ithelium (RPE) and the boundary between the inner segments (IS)and outer segments (OS) of the photoreceptors (the IS/OS bound-ary). Stage 2 was considered to be mounds of accumulated mate-rial sufficient to alter the contour of the IS/OS boundary. In stage3 the material was thicker, adopted a conical appearance, andbroke through the IS/OS boundary (Fig 1). To determine theproportion of each of the stages, drusen were sampled as follows.A table of random numbers was generated in Microsoft Excel(Microsoft Corporation, Redmond, WA) using the Rand() func-tion. These random numbers range from 0 to 0.999�. The scanline for each eye was selected at random. A reticule was con-structed that divided any given scan line into 10 regions labeled 0to 9. The most significant digit of the random number was read first,which indicated the region in the reticule in which the drusen were tobe counted. If the total number of drusen read was less than 10, thenext most non-repeating significant digit was used. If 5 significantdigits were used, a new scan line was selected at random and theprocess was continued until at least 10 drusen were counted.

drusenoid deposition. Along each column is 1 representative OCT image/OS boundary (A), the middle line represents the line formed by thethe RPE (B), and the bottom line represents the RPE (C). Stage 1 wasE and the boundary between the IS and OS of the photoreceptors. Stageontour of the IS/OS boundary. The representative case shows 3 moundsugh the IS/OS boundary. IS � inner segments; OCT � optical coherence

tinalthe ISses ofhe RPthe c

e thro

the whitish material in A. OCT � optical coherence tomography.

Zweifel et al � Subretinal Drusenoid Deposits

The method of obtaining OCT images from the choroid wasreported previously.11–13 Briefly, the choroid was imaged by po-sitioning the Heidelberg Spectralis instrument close enough to theeye to obtain an inverted image. Seven sections, each composed of100 averaged scans, were obtained in a 5�30-degree rectangleencompassing the macula and optic nerve. The horizontal sectiongoing directly through the center of the fovea was used for cho-roidal thickness measurements. By using the Heidelberg Eye Ex-plorer software, the choroid under the fovea was measured fromthe outer border of the hyperreflective line corresponding to theRPE to the inner scleral border. This type of OCT scan is acommon procedure in our office and is called “enhanced depthimaging OCT.”11–13 In addition to OCT, these patients had colorand autofluorescence imaging and angiography as dictated by theirunderlying retinal condition.

retinal pigment epithelium.

4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™Figure 2. A, The color photograph shows the macula of a 78-year-oldpatient with confluent soft drusen and a second type of drusenoid depositcomposed of a reticular array of yellow-white material. Admixed in thismaterial are more white punctuate deposits. The locations of the OCTslices are shown as green lines. B, A 400% magnification of the areacontained in the box in A shows ribbons of accumulated material withsmall white punctuate spots (arrows). C, Near infrared image in the sameregion as A. Note the greater prominence of the accumulations, whichappear as figures composed of dark material, many with a lighter center. D,A 400% enlargement of the same region as contained within the whitebox in A. Note the lighter centers in D correspond to the white punctuatespots seen in B. E, Autofluorescence image shows areas of decreasedautofluorescence corresponding to the deposits. F, Fluorescein angiogramshows staining of the soft drusen and very subtle staining of the reticulardrusenoid deposit. G, An OCT slice through the superior macula shows anearly confluent accumulation of material in the subretinal space withstage 2 (arrowheads) and stage 3 (white arrows) accumulations. H, An OCTslice through the fovea shows cross-sections of the soft drusen (blackarrows) and a subretinal deposition of material that is clearly differentiablefrom the underlying RPE. Some of this material shows stage 3 configura-tions (white arrows). OCT � optical coherence tomography; RPE �

Figure 3. A, Panoramic montage of an 87-year-old patient with subretinaldrusenoid deposits (the reticular whitish accumulations) extending over achoroidal nevus superior to the optic nerve. Green lines show the locations ofthe OCT slices (B, C). B, The nevus is evident as a hyperreflective structurewith altered choroidal vascular arrangement between the 2 arrows. Note thesubretinal drusenoid deposits. C, Numerous accumulations in the subretinalspace are seen in the superior macula. These accumulations co-localized with

305

Ophthalmology Volume 117, Number 2, February 2010

Early AMD was diagnosed if the patient had 1 or more softdrusen �125 �m or more than 5 drusen �63 and �125 �m orany focal hyperpigmentation, but did not have evidence of lateAMD.14 Geographic atrophy was considered present if therewas a hypoautofluorescent area, seen during autofluorescenceimaging, that exceeded one fourth of disc diameter in size anddid not overlie or abut a disciform scar. Choroidal neovascu-larization was considered present if the patient showed findingsconsistent with new vessel ingrowth during subsequent fluores-cein angiography. Late AMD was considered to be present ifthe patient had either geographic atrophy or choroidal neovas-cularization.

The histologic examination was done on an eye previouslyreported.10 The eye was sectioned, and 10 �m-thick cryosectionswere stained with hematoxylin. Serial cryosections, each approx-imately 11 mm long, were viewed with bright field and differentialinterference contrast microscopy to evaluate the distribution ofsubretinal material.

The data obtained were analyzed with frequency and descrip-tive statistics. Spearman’s rank correlation coefficient was used toevaluate the correlation between the number of pseudodrusen inthe right and left eyes. A P value less than 0.05 was considered tobe significant. The statistical analyses were performed with SPSSsoftware version 16.0 (SPSS, Inc., Chicago, IL).

Results

Our sample cohort included 58 eyes of 33 patients who had a meanage of 81.7 years (�5.9 years), and 20 patients (66.6%) werefemale. A history of current or past smoking was present in 12patients (36.4%), a history of diabetes was present in 4 patients(12.1%), and a history of hypertension was present in 22 patients(66.7%). Non-exudative AMD was present in 26 eyes (44.8%),with 5 eyes having geographic atrophy (8.6%). Choroidal neovas-cularization was present in 28 eyes (48.3%), and 4 eyes had neitherearly nor late AMD. A diagnosis of glaucoma was present in 14eyes (24.1%). The mean visual acuity of the 58 eyes was 20/49(logarithm of the minimum angle of resolution � 0.387, median20/40 [logarithm of the minimum angle of resolution � 0.3],interquartile range 20/30–20/60). The mean subfoveal choroidalthickness as measured using enhanced depth imaging OCT imageswas 149 �m (standard deviation [�] 86 �m). Because of the goodquality of the SD OCT scans in the remaining 10 eyes, it waspossible to measure their choroidal thickness as well, which was138 �m (�91 �m). Of the 58 eyes, 32 had a choroidal thicknessless than 125 �m, which given the age of the patients, met thediagnostic criteria for the diagnosis of age-related choroidal atro-phy.13 There did not appear to be any choroidal correlates to thesubretinal deposits.

The pseudodrusen as seen in the blue channel of the colorphotographs were uniformly seen in the infrared images. However,the infrared scanning laser ophthalmoscopic images revealed manymore pseudodrusen than did the blue channel of the color photo-graphs. These pseudodrusen as seen in the scanning laser ophthal-moscopic infrared images corresponded to specific OCT findings,as will be discussed later, so for grading purposes the distributionand number of pseudodrusen were determined from the infraredimages. Pseudodrusen were found predominantly in the superiormacula outside of the fovea (56% of all eyes within the outersuperior subfield, 38% within the inner superior subfield, 46% ofall eyes showed pseudodrusen outside the grid in the superiorquadrant). The pseudodrusen were seen to be hypoautofluorescentby fundus autofluorescence imaging.

By using point-to-point correlation of the Heidelberg Spec-

306

IS/OS boundary correlated to the pseudodrusen visualized inthe SD OCT scans (Figs 2– 6). This material was moderatelyreflective and had an appearance similar to the contents of softdrusen. Although this material was found in close proximity tothe RPE, it was possible to differentiate it from the RPEmonolayer. The subretinal drusen adopted a variety of appear-ances, with some appearing as basal laminar drusen or smallhard drusen on a background of early soft drusen (Fig 2) or softdrusen (Figs 3–5), whereas other eyes had no visible drusen(Fig 6). A total of 583 subretinal drusen were evaluated forgrading. The mean number of stage 1 drusen was 3.4 for theright eye and 3.1 for the left eye; the mean number of stage 2drusen was 4 for the right eye and 4.6 for the left eye; and themean number of stage 3 drusen was 1.2 for the right eye and 1.4for the left eye for the patients in this study. There was asignificant correlation between the mean drusen grade for theright and the left eyes (R � 0.401, P � 0.042). Some of thepseudodrusen in the color photographs appeared to have abright center, and this appearance generally correlated to stage3 drusen.

Histologic analysis came from an 86-year-old donor previouslydescribed by Rudolf et al10 (case 3). This eye exhibited pale spotsat the RPE level at gross examination and prominent subretinaldeposits, thin basal laminar deposits, and minimal RPE change atmicroscopic examination. The cryosections obtained from this eyewere stained with hematoxylin and viewed with bright field anddifferential interference contrast microscopy. The refractile mate-rial, as revealed by interference contrast microscopy, was absentfrom adjacent sections washed with acetone as part of processingfor immunofluorescence. The distribution of subretinal drusenoiddeposits (all �20 �m wide) were evaluated by examining 100cryosections that were obtained starting 200 �m inferior to thefovea.

Histologic images of focal subretinal lesions are shown inFigure 7. Subretinal aggregations at their largest encroached on theouter nuclear layer (Fig 7). Contents were flocculent (Fig 7A, C),and some contained refractile material near the apical cap asrevealed by differential interference contrast microscopy (Fig 7B,D). A reconstruction of 100 adjacent sections revealed that thesubretinal aggregations were actually linked together in sinuousbands (Fig 7E). Note the similarity of the material mapped inFigure 7E to that seen by color photography in different patients inFigures 2B, 3A, or 4A.

Discussion

The patients in this study had drusenoid collections ofmaterial that was easier to visualize in the blue channel offundus color photographs, consistent with the originaldescription of pseudodrusen.1 These patients were foundto have even more deposits by examining the near infra-red images obtained with a scanning laser ophthalmo-scope. The SD OCT image correlating to the depositsrevealed discrete collections of hyperreflective materialnot under but above the RPE, in the subretinal space. Theaccumulated material could be graded by the thickness ofthe accumulation. The thicker aggregates correspondedwith what appeared to be white punctate inclusions in thecolor photograph. The material appeared to be hypo-autofluorescent by fundus autofluorescence imaging, im-plying a paucity of retinoids.

A group of 3 eyes has recently been reported with

Zweifel et al � Subretinal Drusenoid Deposits

retinal space.10 The accumulated material in those eyes wasfound to have membranous debris, the principal constituentof soft drusen, as well as unesterified cholesterol, apolipro-tein E, complement factor H, and vitronectin, all of whichare common drusen components.15–19 The material did notshow opsin immunoreactivity, suggesting the material wasnot simply a condensate of shed outer segments.10 Furtherexamination of these eyes, as part of the current study,showed the subretinal material to be interconnected (Fig7E), suggestive of the reticular arrangement ascribed tothese collections. A reticular pattern was not apparent ongross examination of this eye, likely because blue light wasnot used. When viewed with differential interference con-trast microscopy, a contrast-enhancing method to probetissue on the basis of optical path length gradients to cross-polarized light, the subretinal material was seen to containor be closely associated with refractile material. It is tempt-ing to speculate the more white punctate nature of thesubretinal drusenoid deposits may somehow be related tothe refractile material.

Sarks et al20 and Bressler et al21 examined the evolution

Figure 4. A, The color photograph shows a lacy arrangement of subtle whas a green line. B, The accumulated material is evident in the blue chandark material. D, Late-phase fluorescein angiogram shows subtle variationthe subretinal space (stage 1, black arrows; stage 2, arrowhead; stage 3, whcoherence tomography; RPE � retinal pigment epithelium.

of geographic atrophy and described the accumulation of

subretinal membranous debris, which they speculated arosefrom photoreceptor outer segments that were not phagocy-tized. They also described the ophthalmoscopic appearanceof patients developing geographic atrophy and mentioned“small drusen-like dots” that were 25 to 50 �m in diameter.Careful examination of the photographs demonstratingthese dots show that they appear to be the same collectionsas examined in the present series. Arnold et al2 did not linkthe membranous debris and the small dots with pseudo-drusen, because in a subsequent description of the histopa-thology of one case of pseudodrusen they reported only thefindings present in the choroid. That article included sec-tions lacking neurosensory retina and consequently couldnot show the subretinal space. Although Arnold et al founda marked decrease in the thickness of the choroid, they didnot find any specific correlation between the choroid and thepseudodrusen appearance. They hypothesized that pseudo-drusen appearance may have been due to fibrotic replace-ment of the middle layer of the choroid.2 By using high-resolution OCT, in a method that images the choroid, wecould find no choroidal correlate to the pseudodrusen ap-

accumulations in a 78-year-old patient. The OCT slice location is shownthe color photograph. C, Near infrared imaging shows a denser array of

e background fluorescence. E, An OCT scan shows drusenoid deposits inrows). Note the lack of sub-RPE drusen in this section. OCT � optical

itishnel ofin thite ar

pearance. Point-to-point correlation established that the de-

307

Ophthalmology Volume 117, Number 2, February 2010

Figure 5. A, This 77-year-old patient had a history of severe visual loss in the right eye and a recent loss in the left eye. The patient had some cataractouschanges that degraded the color fundus photographic image. Note the small areas of hemorrhage. Several drusen are visible, but the pseudodrusen presentare very subtle. The locations of the OCT slices (E, F) are shown as green lines. B, The near infrared image shows dark opacities that correspond tosubretinal drusenoid deposits seen by OCT. The early (C) and late (D) phase fluorescein angiogram demonstrates 2 areas of predominantly classicchoroidal neovascularization. E, The OCT slice shows the choroidal neovascularization and subretinal drusenoid deposits. F, The lower OCT slice

308

Zweifel et al � Subretinal Drusenoid Deposits

posits of material accounting for the pseudodrusen appear-ance were in the subretinal space. On the basis of theclinical and histologic findings, we propose that the term“subretinal drusenoid deposit” is more appropriate than“pseudodrusen.”

The origin of this material is not known at present, but onthe basis of clinical and histologic observations, possiblemechanisms may be proposed. In AMD, altered polarizationof the RPE can be observed.10 To the extent that RPE andvectoral transport mediated by or related to the RPE exists,a misdirection of transport may occur into the subretinalspace. An alternate, but not necessarily mutually exclusive,possibility is related to recent theories of drusen formationin Bruch’s membrane. Under ordinary circumstances, thereis a vectoral flow of lipids and other metabolites necessaryfor RPE function. Hageman et al22 postulated that cellulardebris accumulated in Bruch’s membrane to stimulate achronic inflammatory stimulus, which may serve as a nu-

Figure 6. A, This 90-year-old patient has a history of choroidal neovasculachoroidal vessels, the peripapillary atrophy despite the lack of myopia, andB, The blue channel photograph shows these few drusen. C, The near infrasuperotemporal to the optic nerve. D, The deposits above the level ofcoherence tomography; RPE � retinal pigment epithelium.

cleation site for drusen formation. Within some drusen

are bulbous processes possibly originating from dendriticcells.22 These events in Bruch’s membrane have manyanalogues in the subretinal space. A large vectoral flow ofmaterials passes from the RPE to photoreceptors and backthrough the shed outer segments.23–25 Outer segments arephagocytized, and polyunsaturated fatty acids are recycledback to the photoreceptors. The subretinal space has micro-glial cells, which like the choroid-resident dendritic cells arebone marrow-derived and capable of inflammatory reac-tions.26,27 There is cell debris in the subretinal space, relatedto normal function of photoreceptors and their age-relatedloss, which happens preferentially to rods.28 CX3CR1�/�

mice exhibit lipid-bloated microglial cells in the subretinalspace that confer an appearance of drusen-like pale spots inthe fundus.29

Past studies have evaluated different types of drusenunder the rubric of “reticular pseudodrusen” or “reticulardrusen” and considered their contribution to risk for late

n and age-related choroidal atrophy. Note the ready visualization of largerscattered drusen. The location of the OCT slice is shown as a green line.

anning laser ophthalmoscopic image shows numerous dark punctuate spotsPE, corresponding to the dark dots in (C), are visible. OCT � optical

rizatioa few

red scthe R

AMD. Arnold et al2 studied a group of patients with sub-

309

er; R

Ophthalmology Volume 117, Number 2, February 2010

retinal drusenoid deposits, as did Cohen et al.7 Both groupsfound a high prevalence of late AMD among patients withpseudodrusen, but the association does not necessarily meanthe pseudodrusen (or more accurately, the subretinal druse-noid deposits) were actual risk factors for the developmentof the late AMD. Articles by other investigators examining“reticular drusen” evaluated ordinary soft drusen arrangedin a reticular pattern.8,9 For example, the Wisconsin Read-ing Center’s definition of reticular drusen (available at:http://eyephoto.ophth.wisc.edu/ResearchAreas/AREDS/CHAPTER15B.html#Drusen, accessed June 9, 2009), whichmentions Standard slide 10 (available at: http://eyephoto.ophth.wisc.edu/ResearchAreas/AREDS/AREDS_Stds/Astd10.htm, accessed on June 9, 2009), shows an imagewith a strong yellow cast. In this example only soft drusenare visible. As such, reported findings for reticular drusenprobably do not apply to subretinal drusenoid deposits.Given the high prevalence of AMD in the present series ofeyes with reticulated drusen and in similar previously re-

Figure 7. Histologic analysis of subretinal drusenoid debris; 10 �m-thick cand differential interference contrast (B, D) microscopy. Bar, 50 �m. Adrusenoid deposit (arrow) had a loosely packed flocculent base with dense ccap had refractile material (arrowhead). C, In a separate section, a subretilayer. D, There is refractile material on the upper surface of the depositevaluated by examining 100 cryosections, each approximately 11 mm in leby an X. Typical drusen (all �45 �m) are shown as black dots. The subreganglion cell layer; INL � inner nuclear layer; ONL � outer nuclear lay

ported articles,2,7 subretinal drusenoid deposits may well be

310

a risk factor for late AMD. Therefore, classifying them asbeing a component of early AMD may be appropriate. It islikely that because no epidemiologic study has had a clas-sification that included subretinal drusenoid deposits orpseudodrusen, and because the cases illustrated in this arti-cle show that subretinal drusenoid deposits could be con-fused with early soft drusen, hard drusen, or even basallaminar drusen, patients with subretinal drusenoid depositshave so far been largely misclassified.

Limitations

There are important limitations to this study. It was aretrospective study of patients with various underlying dis-eases, and the subretinal drusenoid deposits were usually anincidental finding. The patients were seen in a retinal referralpractice, and this may have introduced biases in the prevalenceof associated ocular conditions, such as late AMD. Currently,

ctions were stained with hematoxylin and viewed with bright field (A, C)e the accumulation of material above the RPE (arrow). This subretinalote the decrease in the number of overlying photoreceptor nuclei. B, The

rusenoid deposit (arrow) extended from the RPE up to the outer nuclearhe distribution of subretinal drusenoid deposits (all �20 �m wide) wasthat were obtained starting 200 �m inferior to the fovea, which is markeddrusenoid deposits formed an interconnected branching pattern. GCL �

PE � retinal pigment epithelium.

ryose, Notap. Nnal d. E, Tngth,tinal

there is no automated software to measure the height or num-

Zweifel et al � Subretinal Drusenoid Deposits

ber of subretinal drusenoid deposits as seen by OCT, andtherefore all measurements were performed manually. How-ever, the results of this study describe the clinical features ofpatients with drusenoid deposits above the RPE, in the sub-retinal space. Correlative results from eye bank eyes with whatappear to be similar findings are also presented. The presentpatient group may not represent all types of pseudodrusendescribed in the literature, and we therefore propose the term“subretinal drusenoid deposits” for patients with the findingsdescribed here. The term “pseudodrusen” is not correct in anycase and does not accurately convey the histologically deter-mined contents of the deposits. The risk posed by subretinaldrusenoid deposits needs to be appraised, and by extensionother types of drusen would need to be reappraised as well.Drusen classification using only color photographs has beenused in the past, but the current findings suggest color photo-graphs alone are not sufficient to accurately assess subretinaldrusenoid deposits or to differentiate them from soft drusen.

References

1. Mimoun G, Soubrane G, Coscas G. Macular drusen [inFrench]. J Fr Ophtalmol 1990;13:511–30.

2. Arnold JJ, Sarks SH, Killingsworth MC, Sarks JP. Reticularpseudodrusen: a risk factor in age-related maculopathy. Retina1995;15:183–91.

3. Arnold JJ, Quaranta M, Soubrane G, et al. Indocyanine greenangiography of drusen. Am J Ophthalmol 1997;124:344–56.

4. Prenner JL, Rosenblatt BJ, Tolentino MJ, et al, CNVPT Re-search Group. Risk factors for choroidal neovascularizationand vision loss in the Fellow Eye Study of CNVPT. Retina2003;23:307–14.

5. Einbock W, Moessner A, Schnurrbusch UE, FAM StudyGroup. Changes in fundus autofluorescence in patients withage-related maculopathy. Correlation to visual function: aprospective study. Graefes Arch Clin Exp Ophthalmol 2005;243:300–5.

6. Smith RT, Chan JK, Busuoic M, et al. Autofluorescencecharacteristics of early, atrophic, and high-risk fellow eyes inage-related macular degeneration. Invest Ophthalmol Vis Sci2006;47:5495–504.

7. Cohen SY, Dubois L, Tadayoni R, et al. Prevalence of retic-ular pseudodrusen in age-related macular degeneration withnewly diagnosed choroidal neovascularisation. Br J Ophthal-mol 2007;91:354–9.

8. Wang JJ, Rochtchina E, Lee AJ, et al. Ten-year incidence andprogression of age-related maculopathy: the Blue MountainsEye Study. Ophthalmology 2007;114:92–8.

9. Klein R, Meuer SM, Knudtson MD, et al. The epidemiology ofretinal reticular drusen. Am J Ophthalmol 2008;145:317–26.

10. Rudolf M, Malek G, Messinger JD, et al. Sub-retinal druse-noid deposits in human retina: organization and composition.Exp Eye Res 2008;87:402–08.

11. Spaide RF, Koizumi H, Pozonni MC. Enhanced depth imag-ing spectral-domain optical coherence tomography. Am JOphthalmol 2008;146:496–500.

12. Margolis R, Spaide RF. A pilot study of enhanced depthimaging optical coherence tomography of the choroid in nor-

13. Spaide RF. Age-related choroidal atrophy. Am J Ophthalmol2009;147:801–10.

14. Age-Related Eye Disease Study Research Group. A simplifiedseverity scale for age-related macular degeneration: AREDSreport no. 18. Arch Ophthalmol 2005;123:1570–4.

15. Malek G, Li CM, Guidry C, et al. Apolipoprotein B incholesterol-containing drusen and basal deposits of humaneyes with age-related maculopathy. Am J Pathol 2003;162:413–25.

16. Curcio CA, Presley JB, Malek G, et al. Esterified and unes-terified cholesterol in drusen and basal deposits of eyes withage-related maculopathy. Exp Eye Res 2005;81:731–41.

17. Mullins RF, Russell SR, Anderson DH, Hageman GS. Drusenassociated with aging and age-related macular degenerationcontain proteins common to extracellular deposits associatedwith atherosclerosis, elastosis, amyloidosis, and dense depositdisease. FASEB J 2000;14:835–46.

18. Li CM, Clark ME, Chimento MF, Curcio CA. Apolipoproteinlocalization in isolated drusen and retinal apolipoprotein geneexpression. Invest Ophthalmol Vis Sci 2006;47:3119–28.

19. Hageman GS, Mullins RF, Russell SR, et al. Vitronectin is aconstituent of ocular drusen and the vitronectin gene is ex-pressed in human retinal pigmented epithelial cells. FASEB J1999;13:477–84.

20. Sarks JP, Sarks SH, Killingsworth MC. Evolution of geo-graphic atrophy of the retinal pigment epithelium. Eye 1988;2:552–77.

21. Bressler SB, Bressler NM, Sarks SH, Sarks JP. Age-relatedmacular degeneration: nonneovascular early AMD, interme-diate AMD, and geographic atrophy. In: Ryan SJ, ed-in-chief,Hinton DR, Schachat AP, Wilkinson CP, eds. Retina. 4th ed.vol. 2. Philadelphia, PA: Elsevier Mosby; 2006:1041–74.

22. Hageman GS, Luthert PJ, Victor Chong NH, et al. An inte-grated hypothesis that considers drusen as biomarkers ofimmune-mediated processes at the RPE-Bruch’s membraneinterface in aging and age-related macular degeneration. ProgRetin Eye Res 2001;20:705–32.

23. Chen H, Anderson RE. Metabolism in frog retinal pigment epi-thelium of docosahexaenoic and arachidonic acids derived fromrod outer segment membranes. Exp Eye Res 1993;57:369–77.

24. Gordon WC, Rodriguez de Turco EB, Bazan NG. Retinalpigment epithelial cells play a central role in the conservationof docosahexaenoic acid by photoreceptor cells after sheddingand phagocytosis. Curr Eye Res 1992;11:73–83.

25. Rodriguez de Turco EB, Parkins N, Ershov AV, Bazan NG.Selective retinal pigment epithelial cell lipid metabolism andremodeling conserves photoreceptor docosahexaenoic acidfollowing phagocytosis. J Neurosci Res 1999;57:479–86.

26. Kezic J, McMenamin PG. Differential turnover rates ofmonocyte-derived cells in varied ocular tissue microenviron-ments. J Leukoc Biol 2008;84:721–9.

27. Xu H, Chen M, Mayer EJ, et al. Turnover of resident retinalmicroglia in the normal adult mouse. Glia 2007;55:1189–98.

28. Curcio CA, Medeiros NE, Millican CL. Photoreceptor loss inage-related macular degeneration. Invest Ophthalmol Vis Sci1996;37:1236–49.

29. Raoul W, Feumi C, Keller N, et al. Lipid-bloated subretinalmicroglial cells are at the origin of drusen appearance inCX3CR1-deficient mice. Ophthalmic Res 2008;40:115–9.

30. Jacques SL, McAuliffe DJ. The melanosome: threshold tem-perature for explosive vaporization and internal absorptioncoefficient during pulsed laser irradiation. Photochem Photo-

311

Ophthalmology Volume 117, Number 2, February 2010

Footnotes and Financial Disclosures

Originally received: May 2, 2009.Final revision: June 17, 2009.Accepted: July 7, 2009.Available online: October 7, 2009. Manuscript no. 2009-590.1 LuEsther T. Mertz Retinal Research Center, Manhattan Eye, Ear, andThroat Hospital, New York, New York.2 Vitreous-Retina-Macula Consultants of New York, New York.3 Department of Ophthalmology, Callahan Eye Foundation Hospital, Uni-versity of Alabama School of Medicine, Birmingham, Alabama.4 Department of Ophthalmology and Pathology, Duke University, Durham,North Carolina.

Financial Disclosure(s):The author(s) have made the following disclosure(s):

Dr. Spaide has been a consultant for Topcon America, Paramus, New

312

Research Center, Zuidland, The Netherlands; and receives researchsupport from Genentech, Inc., South San Francisco, California. Noother authors have any financial interests. The sponsor or fundingorganization had no role in the design or conduct of this research.

Supported in part by The Macula Foundation, Inc. Dr. Imamura was fundedby grants from Koureisha Ganshikkan Kenkyu Zaidan, Mishima Saiichi-kinen Gankakenkyu Kokusaikouryu Kikin, and Takeda Kagaku ShinkouZaidan.

The abstract was submitted to the upcoming American Academy of Oph-thalmology Annual Meeting, San Francisco, California, October 2009.

Correspondence:Richard F. Spaide, MD, 460 Park Avenue, 5th Floor, New York, NY

Zweifel et al � Subretinal Drusenoid Deposits

Appendix 1

Why are Pseudodrusen Easier to See with Blue Light?A possible explanation relies on known absorption spectraof melanosomes. The absorption coefficient increases withdecreasing wavelength for melanosomes in the RPE and isexpressed by the equation:30

�a � 6.49 � 1012 ⁄ �3.48 cm-1

The Beer-Lambert law is I � IOe��ax, where I is theintensity at any given depth � based on the incident inten-sity IO and the extinction coefficient �a. Since incidentintensity of the illuminating light and the path length foreach wavelength is constant the ratio of the amount of redlight to blue light transmitted through melanosomes, substi-tuting into the Beer-Lambert equation:

Ratio of transmitted light

Ired

Iblue� e

�x� 6.49 �1012

�red3.48

�6.49 �1012

�blue3.48 �

�e

x6.49 �1012� �blue3.48��red

3.48

�red3.48�blue

3.48 �

The choice for what is considered red or blue is some-what arbitrary, but by selecting the peak absorption curvefor the red and blue cone pigment spectra (using represen-tative values of 460 nm and 560nm, respectively), and if thecumulative thickness of the melanosomes is one tenth of theRPE cell thickness, which is 10–14 microns thick, the ratioof transmitted blue light is approximately 78–84% of thered light transmission. If the material in ordinary soft drusenand subretinal drusenoid deposits have the same reflectancespectra, then by the location of soft drusen under the retinalpigment epithelium they will appear more yellow. The lightreaching the soft drusen will have some of the blue lightattenuated by the filtering effect and the light returning willhave the blue light attenuated to a similar proportion by thedouble pass through the retinal pigment epithelium. Obser-vation of the drusen types with blue light, or equivalently bylooking at the blue channel of a color photograph will showthe disparity of blue light returning from each type ofdrusen. In the case of soft drusen, the absolute amount ofblue light returning will be reduced as compared with thatfrom subretinal drusenoid deposits by the amount of bluelight attenuation caused by 2 passes through the retinalpigment epithelium.

312.e1