Downloaded from http://bjo.bmj.com/ on August 18, 2017 - Published by group.bmj.com Clinical science Prevalence and quantification of geographic atrophy associated with newly diagnosed and treatment- naïve exudative age-related macular degeneration Anne Sikorav,1 Oudy Semoun,1 Sandrine Zweifel,2 Camille Jung,3 Mayer Srour,1 Giuseppe Querques,1,4 Eric H Souied1

10–14 ►► Additional material is ABSTRACT develop, causing further vision loss. Most published online only. To view Objective To identify and quantify geographic atrophy studies focus on either wet or dry form separately please visit the journal online (GA) associated with neovascular age-related macular and there is a paucity of reports on both forms of (http://​dx.​doi.​org/​10.​1136/​ bjophthalmol-​2015-​308065). degeneration (AMD) at initial presentation using a AMD concomitantly in the same eye. There are fundus autofluorescence (FAF) semi-automated software recent published data regarding the prevalence of – 1Department of Ophthalmology, and to correlate the results with demographic and GA complicating exudative AMD at diagnosis,14 16 University Paris Est, Centre clinical data. with diverse results. Hospitalier Intercommunal de Creteil, Creteil, France Design Retrospective, observational study. In both atrophic and exudative AMDs, RPE 2Department of Ophthalmology, Methods The study population consisted of treatment- damage may be an important factor for visual loss. University Hospital Zurich, naïve patients with newly diagnosed neovascular AMD. The status of RPE can be studied by blue light Zurich, Switzerland Best-corrected visual acuity, fundus photographs, infrared fundus autofluorescence (FAF) imaging, which is a 3 Clinical Research Center, reflectance, FAF and spectral-domain optical coherence non-invasive imaging method that allows for topo- University Paris Est fl Creteil, Centre Hospitalier tomography were performed, associated with uorescein graphic mapping of lipofuscin distribution in the 17 18 intercommunal de Creteil, and indocyanine green angiographies. Identification of RPE cell monolayer in vivo. Various patterns Creteil, France GA was independently performed by three readers. of FAF in different stages of AMD have been 4 – Department of Ophthalmology, Quantification of atrophy areas was done using reported.19 22 Due to the absence of lipofuscin, University Vita Salute, San Raffaele Scientific Institute, RegionFinder Software (RFA), a semi-automated software atrophic areas in eyes with GA have a reduced Milan, Italy embedded in Spectralis device (Heidelberg Engineering, signal. The increased autofluorescence intensity in Germany). the perilesional zone of atrophy helps to predict 23 24 Correspondence to Results We included 206 eyes of 173 consecutive future atrophy progression. Using Region Dr Oudy Semoun, Department patients (72% female, mean age: 79.7±9.1 years). Type Finder Analyser (RegionFinder Software (RFA), of Ophthalmology, Centre I choroidal neovascularisation (CNV) was observed in Heidelberg Engineering, Heidelberg, Germany), a Hospitalier Intercommunal de Creteil, 40, avenue de 44.2% of eyes, type II CNV was observed in 20.9% and novel semi-automated customised software based Verdun, Creteil 94010, France; ​ mixed CNV lesion was observed in 11.7%. Polypoidal on FAF imaging and embedded in Spectralis oudysemoun@​hotmail.​com choroidal vasculopathy was diagnosed in 7.7% and type (Spectralis HRA+OCT; Heidelberg Engineering, III CNV was diagnosed in 15.5%. Analysis of FAF frames Heidelberg, Germany) atrophic patches can be The present research has showed that GA was associated with nAMD in 46/206 quantified and the spread of the total size can be been presented in ARVO 2014 25 26 annual meeting as a poster: eyes (22.3%). Taking into account data both from determined over time. “Incidence and quantification of Region Finder and multimodal imaging, our results In this retrospective observational case series, our geographic atrophy associated suggest that GA was present in 24.3% of eyes newly purpose was to determine the prevalence and to with neovascular age related diagnosed with exudative AMD. Mean size of GA was quantify GA complicating exudative AMD at first macular degeneration at initial 2 – presentation” (Poster number: 1.23±1.76 mm (range 0.03 7.39). presentation, using a multimodal analysis including D0050, programme number: Conclusion GA is associated with nAMD in 1/4 of RFA. 4940). cases at initial presentation. Combined imaging, including RFA is an effective tool to identify and quantify METHODS Received 2 November 2015 GA at diagnosis. We reviewed the charts of consecutive treatment- Revised 13 July 2016 Accepted 14 July 2016 naïve patients with exudative AMD, who presented Published Online First at University Eye clinic of Creteil, France, between 8 August 2016 INTRODUCTION August 2012 and August 2013. French Society of Age-related macular degeneration (AMD) is the Ophthalmology Ethics Committee approval was leading cause of blindness in patients aged over obtained for the retrospective review of the data. – 65 years in industrialised countries.1 3 The first Both eyes could be included in case of bilateral visible alterations of AMD are drusen and retinal exudative AMD. pigment epithelial (RPE) irregularities. This early Exclusion criteria included age under 50 years, stage of age-related maculopathy (ARM) may pro- any previous treatment for exudative AMD (such as gress to either geographic atrophy (GA) or exuda- laser photocoagulation, photodynamic therapy, tive AMD. The prevalence of GA is 3.5% in intravitreal injections of steroids or anti-vascular subjects older than 75 years and accounts for endothelial growth factor (VEGF)), myopia >−6D approximately half the prevalence of exudative (spherical equivalent), active intraocular inflamma- 4–6 To cite: Sikorav A, AMD. Approximately 11% of patients with atro- tion or any other retinopathy (such as hereditary Semoun O, Zweifel S, phic AMD eventually progress to the exudative retinal dystrophy, diabetic retinopathy, retinal – et al. Br J Ophthalmol form at 4 years of evolution.7 9 In some eyes with venous occlusion or epiretinal membrane) and 2017;101:438–444. choroidal neovascularisation (CNV), GA may also CNV attributable to other causes than AMD.

438 Sikorav A, et al. Br J Ophthalmol 2017;101:438–444. doi:10.1136/bjophthalmol-2015-308065 Downloaded from http://bjo.bmj.com/ on August 18, 2017 - Published by group.bmj.com Clinical science

Missing information about FAF or media opacities or fundus atrophy. Measures of atrophy using RFA were performed by haemorrhages that prevented adequate imaging of FAF were not three independent retina specialists (AS, SZ and OS). The first considered as exclusion criteria in our study if additional two readers (AS and SZ) measured the size of atrophy together. imaging data (fluorescein angiography (FA) and spectral-domain Reader 3 (OS) had no insight in atrophy quantification of the optical coherence tomography (SD-OCT)) for grading of GA two previous readers and the values for reader 3 were all inde- were available and applicable. pendently measured. Grading of atrophy was evaluated as All patients underwent at baseline a detailed medical and follow: presence of atrophy on one imaging modality (FAF or ocular history, a complete ophthalmologic examination, which OCT or FA) with the readers unanimity (1), presence of atrophy included measurement of best-corrected visual acuity (BCVA) diagnosed by multimodal imaging with the readers unanimity using standard ETDRS charts, expressed as logarithm of the (2), presence of atrophy diagnosed by multimodal with the minimum of angle of resolution (logMAR),27 slit-lamp biomicro- readers majority (3), absence of atrophy, diagnosed by multi- scopy, intraocular pressure assessment, fundus biomicroscopy. modal imaging, with the readers majority (4), and grading not Retinal imaging was performed with a confocal scanning laser applicable (5). ophthalmoscope system (Spectralis HRA+OCT, Heidelberg Engineering, Germany) and included the acquisition of infrared Statistics reflectance (IRR 820 nm), FAF (excitation 488 nm, emission Qualitative variables were described in percentages and quantita- 500–700 nm), FA, indocyanine green angiography (ICGA) and tive variables were described by their mean with SD. SD-OCT. Images were recorded with a minimum of resolution Comparisons of qualitative variables were performed using the of 768×768 pixels. The field of view was set 30°×30° and χ2 or Fisher’s exact tests. Comparisons of medians were per- centred on the macula, defined as a 6 mm diameter area around formed using the Kruskal-Wallis or Mann-Whitney tests. the fovea. Univariate and multivariate analyses with logistic regression A modified ETDRS grid, with circle diameters of 1200, 3600 were used to determine factors associated with atrophy at initial and 7200 mm (foveal, juxtafoveal and extrafoveal, respectively), presentation. In case of atrophy, predictive factors of visual was placed on the foveal centre using the Spectralis software. acuity and size of atrophy were determined using univariate and CNV lesion type (occult or type 1, classic or type 2, mixed, multivariate analyses with linear regression. Variables analysed retinal angiomatous proliferation (RAP) or type 3 and polypoi- were the following: age, sex, visual acuity, lens status, vitamin dal choroidal vasculopathy (PCV)) and their localisation using supplementation, CNV type, size and localisation of atrophy the ETDRS grid (subfoveal, juxtafoveal, extrafoveal) were first and fellow eye status. Variables with p value <0.20 were initially classified. CNV could be localised in one or more foveal areas. introduced in the model and only variable with a significant The CNV classification was performed as a part of the retro- p value was taken into account for accuracy. spective chart review. Reader 1 (AS, junior physician) and reader 2 (SZ, senior phys- Graders were required to indicate if there were signs of GA at ician) reviewed all images. Reader 3 (OS, senior physician) per- the initial visit evaluating multimodal imaging including FAF, formed a control reading in case of disagreement on the colour photographs, SD-OCT, FA and ICGA. GA was consid- presence of atrophy. The first two readers (AS and SZ) measured ered present if there were hypoautofluorescent lesion on the the size of atrophy together. These measurements were com- FAF,22 correlating thinning or loss of the RPE hyper-reflective pared with the measurement by reader 3 (OS), which was per- line and of the outer and/or internal retinal layers on the formed independently. SD-OCT, associated with choroidal signal enhancement on The inter-rater agreement was assessed by Bland and Altman SD-OCT.28 The corresponding pathology in the FA had to be a graph and the calculation of the intraclass correlation coefficient sharply delineated hyperfluorescent lesion without diffusion (ICC). Cohen’s κ-coefficient was calculated to measure the inde- consistent with a window defect. Multimodal imaging allowed pendent expert’s agreement for the diagnosis and classification distinguishing whether hypoautofluorescent lesions were due to of CNV. The chosen level of statistical significance was p<0.05. blocking effect by xanthophyll pigments, exudates, haemor- Statistical analyses were performed using STATA (V.13)statistical rhages or due to atrophy.29 software (StataCorp, College Station, Texas, USA). Image quality of FAF was analysed by two retina specialists (AS and SZ) as follows: missing FAF image (0), FAF image present and reliable grading (1), poor FAF image due to haem- RESULTS orrhage but reliable grading still possible (2), poor image quality Demographics and clinical data of the studied population of FAF image with limited grading possibilities (3). Only We included 206 eyes of 173 consecutive patients (72% female, hypoautofluorescent areas between the upper and lower tem- mean age: 79.7±9.1 years). Both eyes were included in 33 poral vascular arcades were taken into account, as well as peri- cases. We counted 91/206 eyes (44.2%) presenting with occult papillary atrophy in case of a peripapillary CNV lesion; CNV; 43/206 eyes (20.9%) with classic CNV and 24/206 eyes localisations of atrophic areas were also investigated using the (11.7%) with mixed CNV. PCV were diagnosed in 16/206 eyes ETDRS grid (foveal, juxtafoveal, extrafoveal). (7.7%) and chorioretinal anastomosis in 32/206 eyes (15.5%). A Presence of macular atrophic areas was also assessed in the total of 126/206 eyes (61.2% of cases) had subfoveal CNV, 136/ fellow eyes, if data were available. 206 (66.0%) had juxtafoveal CNV and 76/206 (36.9%) had Atrophic areas were quantified using the RFA.30 Given the extrafoveal CNV. The inter-rater agreement was exact for classi- digital image resolution of 768×768 pixels for a 30°×30° fication of CNV (k=1.00) and reproducibility of grading for the picture, one pixel roughly corresponded to 12 mm. As described CNV localisation was similar (k=0.92). in our previous paper, the minimal lesion size was defined as an Mean initial visual acuity in studied eyes was 0.63±0.46 atrophic areas measuring 0.02 mm2, quantified using RFA.26 logMAR and mean initial contralateral visual acuity was 0.54 All images were reviewed by two independent retina specia- ±0.61 logMAR. A total of 52 patients (26.5%) were receiving lists (AS and SZ). A third reading was performed by a senior an oral vitamin supplementation (AREDS II supplement) at physician (OS) in case of disagreement on the presence of diagnosis.

Sikorav A, et al. Br J Ophthalmol 2017;101:438–444. doi:10.1136/bjophthalmol-2015-308065 439 Downloaded from http://bjo.bmj.com/ on August 18, 2017 - Published by group.bmj.com Clinical science

good: ICC=0.962 (95% CI (0.940 to 0.984)). The Bland - Table 1 Characteristics of patient population Altman graph is presented as online supplementary file. Parameter Quantity

No. of eyes 206 Risks factors associated with atrophy at baseline Univariate analysis found that initial BCVA was significantly No. of patients 173 Mean age (years) 79.7±9.1 lower when atrophy was present in eyes with exudative AMD Female/male 72%/28% (0.72 vs 0.57 logMAR, 20/105 vs 20/74 p=0.004). Presence of atrophy was more frequently observed in older patients (82.0 vs CNV type fi Type 1 91 (44.2%) 78.8 years, p=0.019). No signi cant difference was found Type 2 43 (20.9%) regarding gender (p=0.149), pseudophakic status (p=0.330) or Combined 24 (11.7%) vitamin supplementation (p=0.537) (table 4). Type 3 32 (15.5%) When atrophy was associated with wet AMD, fellow eyes Polypoidal choroidal vasculopathy 16 (7.7%) also presented more frequently with CNV (48%, p=0.004), CNV localisation with a mean lower visual acuity (0.67 vs 0.49 logMAR, 20/94 vs 20/62, p=0.019); nevertheless, incidence of GA or ARM in Subfoveolar 126 (61.2%) fi Juxtafoveolar 136 (66.0%) the fellow eye was not signi cantly increased (p=0.184 and Extrafoveolar 76 (36.9%) p=0.204, respectively). Multivariate analysis indicated that atrophy at baseline in the Mean VA (logMAR) of affected eye at baseline 0.63±0.46 fi Presence of GA on FAF alone 46 (22.3%) study eye was signi cantly associated with CNV or atrophy in the fellow eye (OR=3.91 95% CI (1.86–8.24), p=0.001 and Presence of GA on multimodal imaging 50 (24.3%) – Presence of foveal GA 20 (43.5%) OR=3.16 95% CI (1.26 7.92), p=0.014 respectively), whereas Mean VA (logMAR) of fellow eye at baseline 0.54±0.61 mixed CNV in the study eye were associated with less atrophy in the study eye (OR=0.21 95% CI (0.05–0.97), p=0.046). CNV, choroidal neovascularisation; FAF, fundus autofluorescence; GA, geographic atrophy; logMAR, logarithm of the minimum of angle of resolution; VA, visual acuity. Regarding CNV type, subset analysis of frequency for each CNV type associated with GA found a higher incidence of type 3 (31.25%), followed by type 2 CNV (30.23%), type 1 (29.55%), mixed CNV (9.09%) and PCV (6.25%). The fellow eye presented CNV in 66/206 cases (32%), Mean size of GA was significantly higher in group of patient early AMD in 53/206 cases (25.7%) or GA in 32/206 cases with type 1 CNV, type 2 CNV or mixed CNV than in group of (15.5%). No abnormalities were present in the 22/206 remain- patients with either type 3 CNV or PCV (1.6±1.9 mm2 vs ing cases (10.7%). Demographic and clinical data are sum- 0.2±1.3 mm2 p=0.015). We did not find that a specific CNV marised in table 1. subtype was preferentially associated with GA (table 5).

Predictive factors of visual acuity and size of atrophy, in Analysis of qualitative and quantitative parameters of GA case of atrophy at baseline Evaluation of GA based on FAF showed that GA was present in Multivariate analysis reported that both type 2 CNV and retro- fi 46/206 eyes (22.3%) with neovascular AMD ( gure 1). The foveolar localisation of atrophy were predictive for lower visual localisation of atrophy was retrofoveal in 43.5% of cases, juxta- acuity when atrophy was associated with exudative AMD at foveal in 86.9% of cases and extrafoveal in 19.5% of cases (for baseline (ß=0.19 (95% CI (0.034–0.34), p=0.017 and ß=0.18 the same patient, atrophy could affect one or more of these (95% CI (0.058–0.311), p=0.004, respectively). No predictive fi areas). There was no signi cant correlation between atrophy factors of size of atrophy were found in our study. localisation and CNV localisation, whatever the CNV type. Mean size of GA was 1.23±1.76 mm2 (range 0.03–7.39). Quantification of GA on FAF frames was not applicable or was DISCUSSION with limited reliability in cases of macular haemorrhage (8%), This study, conducted in a tertiary referral centre found that GA poor image quality of FAF frames (20.8%) or absence of FAF is present in 24.3% of patients with newly diagnosed exudative imaging (6.5%). In these cases, multimodal imaging, with infra- AMD. The reported prevalence of GA associated with exudative red, OCT and FA and ICGA, identified the presence of atrophy AMD at baseline is heterogeneous in the literature, from 7.3% in 2% of eyes. in the Comparison of AMD Treatment Trial (CATT) study,15 Using data from RFA and multimodal imaging, which is a 18.1% in the study of Xu et al16 to 58.6% in the study of more reliable approach compared with evaluation of FAF Kumar et al.14 Tanaka et al reported that GA within or overlap- imaging only,26 GA was present in 24.3% of study eyes newly ping the boundary of the entire CNV was present in 5% and diagnosed with exudative AMD. In 2.9% of eyes (corresponding outside this boundary in 10% of eyes with treatment-naïve to patients with poor image quality of FAF frames or with exudative AMD. advanced neovascular lesion with extensive subretinal fibrosis Previously reported mean area of prevalent GA are also which made a grading impossible) neither multimodal reading heterogeneous—from 0.27±0.99 mm2 (range 0–8.1)16 to nor reading by the third grader allowed to decide whether 2.58 mm2 (SD 0.99 mm2).15 atrophy was present or not. Grading of atrophy, when present, All these reports differ from our results. These findings do not is reported in table 2. necessarily contradict ours, given the differences of criteria and Mean and median areas of GA measured using RFA were imaging tools used to characterise GA, as well as the study design comparable for the three graders (table 3 and figure 2). The and methods. In the CATT study, determination of atrophy was correlation coefficient was R2=0.961 for the mean size of obtained with only fundus colour photographs and FA and with a GA between graders 1+2 (AS+SZ) and grader 3 (OS). GA larger diameter threshold of 250 m.InXuet al16 study, deter- Inter-observer reproducibility, as assessed with the ICC, was mination of atrophy was obtained with near IRR imaging and

440 Sikorav A, et al. Br J Ophthalmol 2017;101:438–444. doi:10.1136/bjophthalmol-2015-308065 Downloaded from http://bjo.bmj.com/ on August 18, 2017 - Published by group.bmj.com Clinical science

Figure 1 Example of geographic atrophy (GA) associated to type 3 neovascularization. Multicolour imaging (top left), fundus autofluorescence (FAF) (middle top), Region Finder Analyser (top right), late phase indocyanine green angiography (ICGA) (bottom left), late phase fluorescein angiography (FA) (middle bottom) and spectral-domain optical coherence tomography (SD OCT) scan (bottom right) of the left eye of a 87-year-old-woman showing GA with chorioretinal anastomosis (CRA) at phase I (erosion sign) at baseline. Haemorrhages (arrowheads) and GA (arrows) are seen as low FAF. Areas of absent autofluorescence corresponding to GA are measured using Region Finder Analyser. Typical ‘hot spot’ is not well defined on late ICGA. FA shows areas of hyperfluorescence with well-demarcated margins corresponding to GA and a focal hyperfluorescence corresponding to CRA (asterisk). The corresponding (eye-tracked) SD-OCT shows both GA with retinal pigment epithelial (RPE) loss and increased choroidal reflectivity and CRA, with small, focal RPE erosion over a localised RPE elevation filled with a hyper-reflective material.

VEGF. Subfoveal GA at baseline was an exclusion criterion in Table 2 Grading of atrophy frequency both CATTand Xu et al studies. Parameter Quantity (%) Finally, in both Kumar et al and CATT studies, patients whose atrophy could not be evaluated (missing or ungradable imaging) Presence of atrophy on one imaging (FAF or OCT or FA) with the 86.0 were not included in the analysis. readers unanimity In a recent study, Kuroda et al have specifically evaluated the Presence of atrophy diagnosed by multimodal imaging with the 5.8 readers unanimity prevalence of RPE atrophy in Japanese patients with neovascu- Presence of atrophy diagnosed by multimodal imaging with the 5.3 lar AMD. GA was observed in 5.1% of patients at baseline, readers majority which is lower compared with our results. This can be explained Impossibility to decide 2.9 by the fact that Asian patients have different subtype distribu- tion of AMD than white patients. Indeed, they have excluded FA, fluorescein angiography; FAF: fundus autofluorescence; OCT, optical coherence tomography. RAP owing to the fact that it represents a rare subtype of neo- vascular AMD in Japanese populations.31 In another recent study, Schutze et al have evaluated RPE atrophy progression during antiangiogenic therapy of neovascu- Table 3 Mean and median size of geographic atrophy (GA) lar AMD over 2 years using polarization-sensitive OCT. Patients complicating exudative age-related macular degeneration at with GA at baseline (defined as atrophic RPE lesions >0.1 mm2 baseline, measured by readers 1+2 (AS+SZ) and reader 3 (OS) using a polarization-sensitive OCT-related algorithm) were Readers 1+2 Reader 3 excluded from their analysis but primary RPE-related alterations fi 2 2 2 such as RPE porosity and focal atrophy were identi ed in 23% Mean baseline area of GA (mm ) 1.23±1.76 mm 1.22±1.84 mm 32 (range 0.03–7.39) and 55%, respectively, of treatment-naïve eyes at baseline. Median baseline size of GA (mm2) 0.44 mm2 0.42 mm2 The originality of the present study relies on the facts that all (IQR 0.16–1.48) (IQR 0.17–1.55) consecutive treatment-naïve patients seen in our department were analysed, including patients with media opacities, fundus haemorrhages or with retinal scar, that could interfere with FAF analysis. Furthermore, atrophy was evaluated based on a multi- SD-OCT and with a GA larger diameter threshold of 250 m. modal approach. Finally, minimal GA size chosen for inclusion Kumar et al,14 defined atrophy as being present if there was a con- (0.02 mm2) was related to the definition of the digital image of fluent absence of autofluorescence signal of at least >500 mm FAF imaging, as previously described.26 based only on FAF imaging analysis. Moreover, some of these Multimodal imaging with simultaneous evaluation of FAF, patients have been previously treated with photodynamic near infrared fundus autofluorescence (NIR-FAF), FA and therapy, injections of triamcinolone, dexamethasone or anti- SD-OCT seems to represent a better and more sensitive

Sikorav A, et al. Br J Ophthalmol 2017;101:438–444. doi:10.1136/bjophthalmol-2015-308065 441 Downloaded from http://bjo.bmj.com/ on August 18, 2017 - Published by group.bmj.com Clinical science

Our study raises the question whether GA associated with exudative AMD at baseline is similar to GA in dry AMD in terms of physiopathological pathways, histology, genetics and clinical phenotype. These two phenotypes could also have dif- ferent functional impact and different treatment approaches might need to be applied in future. Recent reports have hypothesised that both GA and exudative AMD are multifactor- ial diseases involving common environmental and genetic risk factors and could share common physiopathological path- ways.36 37 One study found that patients with CNV in one eye and GA in the fellow eye develop CNV in the fellow eye in 30%–50% after 5 years of follow-up.8 Another study reported that patients with CNV in one eye and GA in the fellow eye may develop CNV in this eye in approximately 50% at 5 years.38 The CATT study,12 established that 18.3% of patients with no GA at baseline treated for exudative AMD developed Figure 2 Mean and median sizes of geographic atrophy measured GA at 2 years. However, as mentioned above, the evaluation of using Region Finder Analyser by graders 1+2 (AS+SZ) in ordinate and GA at study enrolment was based on colour fundus photog- grader 3 (OS) in abscissa. raphy only and it is very likely that in some patients some atro- phic areas were already present at baseline but not diagnosed due to haemorrhage or fibrosis. Grob et al,39 investigated geno- types in patients with simultaneous GA and CNV in the same Table 4 Characteristics of the eyes that did (group 1) and did not eye and compared them with patients with only GA or CNV. (group 2) present geographic atrophy at initial presentation for They found that none of the currently known at risk genotypes exudative age-related macular degeneration with univariate analysis were significantly associated with occurrence of simultaneous Group 1 Group 2 p Value GA and CNV.The authors concluded that GA and CNV may be one disease on a continuum. It can therefore also be assumed Female/male 47/ 10 105/44 0.149 30.1%/18.5% 69.1%/81.5% that patients with GA will at some point potentially develop Age 82.0±9.2 78.8±8.9 0.019 CNV and patients with CNV will at some point develop GA. BCVA baseline (logMAR) 0.72±0.46 0.57±0.41 0.004 GA associated with exudative AMD and pure GA may therefore Pseudophakic or aphakic/phakic 4/48 25/129 0.330 be two phenotypes representing a single disease and have 36 37 39 13.8%/27.1% 86.2%/72.9% similar biochemical pathways. Vitamin supplementation/no 18/39 41/108 0.537 Considering clinical presentation, to date, no consensual cri- vitamin supplementation 30.5%/26.5% 69.5%/73.5% teria are clearly defined in the literature to differentiate GA et al Only eyes where grading was possible were included. associated with CNV and pure GA. Similar to Kumar BCVA, best-corrected visual acuity; logMAR, logarithm of the minimum of angle of study,14 we observed in our study that pure GA may usually resolution. develop in a perifoveal location with preservation of the fovea, whereas GA associated with exudative AMD involves more fre- quently the foveal region at baseline, with a lower BCVA as con- Table 5 Subset analysis of frequency for each CNV type sequence. Moreover, a growth rate of GA associated with – 2 15 16 associated with geographic atrophy (GA) exudative AMD was found to be 0.45 0.58 mm /year, slower than in GA alone (1.3–2.8 mm2/year).74041Finally, GA Frequency association Mean size of GA associated with exudative AMD progresses in a centripetal way, CNV type with GA (%) associated (mm2) compared with pure GA.14 It could be hypothesised that GA Type 1 29.55 1.78±2.17 associated with exudative AMD and GA have common genetic CRA 31.25 0.17±0.10 and physiopathological pathways, but different clinical presenta- Type 2 30.23 1.07±1.24 tion and progression. Combined CNV 9.09 2.34±2.94 We did not observe any significant difference in the presence PCV 6.25 0.44±0 of GA depending on the crystalline lens status and the vitamin CNV, choroidal neovascularisation; CRA, chorioretinal anastomosis; PCV, polypoidal supplementation, suggesting that these two factors may not par- choroidal vasculopathy. ticipate in the presence of GA. GA was found to be slightly more present at initial diagnosis of exudative AMD in eyes with a type 3 CNV, than type 1 and type 2 CNV. McBain et al42 also approach to identify and quantify GA, as well as foveal-sparing found similar high rate (36%) of GA in treatments-naïve compared with evaluation of FAF imaging only.25 29 33 We did patients with type 3 CNV at baseline. The size of GA was sig- not consider NIR-FAF in our analysis because it was not per- nificantly greater in the group of patients with type 1 CNV, formed systematically for all our patients. type 2 CNV or mixed CNV than in the group of patients with RFA, based on FAF, offers an accurate, reproducible and time- either type 3 CNV or PCV, but no difference was observed efficient tool for identification and quantification of GA23 30 among the different CNV subtype. This result could be and can be used in daily clinical practice.26 Other additional explained by our relatively small sample size. There was no sig- imaging methods, including SD-OCT with integrated microperi- nificant correlation between atrophy localisation and CNV metry,34 or the ‘Index of Retinal Autofluorescence’ applied to localisation, whatever the CNV type, indicating that CNV and FAF imaging,35 may further improve the accuracy and reprodu- atrophic areas are not always co-located at initial diagnosis of cibility of atrophy quantification. wet AMD.

442 Sikorav A, et al. Br J Ophthalmol 2017;101:438–444. doi:10.1136/bjophthalmol-2015-308065 Downloaded from http://bjo.bmj.com/ on August 18, 2017 - Published by group.bmj.com Clinical science

Our study has several limitations. It is a cross-sectional and 10 Rosenfeld PJ, Shapiro H, Tuomi L, et al. Characteristics of patients losing vision after retrospective study, performed in a tertiary referral centre, 2 years of monthly dosing in the phase III ranibizumab clinical trials. Ophthalmology 2011;118:523–30. which present the biases inherent to these types of studies. 11 Martin DF, Maguire MG, Fine SL, et al., Comparison of Age-related Macular Considering imaging, we have not analysed the different var- Degeneration Treatments Trials Research Group. Ranibizumab and bevacizumab for iants of abnormal autofluorescence, and especially presence of treatment of neovascular age-related macular degeneration: two-year results. hyperautofluorescence, which could be predictive of future Ophthalmology 2012;119:1388–98. alteration of the RPE,22 43 44 and thus may have an impact on 12 Grunwald JE, Daniel E, Huang J, et al. Risk of geographic atrophy in the comparison of age-related macular degeneration treatments trials. Ophthalmology visual function or response to treatment; this study did not take 2014;121:150–61. into consideration the duration of symptoms for each patient, 13 Investigators IS, Chakravarthy U, Harding SP, et al. Ranibizumab versus which might have an effect on the FAF findings. NIR-FAF bevacizumab to treat neovascular age-related macular degeneration: one-year imaging was not performed systematically for all our patients findings from the IVAN randomized trial. Ophthalmology 2012;119:1399–411. 14 Kumar N, Mrejen S, Fung AT, et al. Retinal pigment epithelial cell loss assessed by and thus was not taken into account in our multimodal analysis, fundus autofluorescence imaging in neovascular age-related macular degeneration. despite previous study reporting that NIR-FAF allows earlier Ophthalmology 2013;120:334–41. detection of foveal sparing in patients with GA.29 Finally, it is 15 Grunwald JE, Pistilli M, Ying GS, et al. Growth of geographic atrophy in the important to make a differentiation between the digital and the comparison of age-related macular degeneration treatments trials. Ophthalmology – anatomical resolution. Our results are based on an image ana- 2015;122:809 16. 16 Xu L, Mrejen S, Jung JJ, et al. Geographic atrophy in patients receiving anti-vascular lysis and not an anatomical analysis. Therefore, the results are endothelial growth factor for neovascular age-related macular degeneration. Retina limited by the resolution of the imaging technology. (Philadelphia, Pa) 2015;35:176–86. In conclusion, we report here that GA is associated with neo- 17 Delori FC, Dorey CK, Staurenghi G, et al. In vivo fluorescence of the ocular fundus vascular AMD in 1/4 of cases at initial presentation, with 43.5% exhibits retinal pigment epithelium lipofuscin characteristics. Invest Ophthalmol Vis fi Sci 1995;36:718–29. early foveal involvement. The size of GA is signi cantly more 18 von Ruckmann A, Fitzke FW, Bird AC. Distribution of fundus autofluorescence with important in type 1, type 2 and mixed CNV. Multimodal a scanning laser ophthalmoscope. Br J Ophthalmol 1995;79:407–12. imaging, including RFA, is an effective tool to identify and 19 Spaide RF. Fundus autofluorescence and age-related macular degeneration. quantify atrophic retinal areas at diagnosis in everyday practice. Ophthalmology 2003;110:392–9. fl The physiopathological mechanisms of the association of GA 20 McBain VA, Townend J, Lois N. Fundus auto uorescence in exudative age-related macular degeneration. Br J Ophthalmol 2007;91:491–6. and CNV need to be further characterised, as well as the long- 21 Bindewald A, Schmitz-Valckenberg S, Jorzik JJ, et al. Classification of abnormal term impact of initial atrophy and its functional consequences. fundus autofluorescence patterns in the junctional zone of geographic atrophy A prospective follow-up of our study is currently being per- in patients with age related macular degeneration. Br J Ophthalmol formed. Controlled clinical trials are also necessary to shed light 2005;89:874–8. 22 Holz FG, Bellman C, Staudt S, et al. Fundus autofluorescence and development of on the actual controversy surrounding the possible role of geographic atrophy in age-related macular degeneration. Invest Ophthalmol Vis Sci anti-VEGF in the presence of GA. 2001;42:1051–6. 23 Holz FG, Bindewald-Wittich A, Fleckenstein M, et al. Progression of geographic Contributors Design and conduct of the study: AS and OS; collection, atrophy and impact of fundus autofluorescence patterns in age-related macular management and analysis: AS, OS, MS, SZ and CJ; interpretation of the data: degeneration. Am J Ophthalmol 2007;143:463–72. AS and OS and preparation, review or approval of the study: AS, OS, GQ and EHS. 24 Schmitz-Valckenberg S, Bindewald-Wittich A, Dolar-Szczasny J, et al. Correlation Funding Each of the coauthors has seen and has agreed with each of the changes between the area of increased autofluorescence surrounding geographic atrophy made to this manuscript in the revision and to the way his or her name is listed. and disease progression in patients with AMD. Invest Ophthalmol Vis Sci 2006;47:2648–54. Competing interests None declared. 25 Schmitz-Valckenberg S, Fleckenstein M, Gobel AP, et al. Optical coherence Ethics approval French Society of Ophthalmology Ethics Comittee. tomography and autofluorescence findings in areas with geographic atrophy due to age-related macular degeneration. Invest Ophthalmol Vis Sci 2011;52:1–6. Provenance and peer review Not commissioned; externally peer reviewed. 26 Panthier C, Querques G, Puche N, et al. Evaluation of semiautomated measurement of geographic atrophy in age-related macular degeneration by fundus autofluorescence in clinical setting. Retina 2013;34:576–82. REFERENCES 27 Holladay JT. Visual acuity measurements. J Cataract Refract Surg 2004;30:287–90. 1 SanGiovanni JP, Chew EY, Clemons TE, et al., Age-Related Eye Disease Study 28 Moussa K, Lee JY, Stinnett SS, et al. Spectral domain optical coherence Research Group. The relationship of dietary carotenoid and vitamin A, E, and C tomography-determined morphologic predictors of age-related macular intake with age-related macular degeneration in a case-control study: AREDS Report degeneration-associated geographic atrophy progression. Retina (Philadelphia, Pa) No. 22. Arch Ophthalmol 2007;125:1225–32. 2013;33:1590–9. 2 Bressler NM. Age-related macular degeneration is the leading cause of blindness. 29 Forte R, Querques G, Querques L, et al. Multimodal evaluation of foveal sparing in JAMA 2004;291:1900–1. patients with geographicatrophy due to age-related macular degeneration. 3 Klein R, Klein BE, Jensen SC, et al. The five-year incidence and progression of Retina (Philadelphia, Pa) 2013;33:482–9. age-related maculopathy: the Beaver Dam Eye Study. Ophthalmology 30 Schmitz-Valckenberg S, Brinkmann CK, Alten F, et al. Semiautomated image 1997;104:7–21. processing method for identification and quantification of geographic atrophy 4 Klein R, Klein BE, Knudtson MD, et al. Fifteen-year cumulative incidence of in age-related macular degeneration. Invest Ophthalmol Vis Sci age-related macular degeneration: the Beaver Dam Eye Study. Ophthalmology 2011;52:7640–6. 2007;114:253–62. 31 Kuroda Y, Yamashiro K, Tsujikawa A, et al. Retinal pigment epithelial atrophy in 5 Augood CA, Vingerling JR, de Jong PT, et al. Prevalence of age-related maculopathy neovascular age-related macular degeneration after Ranibizumab treatment. in older Europeans: the European Eye Study (EUREYE). Arch Ophthalmol Am J Ophthalmol 2016;161:94–103.e1. 2006;124:529–35. 32 Schutze C, Wedl M, Baumann B, et al. Progression of retinal pigment epithelial 6 Holz FG, Strauss EC, Schmitz-Valckenberg S, et al. Geographic atrophy: atrophy in antiangiogenic therapy of neovascular age-related macular degeneration. clinical features and potential therapeutic approaches. Ophthalmology Am J Ophthalmol 2015;159:1100–14.e1. 2014;121:1079–91. 33 Sayegh RG, Simader C, Scheschy U, et al. A systematic comparison of 7 Schatz H, McDonald HR. Atrophic macular degeneration. Rate of spread of spectral-domain optical coherence tomography and fundus autofluorescence geographic atrophy and visual loss. Ophthalmology 1989;96:1541–51. in patients with geographic atrophy. Ophthalmology 2011;118:1844–51. 8 Sunness JS, Gonzalez-Baron J, Bressler NM, et al. The development of choroidal 34 Querques L, Querques G, Forte R, et al. Microperimetric correlations of neovascularization in eyes with the geographic atrophy form of age-related macular autofluorescence and optical coherence tomography imaging in dry age-related degeneration. Ophthalmology 1999;106:910–19. macular degeneration. Am J Ophthalmol 2012;153:1110–15. 9 McCarthy LC, Newcombe PJ, Whittaker JC, et al. Predictive models of choroidal 35 Schachar IH, Zahid S, Comer GM, et al.Quantification of fundus autofluorescence neovascularization and geographic atrophy incidence applied to clinical trial design. to detect disease severity in nonexudative age-related macular degeneration. Am J Ophthalmol 2012;154:568–78.e12. JAMA Ophthalmol 2013;131:1009–15.

Sikorav A, et al. Br J Ophthalmol 2017;101:438–444. doi:10.1136/bjophthalmol-2015-308065 443 Downloaded from http://bjo.bmj.com/ on August 18, 2017 - Published by group.bmj.com Clinical science

36 Spaide RF, Armstrong D, Browne R. Continuing medical education 40 Sunness JS, Gonzalez-Baron J, Applegate CA, et al. Enlargement of atrophy and review: choroidal neovascularization in age-related macular visual acuity loss in the geographic atrophy form of age-related macular degeneration—what is the cause? Retina (Philadelphia, Pa) degeneration. Ophthalmology 1999;106:1768–79. 2003;23:595–614. 41 Klein R, Davis MD, Magli YL, et al. The Wisconsin age-related maculopathy grading 37 Zarbin MA. Current concepts in the pathogenesis of age-related macular system. Ophthalmology 1991;98:1128–34. degeneration. Arch Ophthalmol 2004;122:598–614. 42 McBain VA, Kumari R, Townend J, et al. Geographic atrophy in retinal 38 Risk factors for choroidal neovascularization in the second eye of patients with angiomatous proliferation. Retina (Philadelphia, Pa) 2011;31:1043–52. juxtafoveal or subfoveal choroidal neovascularization secondary to age-related 43 Dandekar SS, Jenkins SA, Peto T, et al. Autofluorescence imaging of choroidal macular degeneration. Macular Photocoagulation Study Group. Arch Ophthalmol neovascularization due to age-related macular degeneration. Arch Ophthalmol 1997;115:741–7. 2005;123:1507–13. 39 Grob S, Luo J, Hughes G, et al. Genetic analysis of simultaneous 44 Heimes B, Lommatzsch A, Zeimer M, et al. Foveal RPE autofluorescence as a geographic atrophy and choroidal neovascularization. Eye (Lond) prognostic factor for anti-VEGF therapy in exudative AMD. Graefes Arch Clin Exp 2012;26:1106–13. Ophthalmol 2008;246:1229–34.

444 Sikorav A, et al. Br J Ophthalmol 2017;101:438–444. doi:10.1136/bjophthalmol-2015-308065 Downloaded from http://bjo.bmj.com/ on August 18, 2017 - Published by group.bmj.com

Prevalence and quantification of geographic atrophy associated with newly diagnosed and treatment-naïve exudative age-related macular degeneration Anne Sikorav, Oudy Semoun, Sandrine Zweifel, Camille Jung, Mayer Srour, Giuseppe Querques and Eric H Souied

Br J Ophthalmol 2017 101: 438-444 originally published online August 8, 2016 doi: 10.1136/bjophthalmol-2015-308065

Updated information and services can be found at: http://bjo.bmj.com/content/101/4/438

These include: References This article cites 44 articles, 8 of which you can access for free at: http://bjo.bmj.com/content/101/4/438#BIBL Email alerting Receive free email alerts when new articles cite this article. Sign up in the service box at the top right corner of the online article.

Notes

To request permissions go to: http://group.bmj.com/group/rights-licensing/permissions

To order reprints go to: http://journals.bmj.com/cgi/reprintform

To subscribe to BMJ go to: http://group.bmj.com/subscribe/