RESPONSE OF PIGMENT EPITHELIAL DETACHMENTS TO INTRAVITREAL AFLIBERCEPT AMONG PATIENTS WITH TREATMENT-RESISTANT NEOVASCULAR AGE-RELATED MACULAR DEGENERATION GEOFFREY K. BROADHEAD, MBBS, MPH,*† THOMAS HONG, MSCMED, BAPPSC,* MEIDONG ZHU, MBBS, PHD,*† HAITAO LI, MBBS, PHD,* TIMOTHY E. SCHLUB, BSC (HONS), PHD,‡ WIJEYANTHY WIJEYAKUMAR, MOTH, BSC,*† ANDREW A. CHANG, FRANZCO, PHD*† Purpose: To assess the effect of intravitreal aflibercept on pigment epithelial detachment (PED) in patients with treatment-resistant neovascular age-related macular degeneration. Methods: Forty-six patients with vascularized PEDs participating in a wider, prospective clinical trial of treatment-resistant neovascular age-related macular degeneration received 2-mg aflibercept as 3 loading doses 1 month apart, followed by further 2-monthly doses over a total 12-month period. Change in PED dimensions and reflective properties were assessed by optical coherence tomography. Reflectivity was subclassified as solid (hyperreflective), hollow (hyporeflective), or mixed (elements of both). Results: Aflibercept reduced PED height, width, and length at 48 weeks compared with baseline values (P # 0.01 for all). Reductions in PED height were correlated with reductions in central macular thickness at 48 weeks (R2 = 0.36, P , 0.001). There was no significant correlation between PED height decrease and visual acuity changes at 48 weeks. Solid PEDs were less likely to experience reductions in all three dimensions than either hollow or mixed PEDs. Conclusion: Aflibercept is effective in reducing PED dimensions in treatment-resistant patients and is most effective in PEDs demonstrating some hyporeflective optical coherence tomography characteristics. Reduction in PED dimensions correlated with central macular thickness, but not with visual acuity changes. The role of PEDs as markers of disease requires further investigation; however, lesions should be monitored for retinal fluid recurrence. RETINA 35:975–981, 2015
P
igment epithelium detachment (PED) is a wellknown complication of neovascular age-related macular degeneration (AMD). These lesions occur when the retinal pigment epithelium (RPE) detaches from the underlying Bruch membrane, and are often associated with choroidal neovascularization (CNV).1 Vascularized PEDs occur in conjunction with CNV, with fibrovascular material, or hemorrhage accumulating within the PED.2 These lesions can progress to form disciform scars, resulting in poor visual outcomes. Vascularized PEDs may be complicated by RPE rips and tears.3
Previous studies have shown that PEDs respond poorly to treatments for CNV, including photodynamic therapy4 and intravitreal anti-vascular endothelial growth factor (anti-VEGF) agents.5,6 Some studies have suggested that the reflective properties of PEDs on spectral domain optical coherence tomography (SD-OCT) may predict the response to anti-VEGF therapy.7 Less reflective PEDs have been reported as responding better to anti-VEGF therapies. These reflective properties may relate to differences in the properties of the underlying neovascular complex.8 Pigment epithelial detachment location has also been 975
976
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 5
associated with visual outcomes, and subfoveal PEDs have been associated with worse vision.9 Aflibercept (VEGF-Trap Eye/Eylea; Regeneron, Tarrytown, NY), a newer anti-VEGF agent, has been shown to be effective in treating neovascular AMD10 and is reported to have broader antiangiogenic range of activity than other anti-VEGF therapies.11,12 There is limited evidence to suggest that aflibercept may achieve improvement in structural PED characteristics, and that this may correlate with visual gains, in patients who are resistant to other anti-VEGF therapies.4,13 Therefore, we prospectively assessed the effect of aflibercept on PED characteristics and evaluated the correlation of these changes with visual and anatomical outcomes in patients with treatmentresistant neovascular AMD. Methods The study population, inclusion and exclusion criteria, and methods of a registered, prospective open-label clinical trial evaluating the use of aflibercept in treatment-resistant neovascular AMD have been previously reported.14 Participants from that trial who met the following criteria were included in this study: 1) known CNV secondary to AMD as demonstrated by fluorescein angiography; 2) persistent intraretinal or subretinal fluid on SD-OCT for at least 6 months despite at least 4 intravitreal anti-VEGF injections in the past 6 months; 3) persistent PED on SD-OCT for at least 6 months despite at least 4 antiVEGF injections in the past 6 months; 4) the presence of PED on SD-OCT at baseline visit with maximum PED height .100 mm; and 5) hyperfluorescence on fluorescein angiography contiguous with the PED indicative of vascularization of the PED. All patients were treated with aflibercept (2.0 mg) given as 3 initial loading doses (Weeks 0, 4, and 8) followed by further treatments every 8 weeks (Weeks 16, 24, 32, 40, and 48). All patients underwent review From the *Sydney Institute of Vision Science, Sydney, Australia; †Save Sight Institute, The University of Sydney, Sydney, Australia; and ‡Sydney School of Public Health, The University of Sydney, Sydney, Australia. Supported by Bayer Corporation Global. Paper presented at the Royal Australian College of Ophthalmologists Annual Congress, Hobart, Australia, November 2013. A. A. Chang has acted as a consultant for Alcon, Bayer and Novartis. The other authors do not have any financial/conflicting interests to disclose. G. K. Broadhead and T. Hong contributed equally to this work. Reprint requests: Andrew A. Chang, FRANZCO, PhD, Sydney Retina Clinic and Day Surgery, Level 13, Park House, 187 Macquarie Street, Sydney, New South Wales 2000, Australia; e-mail: [email protected]
1 week after initial injection, then every 4 weeks regardless of the injection status. At each visit, a comprehensive ophthalmic examination was performed including best-corrected visual acuity in Early Treatment of Diabetic Retinopathy Study letters and SD-OCT (Heidelberg Spectralis; Heidelberg Industries, Heidelberg, Germany). Fluorescein angiography and indocyanine green angiography were conducted at the baseline visit to confirm the presence of active neovascularization associated with the PED (Figure 1) and to exclude confounding lesions such as polypoidal vasculopathy. Continuity of OCT scan position was ensured through the use of inbuilt eye tracking and image recognition software (Tru-Track and AutoRescan; Heidelberg Industries, Heidelberg, Germany). Central macular thickness (CMT) was defined as the distance between the internal limiting membrane and Bruch membrane, and manual resegmentation of those layers was performed where necessary.14 The presence of PED resolution and RPE tears was recorded at each visit. Two independent graders reviewed PED dimensions and reflectivity characteristics, discrepancies were adjudicated by the third independent assessor. Pigment epithelial detachments were graded as hollow, solid, or mixed based on their SD-OCT reflectivity according to the method previously described by Punjabi et al7 (Figure 2). The location of PEDs was assessed at baseline and categorized as subfoveal (within 500 mm of fovea) or extrafoveal (.500 mm from fovea). Pigment epithelial detachment size was measured in three maximum dimensions (height, width, and length) at each visit using inbuilt callipers within the SD-OCT software. On OCT scans, PED width was measured as the maximal horizontal PED diameter and PED length was measured as the maximal PED vertical diameter. Height was measured as the vertical distance from Bruch membrane to the RPE border. Length and width were measured as the maximal distance from RPE elevation to RPE elevation in any scan. Statistical analyses were performed using R version 2.15.2 (R Foundation for Statistical Computing, Vienna, Austria). Data were described as mean ± standard deviation or frequency. The effect of baseline characteristics, including injection frequency, total injection number, lesion duration, and angiographic subtype (retinal angiomatous proliferation vs. occult CNV) on PED response was also analyzed. For all correlation analyses and the comparison of change based on PED reflectivity and PED location, change at the 48-week time point was used for analysis. Student’s t-tests and the nonparametric Mann–Whitney U tests were used to compare
AFLIBERCEPT AND PEDs IN REFRACTORY AMD BROADHEAD ET AL
977
Fig. 1. Angiography demonstrating a vascularized extrafoveal PED. A. Fundus photograph, (B) autofluorescence image, (C) earlyphase indocyanine green angiography with a discrete juxtafoveal area of hypofluorescence consistent with a PED, (D) latephase indocyanine green angiography showing patchy central hyperfluorescence, (E) earlyphase fluorescein angiography with a discrete juxtafoveal area of hypofluorescence, (F) latephase fluorescein angiography showing marked hyperfluorescence surrounding the margins of the PED, indicative of a vascularized PED.
continuous outcomes against binary covariates. As both tests gave similar values of P, those of the Mann–Whitney U test are presented. Correlations between continuous variables were assessed with Pearson’s correlations and Spearman rank correlations, showing similar results. Pearson’s correlation coefficients and values of P are presented for more meaningful interpretation of correlation coefficients. The average effect of spacing injection frequency on PED dimensions was assessed with a random effects model with the patient as the grouping factor, using the “lme”
function in the “nlme” package in R. A P , 0.05 was considered to be significant.
Results Forty-three patients who met inclusion criteria were included in this study. Baseline characteristics of the patient cohort are summarized in Table 1. Twenty of the participants (47%) were male, and the mean patient age was 78.7 years. Twenty-one PEDs (49%) were
978
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 5
Fig. 2. Examples of PED subtypes: hollow (top), solid (center), and mixed (bottom) PED.
solid, 13 (30%) were hollow, and 9 (21%) were mixed. Thirty-four PEDs (79%) were subfoveal, and 9 (21%) were extrafoveal. Mean baseline PED height, width, and length was 218.5 ± 144.9 mm, 2,337.7 ± 1,234.3 mm, and 2,580.2 ± 1,715.4 mm, respectively. Response of Pigment Epithelial Detachment to Aflibercept Therapy Pigment epithelial detachment height decreased by an average of −51.5 mm at Week 24 and −53.3 mm (range, −323 to 89 mm) at Week 48 compared with baseline (P , 0.01 for all follow-up visits). Pigment epithelial detachment width showed a similar trend, decreasing by an average of −320.6 mm at Week 24 and −276.8 mm (range, −2,798 to 438 mm) at Week 48 (P , 0.05 for all follow-up visits). Pigment epithelial detachment length decreased by −248.8 mm at Week 24 and −301.9 mm (range, −2084 to 439 mm) at Week 48 (P , 0.05 for all follow-up visits) (Figure 3). ComTable 1. Background Characteristics of the Patient Cohort Characteristic Male Age at baseline, years Right eye Mean total Lucentis injections Mean PED duration, months Pseudophakic PED type Solid Hollow Mixed PED location Subfoveal Extrafoveal
Frequency/Mean, n (%) 20 (47) 78.7 21 (49) 34.8 32.1 26 (60.0) 21 (49) 13 (30) 9 (21) 34 (79) 9 (21)
plete PED resolution occurred in 4 patients by the end of the 12-month period, and 42% (18 of 43) patients had PED improvement (.50 mm improvement), 56% (24 of 43) had stable PEDs, and 2% (1 of 43) had worsening (.50 mm increase) of their PED at Week 48. There was no difference in response in any dimension between those receiving anti-VEGF therapy every 4 weeks and every 6 weeks (P . 0.05), or between those with retinal angiomatous proliferation lesions and those with occult CNV related to vascularized PEDs (P . 0.05). No significant difference was seen in response of PED height or length between those who received previous ranibizumab monotherapy and those who received both previous ranibizumab and bevacizumab (P . 0.05), however, a difference was seen in response of PED width (difference of 311 mm response at 48 weeks, P = 0.03) between these 2 groups. Spacing Spacing of injection frequency from monthly to every 2 months resulted in significant increases in PED height, width, and length (average increase in height width and length for 1 month after injection compared with 2 months after injection: 20.1, 135.6, and 91.9 mm, respectively, P , 0.05 for all 3 dimensions). This oscillating pattern of change in PED dimensions in response to spacing of injection frequency is seen in Figure 3. Correlation With Central Macular Thickness and Vision Changes in PED height were correlated with changes in CMT (R2 = 0.36, P , 0.001). This correlation became stronger when only subfoveal PEDs were considered (R2 = 0.64, P , 0.001). Reduction in PED width was also correlated with CMT change (R2 = 0.25, P , 0.001), as were changes in PED length (R2 = 0.16, P , 0.01). The association between both PED width and CMT, as well as PED length and CMT, became less significant when only subfoveal PEDs were considered (R2 = 0.09, P = 0.10 for width, R2 = 0.12, P = 0.05 for length). Best-corrected visual acuity gains of 6.8 and 4.6 letters were observed at Weeks 24 and 48, respectively (P , 0.001 for both). Approximately 47% of patients had a visual gain of $5 letters at Week 48. We observed minor associations between reflectivity characteristics, duration of PED, baseline best-corrected visual acuity, and changes in PED height, width, or length with changes in best-corrected visual acuity, but none of these were statistically significant (P . 0.10 for all).
AFLIBERCEPT AND PEDs IN REFRACTORY AMD BROADHEAD ET AL
979
Fig. 3. Changes in PED height, width, and length compared with baseline examination over 48 weeks.
Pigment Epithelial Detachment Reflectivity Hollow (less reflective) PEDs exhibited a significant reduction in all 3 PED dimensions at Week 48 compared with baseline (mean change in height, width, and length: 70.8, 516.3, and 479.3 mm, respectively, P # 0.01 for all 3 dimensions). Mixed PEDs had a significant reduction in PED height and length (mean change in height and length: 113.2 and 649.6 mm, respectively, P = 0.02 for both dimensions), but not for width (mean change: 551.2 mm, P = 0.11). Solid (more reflective) PEDs had a significant reduction only in PED height (mean change: 16.9 mm, P = 0.048), but not for either length or width (mean change in length and width: 30.2 and 23.4 mm, respectively, P . 0.10 for both dimensions). Significant differences were seen in response to therapy based on the PED subtype, with both mixed and hollow PEDs significantly more likely to experience reductions in almost all 3 PED dimensions compared with solid PEDs (P , 0.05 for hollow or mixed versus solid PEDs for all 3 dimensions, excluding height in hollow to solid [P = 0.05] and width in mixed to solid [P = 0.099]). There was no difference seen in response between hollow and mixed PEDs (P . 0.10 for mixed versus hollow for all 3 dimensions). Pigment Epithelial Detachment Location and Retinal Pigment Epithelium Tears Subfoveal PEDs (79%) were more frequently observed than extrafoveal PEDs (21%; P , 0.001). The reduction in any of the PED dimensions was not associated with PED location (P . 0.10 for all
3 dimensions). No cases of RPE rips or tears were observed during the study period. Grading Variability Two independent graders showed a high level of agreement (mean difference and limits of agreement for height, width, and length were −1.64 mm [−27.7 to 24.4 mm], 7.3 mm [−182 to 196 mm], and 8.6 mm [−342 to 360 mm], respectively, Cohen’s Kappa on reflectivity characteristic = 0.853). Discussion In this study, aflibercept was found to effectively reduce PED size in all dimensions of vascularized PEDs that had been resistant to previous anti-VEGF therapy. In a few cases (4 of 43), complete PED resolution also occurred. Cases of PED resolution and improvement in PED dimensions with aflibercept for treatment-resistant PEDs have previously been reported.4,13 It is interesting to note that the reduction in both PED height and width observed in our study was similar to that seen in the retrospective study of aflibercept by Kumar et al13 in a similar treatmentresistant cohort (mean PED height change: 36 mm, mean width change: 316 mm). Reported changes in PED height in response to other anti-VEGF agents vary considerably. In newly diagnosed vascularized PEDs, Panos et al15 observed a change in a mean PED height of 135 mm after pro re nata ranibizumab for 12 months, whereas Ach et al16 showed a nonsignificant decrease in PED height (mean height decrease: 81.7 mm, P . 0.05) with bevacizumab for
980
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 5
vascularized PEDs. It is not clear why differences in PED height reduction have been observed, although in both this study and the study by Kumar et al, it may relate to the treatment-resistant nature of patients. Part of the discrepancy may also relate to differences in measurement techniques. In our study, we observe low bias across graders with a 1.6-mm average difference in PED height measurements between the 2 graders. The variability of the differences was also low, with 95% of disagreements ,30 mm. There may also be limitations with reliability of the images captured by the OCT, although the use of manual measurements by 2 graders was undertaken to minimize any errors that may arise because of automated measurements. However, the mean change itself is still small (53 mm), and as such it is unclear whether a change of this magnitude is a clinically meaningful outcome. The correlation between changes in PED height and CMT is to be expected, given that PED height forms a component of the CMT. As such, CMT may potentially be a surrogate marker of changes in PED architecture in response to aflibercept therapy. The lack of correlation between improvements in PED dimensions and significant additional visual improvement observed in our study is consistent with previous reports of PED response to anti-VEGF therapy.6,7,15,16 The visual improvement observed in these patients is therefore likely to be a result of improvement in other factors, such as intraretinal fluid or subretinal fluid, rather than reduction in PED size. The degree of visual gain obtained decreased slightly at Week 48 compared with the gain at Week 24 (2 letters), however a decline of this magnitude is not clinically significant. The reason for this decline is unknown. It is possible that the 2-monthly injection frequency, as opposed to the initial monthly loading therapy, contributed to this reduction of visual gain. However, other factors such as the longer follow-up duration (1 year) may have also impacted on the visual outcomes achieved. Vision loss associated with PEDs thus seems to be largely nonreversible, even with structural reduction of the lesion. This lack of correlation between PED reduction and vision may make changes in PED size less useful as a clinical marker of overall treatment response. However, it is clear that these sites bear close monitoring, because they are sites over which other lesions associated with visual impairment, such as subretinal fluid and intraretinal fluid can accumulate. Previous studies have suggested that PED reflectivity may be an important characteristic in response to therapy.7,17 We observed similar findings, with the response of PEDs to aflibercept therapy differing based on OCT reflectivity. Pigment epithelial detachments with significant hyporeflectivity (hollow or mixed)
were more responsive than hyperreflective (solid) PEDs. It has been suggested that the hyporeflective components may reflect the presence of fluid exudate, and the hyperreflective component of mixed and solid PEDs represents fibrinous leakage or fibrovascular proliferation, suggesting active neovascularization.2,8 The increased response of lesions with some hyporeflective component may thus relate to a reduction in the exudative component of these reflective subtypes as a result of inhibition of VEGF-driven vasodilation and vascular leakage. Hyperreflective PEDs may sometimes involve a significant lipid or fibrous component in addition to a neovascular membrane.2 These materials would be expected to be less responsive to aflibercept therapy than exudative substances such as fluid or blood. The presence of such material in some of the solid PEDs in this study would explain why these lesions had smaller changes in response to aflibercept therapy. Mixed PEDs may have shown a greater response to aflibercept than solid lesions because they contain some exudative components. It is interesting that some improvement was still noted in solid PED height, suggesting that treatment of vascularized solid PEDs may still reduce some of the exudation and hemorrhage from these lesions. Analysis of PED reflectivity may in some cases be affected by other findings on OCT. As can be seen from Case 3 in Figure 2, in cases of concurrent RPE thinning, increased visualization of the underlying retinal structures including PED content can occur. This may affect the interpretation of PED reflectivity, with more subtle degrees of reflective content within the PED potentially seen in these cases. The hyperreflective material seen lining the surface of the RPE in Case 1 in Figure 2 may also be interpreted in this manner. However, this material is commonly seen in vascularized PEDs,2 and caution should be applied when interpreting this structure as representing overall PED reflectivity. Retinal pigment epithelium tears may complicate treatment of PEDs during treatment with intravitreal anti-VEGF therapy. Larger vascularized PEDs that have a higher intraluminal pressure are at a significantly greater risk of producing RPE tears after antiVEGF therapy, with acute vision loss.3,18 There were no cases of RPE tears, and this may be influenced by the fact that tears have been shown to be more likely to occur in PEDs during the initial stages of anti-VEGF therapy,3,19 and the patients in our study had been chronically treated with anti-VEGF therapy. The limitations of this study include the small number of patients, particularly with either extrafoveal or mixed PEDs, which makes analysis of the response
AFLIBERCEPT AND PEDs IN REFRACTORY AMD BROADHEAD ET AL
of these less common PED subtypes difficult. The lack of a control group also makes comparison of response with aflibercept as compared with other anti-VEGF agents difficult to define. However, the prospective nature of the study and the standardized manner in which all data were collected provides a greater level of evidence than previous studies.4 Additionally, the use of both fluorescein angiography and indocyanine green angiography allowed for the confirmation of vascularized PEDs in all cases in this study, and excluded potential confounding neovascular lesions such as choroidal polyps. Intravitreal aflibercept is effective in reducing PED size in previously treatment-resistant patients, although changes in PED size do not correlate with changes in vision. Pigment epithelial detachments that demonstrate some hyporeflectivity suggestive of active exudation may be more responsive to aflibercept therapy than hyperreflective PEDs, although the clinical implications of this are still unknown. Given the lack of visual response with PED reduction and the risk of adverse events after injection, the treatment and surveillance of PEDs in the management of neovascular AMD requires further research. Monitoring of these lesions for other signs of active neovascularization such as subretinal fluid and intraretinal exudation is necessary. Key words: neovascular age-related macular degeneration, pigment epithelial detachment, aflibercept.
6.
7.
8.
9.
10.
11. 12.
13.
14.
15.
References 1. Holz FG, Pauleikhoff D, Klein R, Bird AC. Pathogenesis of lesions in late age-related macular disease. Am J Ophthalmol 2004;137:504–510. 2. Mrejen S, Sarraf D, Mukkamala SK, Freund KB. Multimodal imaging of pigment epithelial detachment: a guide to evaluation. Retina 2013;33:1735–1762. 3. Doguizi S, Ozdek S. Pigment epithelial tears associated with anti-VEGF therapy: incidence, long-term visual outcome, and relationship with pigment epithelial detachment in age-related macular degeneration. Retina 2014;34:1156–1162. 4. Patel KH, Chow CC, Rathod R, et al. Rapid response of retinal pigment epithelial detachments to intravitreal aflibercept in neovascular age-related macular degeneration refractory to bevacizumab and ranibizumab. Eye (Lond) 2013;27:663– 667; quiz 668. 5. Keane PA, Heussen FM, Ouyang Y, et al. Assessment of differential pharmacodynamic effects using optical coherence
16.
17.
18.
19.
981
tomography in neovascular age-related macular degeneration. Invest Ophthalmol Vis Sci 2012;53:1152–1161. Parodi MB, Iacono P, Papayannis A, et al. Intravitreal ranibizumab for pigment epithelial detachment with subfoveal occult neovascularization: a prospective 24-month case series. Am J Ophthalmol 2013;155:103–108. Punjabi OS, Huang J, Rodriguez L, et al. Imaging characteristics of neovascular pigment epithelial detachments and their response to anti-vascular endothelial growth factor therapy. Br J Ophthalmol 2013;97:1024–1031. Spaide RF. Enhanced depth imaging optical coherence tomography of retinal pigment epithelial detachment in age-related macular degeneration. Am J Ophthalmol 2009;147:644–652. Ying GS, Huang J, Maguire MG, et al. Baseline predictors for one-year visual outcomes with ranibizumab or bevacizumab for neovascular age-related macular degeneration. Ophthalmology 2013;120:122–129. Heier JS, Brown DM, Chong V, et al. Intravitreal aflibercept (VEGF trap-eye) in wet age-related macular degeneration. Ophthalmology 2012;119:2537–2548. Stewart MW, Rosenfeld PJ. Predicted biological activity of intravitreal VEGF trap. Br J Ophthalmol 2008;92:667–668. Papadopoulos N, Martin J, Ruan Q, et al. Binding and neutralization of vascular endothelial growth factor (VEGF) and related ligands by VEGF trap, ranibizumab and bevacizumab. Angiogenesis 2012;15:171–185. Kumar N, Marsiglia M, Mrejen S, et al. Visual and anatomical outcomes of intravitreal aflibercept in eyes with persistent subfoveal fluid despite previous treatments with ranibizumab in patients with neovascular age-related macular degeneration. Retina 2013;33:1605–1612. Chang AA, Li H, Broadhead GK, et al. Intravitreal aflibercept for treatment-resistant neovascular age-related macular degeneration. Ophthalmology 2014;121:188–192. Panos GD, Gatzioufas Z, Petropoulos IK, et al. Effect of ranibizumab on serous and vascular pigment epithelial detachments associated with exudative age-related macular degeneration. Drug Des Devel Ther 2013;7:565–569. Ach T, Hoeh AE, Ruppenstein M, et al. Intravitreal bevacizumab in vascular pigment epithelium detachment as a result of subfoveal occult choroidal neovascularization in age-related macular degeneration. Retina 2010;30:1420–1425. Arora S, McKibbin M. One-year outcome after intravitreal ranibizumab for large, serous pigment epithelial detachment secondary to age-related macular degeneration. Eye (Lond) 2011;25:1034–1038. Introini U, Torres Gimeno A, Scotti F, et al. Vascularized retinal pigment epithelial detachment in age-related macular degeneration: treatment and RPE tear incidence. Graefes Arch Clin Exp Ophthalmol 2012;250:1283–1292. Cunningham ET Jr, Feiner L, Chung C, et al. Incidence of retinal pigment epithelial tears after intravitreal ranibizumab injection for neovascular age-related macular degeneration. Ophthalmology 2011;118:2447–2452.
P
igment epithelium detachment (PED) is a wellknown complication of neovascular age-related macular degeneration (AMD). These lesions occur when the retinal pigment epithelium (RPE) detaches from the underlying Bruch membrane, and are often associated with choroidal neovascularization (CNV).1 Vascularized PEDs occur in conjunction with CNV, with fibrovascular material, or hemorrhage accumulating within the PED.2 These lesions can progress to form disciform scars, resulting in poor visual outcomes. Vascularized PEDs may be complicated by RPE rips and tears.3
Previous studies have shown that PEDs respond poorly to treatments for CNV, including photodynamic therapy4 and intravitreal anti-vascular endothelial growth factor (anti-VEGF) agents.5,6 Some studies have suggested that the reflective properties of PEDs on spectral domain optical coherence tomography (SD-OCT) may predict the response to anti-VEGF therapy.7 Less reflective PEDs have been reported as responding better to anti-VEGF therapies. These reflective properties may relate to differences in the properties of the underlying neovascular complex.8 Pigment epithelial detachment location has also been 975
976
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 5
associated with visual outcomes, and subfoveal PEDs have been associated with worse vision.9 Aflibercept (VEGF-Trap Eye/Eylea; Regeneron, Tarrytown, NY), a newer anti-VEGF agent, has been shown to be effective in treating neovascular AMD10 and is reported to have broader antiangiogenic range of activity than other anti-VEGF therapies.11,12 There is limited evidence to suggest that aflibercept may achieve improvement in structural PED characteristics, and that this may correlate with visual gains, in patients who are resistant to other anti-VEGF therapies.4,13 Therefore, we prospectively assessed the effect of aflibercept on PED characteristics and evaluated the correlation of these changes with visual and anatomical outcomes in patients with treatmentresistant neovascular AMD. Methods The study population, inclusion and exclusion criteria, and methods of a registered, prospective open-label clinical trial evaluating the use of aflibercept in treatment-resistant neovascular AMD have been previously reported.14 Participants from that trial who met the following criteria were included in this study: 1) known CNV secondary to AMD as demonstrated by fluorescein angiography; 2) persistent intraretinal or subretinal fluid on SD-OCT for at least 6 months despite at least 4 intravitreal anti-VEGF injections in the past 6 months; 3) persistent PED on SD-OCT for at least 6 months despite at least 4 antiVEGF injections in the past 6 months; 4) the presence of PED on SD-OCT at baseline visit with maximum PED height .100 mm; and 5) hyperfluorescence on fluorescein angiography contiguous with the PED indicative of vascularization of the PED. All patients were treated with aflibercept (2.0 mg) given as 3 initial loading doses (Weeks 0, 4, and 8) followed by further treatments every 8 weeks (Weeks 16, 24, 32, 40, and 48). All patients underwent review From the *Sydney Institute of Vision Science, Sydney, Australia; †Save Sight Institute, The University of Sydney, Sydney, Australia; and ‡Sydney School of Public Health, The University of Sydney, Sydney, Australia. Supported by Bayer Corporation Global. Paper presented at the Royal Australian College of Ophthalmologists Annual Congress, Hobart, Australia, November 2013. A. A. Chang has acted as a consultant for Alcon, Bayer and Novartis. The other authors do not have any financial/conflicting interests to disclose. G. K. Broadhead and T. Hong contributed equally to this work. Reprint requests: Andrew A. Chang, FRANZCO, PhD, Sydney Retina Clinic and Day Surgery, Level 13, Park House, 187 Macquarie Street, Sydney, New South Wales 2000, Australia; e-mail: [email protected]
1 week after initial injection, then every 4 weeks regardless of the injection status. At each visit, a comprehensive ophthalmic examination was performed including best-corrected visual acuity in Early Treatment of Diabetic Retinopathy Study letters and SD-OCT (Heidelberg Spectralis; Heidelberg Industries, Heidelberg, Germany). Fluorescein angiography and indocyanine green angiography were conducted at the baseline visit to confirm the presence of active neovascularization associated with the PED (Figure 1) and to exclude confounding lesions such as polypoidal vasculopathy. Continuity of OCT scan position was ensured through the use of inbuilt eye tracking and image recognition software (Tru-Track and AutoRescan; Heidelberg Industries, Heidelberg, Germany). Central macular thickness (CMT) was defined as the distance between the internal limiting membrane and Bruch membrane, and manual resegmentation of those layers was performed where necessary.14 The presence of PED resolution and RPE tears was recorded at each visit. Two independent graders reviewed PED dimensions and reflectivity characteristics, discrepancies were adjudicated by the third independent assessor. Pigment epithelial detachments were graded as hollow, solid, or mixed based on their SD-OCT reflectivity according to the method previously described by Punjabi et al7 (Figure 2). The location of PEDs was assessed at baseline and categorized as subfoveal (within 500 mm of fovea) or extrafoveal (.500 mm from fovea). Pigment epithelial detachment size was measured in three maximum dimensions (height, width, and length) at each visit using inbuilt callipers within the SD-OCT software. On OCT scans, PED width was measured as the maximal horizontal PED diameter and PED length was measured as the maximal PED vertical diameter. Height was measured as the vertical distance from Bruch membrane to the RPE border. Length and width were measured as the maximal distance from RPE elevation to RPE elevation in any scan. Statistical analyses were performed using R version 2.15.2 (R Foundation for Statistical Computing, Vienna, Austria). Data were described as mean ± standard deviation or frequency. The effect of baseline characteristics, including injection frequency, total injection number, lesion duration, and angiographic subtype (retinal angiomatous proliferation vs. occult CNV) on PED response was also analyzed. For all correlation analyses and the comparison of change based on PED reflectivity and PED location, change at the 48-week time point was used for analysis. Student’s t-tests and the nonparametric Mann–Whitney U tests were used to compare
AFLIBERCEPT AND PEDs IN REFRACTORY AMD BROADHEAD ET AL
977
Fig. 1. Angiography demonstrating a vascularized extrafoveal PED. A. Fundus photograph, (B) autofluorescence image, (C) earlyphase indocyanine green angiography with a discrete juxtafoveal area of hypofluorescence consistent with a PED, (D) latephase indocyanine green angiography showing patchy central hyperfluorescence, (E) earlyphase fluorescein angiography with a discrete juxtafoveal area of hypofluorescence, (F) latephase fluorescein angiography showing marked hyperfluorescence surrounding the margins of the PED, indicative of a vascularized PED.
continuous outcomes against binary covariates. As both tests gave similar values of P, those of the Mann–Whitney U test are presented. Correlations between continuous variables were assessed with Pearson’s correlations and Spearman rank correlations, showing similar results. Pearson’s correlation coefficients and values of P are presented for more meaningful interpretation of correlation coefficients. The average effect of spacing injection frequency on PED dimensions was assessed with a random effects model with the patient as the grouping factor, using the “lme”
function in the “nlme” package in R. A P , 0.05 was considered to be significant.
Results Forty-three patients who met inclusion criteria were included in this study. Baseline characteristics of the patient cohort are summarized in Table 1. Twenty of the participants (47%) were male, and the mean patient age was 78.7 years. Twenty-one PEDs (49%) were
978
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 5
Fig. 2. Examples of PED subtypes: hollow (top), solid (center), and mixed (bottom) PED.
solid, 13 (30%) were hollow, and 9 (21%) were mixed. Thirty-four PEDs (79%) were subfoveal, and 9 (21%) were extrafoveal. Mean baseline PED height, width, and length was 218.5 ± 144.9 mm, 2,337.7 ± 1,234.3 mm, and 2,580.2 ± 1,715.4 mm, respectively. Response of Pigment Epithelial Detachment to Aflibercept Therapy Pigment epithelial detachment height decreased by an average of −51.5 mm at Week 24 and −53.3 mm (range, −323 to 89 mm) at Week 48 compared with baseline (P , 0.01 for all follow-up visits). Pigment epithelial detachment width showed a similar trend, decreasing by an average of −320.6 mm at Week 24 and −276.8 mm (range, −2,798 to 438 mm) at Week 48 (P , 0.05 for all follow-up visits). Pigment epithelial detachment length decreased by −248.8 mm at Week 24 and −301.9 mm (range, −2084 to 439 mm) at Week 48 (P , 0.05 for all follow-up visits) (Figure 3). ComTable 1. Background Characteristics of the Patient Cohort Characteristic Male Age at baseline, years Right eye Mean total Lucentis injections Mean PED duration, months Pseudophakic PED type Solid Hollow Mixed PED location Subfoveal Extrafoveal
Frequency/Mean, n (%) 20 (47) 78.7 21 (49) 34.8 32.1 26 (60.0) 21 (49) 13 (30) 9 (21) 34 (79) 9 (21)
plete PED resolution occurred in 4 patients by the end of the 12-month period, and 42% (18 of 43) patients had PED improvement (.50 mm improvement), 56% (24 of 43) had stable PEDs, and 2% (1 of 43) had worsening (.50 mm increase) of their PED at Week 48. There was no difference in response in any dimension between those receiving anti-VEGF therapy every 4 weeks and every 6 weeks (P . 0.05), or between those with retinal angiomatous proliferation lesions and those with occult CNV related to vascularized PEDs (P . 0.05). No significant difference was seen in response of PED height or length between those who received previous ranibizumab monotherapy and those who received both previous ranibizumab and bevacizumab (P . 0.05), however, a difference was seen in response of PED width (difference of 311 mm response at 48 weeks, P = 0.03) between these 2 groups. Spacing Spacing of injection frequency from monthly to every 2 months resulted in significant increases in PED height, width, and length (average increase in height width and length for 1 month after injection compared with 2 months after injection: 20.1, 135.6, and 91.9 mm, respectively, P , 0.05 for all 3 dimensions). This oscillating pattern of change in PED dimensions in response to spacing of injection frequency is seen in Figure 3. Correlation With Central Macular Thickness and Vision Changes in PED height were correlated with changes in CMT (R2 = 0.36, P , 0.001). This correlation became stronger when only subfoveal PEDs were considered (R2 = 0.64, P , 0.001). Reduction in PED width was also correlated with CMT change (R2 = 0.25, P , 0.001), as were changes in PED length (R2 = 0.16, P , 0.01). The association between both PED width and CMT, as well as PED length and CMT, became less significant when only subfoveal PEDs were considered (R2 = 0.09, P = 0.10 for width, R2 = 0.12, P = 0.05 for length). Best-corrected visual acuity gains of 6.8 and 4.6 letters were observed at Weeks 24 and 48, respectively (P , 0.001 for both). Approximately 47% of patients had a visual gain of $5 letters at Week 48. We observed minor associations between reflectivity characteristics, duration of PED, baseline best-corrected visual acuity, and changes in PED height, width, or length with changes in best-corrected visual acuity, but none of these were statistically significant (P . 0.10 for all).
AFLIBERCEPT AND PEDs IN REFRACTORY AMD BROADHEAD ET AL
979
Fig. 3. Changes in PED height, width, and length compared with baseline examination over 48 weeks.
Pigment Epithelial Detachment Reflectivity Hollow (less reflective) PEDs exhibited a significant reduction in all 3 PED dimensions at Week 48 compared with baseline (mean change in height, width, and length: 70.8, 516.3, and 479.3 mm, respectively, P # 0.01 for all 3 dimensions). Mixed PEDs had a significant reduction in PED height and length (mean change in height and length: 113.2 and 649.6 mm, respectively, P = 0.02 for both dimensions), but not for width (mean change: 551.2 mm, P = 0.11). Solid (more reflective) PEDs had a significant reduction only in PED height (mean change: 16.9 mm, P = 0.048), but not for either length or width (mean change in length and width: 30.2 and 23.4 mm, respectively, P . 0.10 for both dimensions). Significant differences were seen in response to therapy based on the PED subtype, with both mixed and hollow PEDs significantly more likely to experience reductions in almost all 3 PED dimensions compared with solid PEDs (P , 0.05 for hollow or mixed versus solid PEDs for all 3 dimensions, excluding height in hollow to solid [P = 0.05] and width in mixed to solid [P = 0.099]). There was no difference seen in response between hollow and mixed PEDs (P . 0.10 for mixed versus hollow for all 3 dimensions). Pigment Epithelial Detachment Location and Retinal Pigment Epithelium Tears Subfoveal PEDs (79%) were more frequently observed than extrafoveal PEDs (21%; P , 0.001). The reduction in any of the PED dimensions was not associated with PED location (P . 0.10 for all
3 dimensions). No cases of RPE rips or tears were observed during the study period. Grading Variability Two independent graders showed a high level of agreement (mean difference and limits of agreement for height, width, and length were −1.64 mm [−27.7 to 24.4 mm], 7.3 mm [−182 to 196 mm], and 8.6 mm [−342 to 360 mm], respectively, Cohen’s Kappa on reflectivity characteristic = 0.853). Discussion In this study, aflibercept was found to effectively reduce PED size in all dimensions of vascularized PEDs that had been resistant to previous anti-VEGF therapy. In a few cases (4 of 43), complete PED resolution also occurred. Cases of PED resolution and improvement in PED dimensions with aflibercept for treatment-resistant PEDs have previously been reported.4,13 It is interesting to note that the reduction in both PED height and width observed in our study was similar to that seen in the retrospective study of aflibercept by Kumar et al13 in a similar treatmentresistant cohort (mean PED height change: 36 mm, mean width change: 316 mm). Reported changes in PED height in response to other anti-VEGF agents vary considerably. In newly diagnosed vascularized PEDs, Panos et al15 observed a change in a mean PED height of 135 mm after pro re nata ranibizumab for 12 months, whereas Ach et al16 showed a nonsignificant decrease in PED height (mean height decrease: 81.7 mm, P . 0.05) with bevacizumab for
980
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 5
vascularized PEDs. It is not clear why differences in PED height reduction have been observed, although in both this study and the study by Kumar et al, it may relate to the treatment-resistant nature of patients. Part of the discrepancy may also relate to differences in measurement techniques. In our study, we observe low bias across graders with a 1.6-mm average difference in PED height measurements between the 2 graders. The variability of the differences was also low, with 95% of disagreements ,30 mm. There may also be limitations with reliability of the images captured by the OCT, although the use of manual measurements by 2 graders was undertaken to minimize any errors that may arise because of automated measurements. However, the mean change itself is still small (53 mm), and as such it is unclear whether a change of this magnitude is a clinically meaningful outcome. The correlation between changes in PED height and CMT is to be expected, given that PED height forms a component of the CMT. As such, CMT may potentially be a surrogate marker of changes in PED architecture in response to aflibercept therapy. The lack of correlation between improvements in PED dimensions and significant additional visual improvement observed in our study is consistent with previous reports of PED response to anti-VEGF therapy.6,7,15,16 The visual improvement observed in these patients is therefore likely to be a result of improvement in other factors, such as intraretinal fluid or subretinal fluid, rather than reduction in PED size. The degree of visual gain obtained decreased slightly at Week 48 compared with the gain at Week 24 (2 letters), however a decline of this magnitude is not clinically significant. The reason for this decline is unknown. It is possible that the 2-monthly injection frequency, as opposed to the initial monthly loading therapy, contributed to this reduction of visual gain. However, other factors such as the longer follow-up duration (1 year) may have also impacted on the visual outcomes achieved. Vision loss associated with PEDs thus seems to be largely nonreversible, even with structural reduction of the lesion. This lack of correlation between PED reduction and vision may make changes in PED size less useful as a clinical marker of overall treatment response. However, it is clear that these sites bear close monitoring, because they are sites over which other lesions associated with visual impairment, such as subretinal fluid and intraretinal fluid can accumulate. Previous studies have suggested that PED reflectivity may be an important characteristic in response to therapy.7,17 We observed similar findings, with the response of PEDs to aflibercept therapy differing based on OCT reflectivity. Pigment epithelial detachments with significant hyporeflectivity (hollow or mixed)
were more responsive than hyperreflective (solid) PEDs. It has been suggested that the hyporeflective components may reflect the presence of fluid exudate, and the hyperreflective component of mixed and solid PEDs represents fibrinous leakage or fibrovascular proliferation, suggesting active neovascularization.2,8 The increased response of lesions with some hyporeflective component may thus relate to a reduction in the exudative component of these reflective subtypes as a result of inhibition of VEGF-driven vasodilation and vascular leakage. Hyperreflective PEDs may sometimes involve a significant lipid or fibrous component in addition to a neovascular membrane.2 These materials would be expected to be less responsive to aflibercept therapy than exudative substances such as fluid or blood. The presence of such material in some of the solid PEDs in this study would explain why these lesions had smaller changes in response to aflibercept therapy. Mixed PEDs may have shown a greater response to aflibercept than solid lesions because they contain some exudative components. It is interesting that some improvement was still noted in solid PED height, suggesting that treatment of vascularized solid PEDs may still reduce some of the exudation and hemorrhage from these lesions. Analysis of PED reflectivity may in some cases be affected by other findings on OCT. As can be seen from Case 3 in Figure 2, in cases of concurrent RPE thinning, increased visualization of the underlying retinal structures including PED content can occur. This may affect the interpretation of PED reflectivity, with more subtle degrees of reflective content within the PED potentially seen in these cases. The hyperreflective material seen lining the surface of the RPE in Case 1 in Figure 2 may also be interpreted in this manner. However, this material is commonly seen in vascularized PEDs,2 and caution should be applied when interpreting this structure as representing overall PED reflectivity. Retinal pigment epithelium tears may complicate treatment of PEDs during treatment with intravitreal anti-VEGF therapy. Larger vascularized PEDs that have a higher intraluminal pressure are at a significantly greater risk of producing RPE tears after antiVEGF therapy, with acute vision loss.3,18 There were no cases of RPE tears, and this may be influenced by the fact that tears have been shown to be more likely to occur in PEDs during the initial stages of anti-VEGF therapy,3,19 and the patients in our study had been chronically treated with anti-VEGF therapy. The limitations of this study include the small number of patients, particularly with either extrafoveal or mixed PEDs, which makes analysis of the response
AFLIBERCEPT AND PEDs IN REFRACTORY AMD BROADHEAD ET AL
of these less common PED subtypes difficult. The lack of a control group also makes comparison of response with aflibercept as compared with other anti-VEGF agents difficult to define. However, the prospective nature of the study and the standardized manner in which all data were collected provides a greater level of evidence than previous studies.4 Additionally, the use of both fluorescein angiography and indocyanine green angiography allowed for the confirmation of vascularized PEDs in all cases in this study, and excluded potential confounding neovascular lesions such as choroidal polyps. Intravitreal aflibercept is effective in reducing PED size in previously treatment-resistant patients, although changes in PED size do not correlate with changes in vision. Pigment epithelial detachments that demonstrate some hyporeflectivity suggestive of active exudation may be more responsive to aflibercept therapy than hyperreflective PEDs, although the clinical implications of this are still unknown. Given the lack of visual response with PED reduction and the risk of adverse events after injection, the treatment and surveillance of PEDs in the management of neovascular AMD requires further research. Monitoring of these lesions for other signs of active neovascularization such as subretinal fluid and intraretinal exudation is necessary. Key words: neovascular age-related macular degeneration, pigment epithelial detachment, aflibercept.
6.
7.
8.
9.
10.
11. 12.
13.
14.
15.
References 1. Holz FG, Pauleikhoff D, Klein R, Bird AC. Pathogenesis of lesions in late age-related macular disease. Am J Ophthalmol 2004;137:504–510. 2. Mrejen S, Sarraf D, Mukkamala SK, Freund KB. Multimodal imaging of pigment epithelial detachment: a guide to evaluation. Retina 2013;33:1735–1762. 3. Doguizi S, Ozdek S. Pigment epithelial tears associated with anti-VEGF therapy: incidence, long-term visual outcome, and relationship with pigment epithelial detachment in age-related macular degeneration. Retina 2014;34:1156–1162. 4. Patel KH, Chow CC, Rathod R, et al. Rapid response of retinal pigment epithelial detachments to intravitreal aflibercept in neovascular age-related macular degeneration refractory to bevacizumab and ranibizumab. Eye (Lond) 2013;27:663– 667; quiz 668. 5. Keane PA, Heussen FM, Ouyang Y, et al. Assessment of differential pharmacodynamic effects using optical coherence
16.
17.
18.
19.
981
tomography in neovascular age-related macular degeneration. Invest Ophthalmol Vis Sci 2012;53:1152–1161. Parodi MB, Iacono P, Papayannis A, et al. Intravitreal ranibizumab for pigment epithelial detachment with subfoveal occult neovascularization: a prospective 24-month case series. Am J Ophthalmol 2013;155:103–108. Punjabi OS, Huang J, Rodriguez L, et al. Imaging characteristics of neovascular pigment epithelial detachments and their response to anti-vascular endothelial growth factor therapy. Br J Ophthalmol 2013;97:1024–1031. Spaide RF. Enhanced depth imaging optical coherence tomography of retinal pigment epithelial detachment in age-related macular degeneration. Am J Ophthalmol 2009;147:644–652. Ying GS, Huang J, Maguire MG, et al. Baseline predictors for one-year visual outcomes with ranibizumab or bevacizumab for neovascular age-related macular degeneration. Ophthalmology 2013;120:122–129. Heier JS, Brown DM, Chong V, et al. Intravitreal aflibercept (VEGF trap-eye) in wet age-related macular degeneration. Ophthalmology 2012;119:2537–2548. Stewart MW, Rosenfeld PJ. Predicted biological activity of intravitreal VEGF trap. Br J Ophthalmol 2008;92:667–668. Papadopoulos N, Martin J, Ruan Q, et al. Binding and neutralization of vascular endothelial growth factor (VEGF) and related ligands by VEGF trap, ranibizumab and bevacizumab. Angiogenesis 2012;15:171–185. Kumar N, Marsiglia M, Mrejen S, et al. Visual and anatomical outcomes of intravitreal aflibercept in eyes with persistent subfoveal fluid despite previous treatments with ranibizumab in patients with neovascular age-related macular degeneration. Retina 2013;33:1605–1612. Chang AA, Li H, Broadhead GK, et al. Intravitreal aflibercept for treatment-resistant neovascular age-related macular degeneration. Ophthalmology 2014;121:188–192. Panos GD, Gatzioufas Z, Petropoulos IK, et al. Effect of ranibizumab on serous and vascular pigment epithelial detachments associated with exudative age-related macular degeneration. Drug Des Devel Ther 2013;7:565–569. Ach T, Hoeh AE, Ruppenstein M, et al. Intravitreal bevacizumab in vascular pigment epithelium detachment as a result of subfoveal occult choroidal neovascularization in age-related macular degeneration. Retina 2010;30:1420–1425. Arora S, McKibbin M. One-year outcome after intravitreal ranibizumab for large, serous pigment epithelial detachment secondary to age-related macular degeneration. Eye (Lond) 2011;25:1034–1038. Introini U, Torres Gimeno A, Scotti F, et al. Vascularized retinal pigment epithelial detachment in age-related macular degeneration: treatment and RPE tear incidence. Graefes Arch Clin Exp Ophthalmol 2012;250:1283–1292. Cunningham ET Jr, Feiner L, Chung C, et al. Incidence of retinal pigment epithelial tears after intravitreal ranibizumab injection for neovascular age-related macular degeneration. Ophthalmology 2011;118:2447–2452.
