Purpose
To study the effect of various baseline factors, particularly the type of drug (ranibizumab vs aflibercept), on the functional and anatomic response of treatment-naïve pigment epithelial detachment (PED) associated with neovascular age-related macular degeneration (neovascular AMD), after 3 intravitreal injections.
Design
Retrospective consecutive case series.
Methods
This study included 102 patients (n = 115 eyes) with treatment-naïve neovascular AMD and PED (>150 μm), who were treated with either ranibizumab (n = 68 eyes) or aflibercept (n = 47 eyes). A multivariate analysis using stepwise linear regression was performed in order to assess factors influencing visual acuity improvement, as well as treatment response of PED height after 3 monthly injections.
Results
Multivariate analysis revealed that better visual improvement was associated with lower best-corrected visual acuity (BCVA) at baseline ( P = .001), presence of subretinal fluid ( P = .001), and retinal angiomatous proliferation ( P = .001); PED reduction was associated with higher PED at baseline ( P = .001), predominantly serous PED ( P = .003), and the use of aflibercept ( P = .022). Drug type was not associated with change in BCVA at 3 months.
Conclusion
Eyes with neovascular AMD and PED showed significant functional and anatomic response after 3 monthly intravitreal anti-VEGF injections. The functional response depended on baseline BCVA, presence of subretinal fluid, and retinal angiomatous proliferation, while anatomic response was influenced by baseline PED height, degree of vascularization, and drug type. Drug type was not associated with change in BCVA, but had a weak effect on anatomic response.
The current gold-standard treatment for neovascular age-related macular degeneration (AMD) using intravitreal injections of anti–vascular endothelial growth factor (anti-VEGF) has been validated by several multicenter trials. Ranibizumab and aflibercept are 2 different anti-VEGF drugs that have shown equal efficacy in terms of visual acuity and central retinal thickness improvement. However, these 2 drugs have different pharmacologic characteristics, which may have an effect on treatment responses in specific subgroups.
Ranibizumab (Lucentis; Novartis Pharma AG, Basel, Switzerland) is a recombinant, humanized monoclonal antibody Fab fragment that neutralizes all isoforms of VEGF-A. Aflibercept (Eylea; Bayer Schweiz AG, Zurich, Switzerland) is a recombinant fusion protein composed of an Fc domain fused to the VEGF-binding domains of VEGF receptors 1 and 2. In addition to binding VEGF-A, aflibercept also binds VEGF-B and placental growth factor (PlGF).
Although it is known that VEGF-A is a key player in the pathogenesis of neovascular AMD, the role of the different members of the VEGF family in the development of neovascular AMD is only partially understood. However, there is some evidence that PlGF may also be implicated in the pathogenesis of choroidal neovascularization (CNV).
While VEGF-A binds to VEGF receptors 1 and 2, PlGF activates only receptor 1. These receptors are unequally distributed in the retinal and choroidal layers, suggesting that the exudative phenotypes might potentially be influenced by different patterns of receptor activation. Thus, the different spectrum of anti-VEGF drugs might result in different responses depending on the phenotype. In fact, some reports have suggested that aflibercept may have a stronger effect on pigment epithelium detachment (PED), as described for ranibizumab refractory cases.
The aim of the present study was to assess factors that influence the functional and structural responses of neovascular AMD with PED, with a particular focus on the effect of drug type (ie, ranibizumab vs aflibercept).
Methods
This is a retrospective consecutive case series that was performed in the Jules-Gonin University Eye Hospital in Lausanne, Switzerland. The study was approved by the Swiss Federal Department of Health for retrospective data analysis, and was performed in accordance with the ethical standards of the Declaration of Helsinki.
In this study, we identified a consecutive series of eyes with treatment-naïve active neovascular AMD with PED that were treated with intravitreal anti-VEGF injections between January 1, 2009 and January 31, 2014. Inclusion criteria were a minimal PED height of 150 μm observed in optical coherence tomography (OCT), good imaging quality, and a full loading dose of 3 monthly intravitreal injections of the same drug (ie, ranibizumab or aflibercept). The reason for using either ranibizumab or aflibercept was not based on clinical judgment, but rather was related to insurance approval and reimbursement of treatment cost. Exclusion criteria were poor image quality, presence of a pigment epithelium tear at baseline, polypoidal choroidal vasculopathy or any other confounding retinopathies, or use of other adjuvant treatment modalities.
Baseline examination and 3-month follow-up visits included the evaluation of best-corrected visual acuity (BCVA) using a letter count on the Early Treatment of Diabetic Retinopathy Study (ETDRS) chart, anterior segment examination with measurement of intraocular pressure, a dilated fundus examination, an OCT examination, a fluorescein angiography (FA), and an indocyanine green angiography (ICGA) (TRC-501X; Topcon, Tokyo, Japan); the latter was performed usually only at baseline. OCT with cube examination was performed using Spectralis (6 mm, 49 lines; Heidelberg Engineering, Heidelberg, Germany) or SD-OCT Cirrus (128 × 512 cube examination; Carl Zeiss Meditec, Inc, Oberkochen, Germany), and the same machine was used at baseline and after 3 months.
Baseline characteristics were recorded for age, sex, angiographic type of CNV, maximum and subfoveal PED height, size of CNV lesion and PED, degree of vascularization of PED on ICGA, and OCT reflectivity under PED. The following data were recorded manually, with the grader being masked for drug type, using the caliper provided by the manufacturer’s software: maximum PED height (ie, the distance between the point just underneath the hyperreflective pigment epithelium band and the Bruch membrane), subfoveal PED height (ie, the distance between the point just underneath the hyperreflective pigment epithelium band and the Bruch membrane at the foveal center), and central retinal thickness (CRT; ie, the distance between the foveal inner limiting membrane and the top of the pigment epithelium layer at the foveal center). We also assessed the presence or absence of intra- or subretinal fluid in the macular area on OCT. The OCT reflectivity under the PED was classified as optically empty/patchy or dense. The degree of PED vascularization was categorized as predominantly serous if less than 50% of the lesion was vascularized, or predominantly vascularized if 50% or more was vascularized, based on ICGA results. The size of PED and CNV lesion were determined based on multimodal imaging studies (ie, FA and ICGA), and were measured as disc areas (DA).
Our primary outcome measure was the change of BCVA at month 3. The secondary outcome measure was the change of maximum PED height at month 3.
Evaluations at baseline and after 3 months were compared using paired t tests. Univariate analyses were performed using Pearson correlation, t test, and analysis of variance to determine baseline factors associated with changes of BCVA and maximum PED height. Stepwise multivariate linear regression analysis was performed to evaluate the effect of baseline predictors on changes of BCVA and maximum PED height after 3 months (ie, dependent variables). Drug type, baseline demographic characteristics (ie, age, sex), and factors that were associated with changes of BCVA and maximum PED height in the univariate analysis ( P < .2) were included in the multiple regression analysis. For data analysis, a Microsoft Excel 2010 spreadsheet and SPSS software for Windows (version 17.0; SPSS, Inc, Chicago, Illinois, USA) were used. A 2-tailed probability of .05 or less was considered statistically significant.
Results
In this study we assessed 115 eyes of 102 patients. The mean age of the patients was 80 ± 7.8 years, and the male-to-female ratio was 0.44. The angiographic type of CNV at baseline was occult in 61.7% (71 of 115), minimally classic in 19.1% (22 of 115), classic in 4.3% (5 of 115), and retinal angiomatous proliferation (RAP) in 14.8% (17 of 115) of the eyes. At baseline, the mean size of CNV was 3.2 ± 1.5 DA, and the mean size of PED was 1.9 ± 1.1 DA. On ICGA, PEDs were predominantly serous in 45.2% (52 of 115), and predominantly vascularized in 54.8% (63 of 115) of the eyes. The OCT reflectivity under the PED was empty/patchy in 62.6% (72 of 115), and dense in 37.4% of the eyes (43 of 115). Intravitreal injections of ranibizumab or aflibercept were performed in 68 and 47 eyes, respectively.
After 3 loading intravitreous injections, there was a significant improvement in BCVA (60 ± 17 ETDRS letters at baseline vs 67 ± 16 ETDRS letters at month 3; P = .001), maximum PED height (368 ± 201 μm at baseline vs 248 ± 172 μm at month 3; P = .001), subfoveal PED height (199 ± 218 μm at baseline vs 124 ± 158 μm at month 3; P = .001), and central retinal thickness (328 ± 170 μm at baseline vs 198 ± 86 μm at month 3; P = .006).
We assessed baseline characteristics associated with BCVA change after the loading dose using univariate analysis. We found that only CNV lesion type and baseline BCVA had a significant effect on BCVA change at 3 months ( P = .025 and P = .001, respectively) ( Table 1 ).
| Categorical Variables | Maximal PED Height | BCVA | |||
|---|---|---|---|---|---|
| Maximum PED Height Decrease | P Value | BCVA Increase | P Value | ||
| Sex | Male (n = 34) | 92 ± 181 | .313 | 5.6 ± 8.9 | .251 |
| Female (n = 81) | 132 ± 191 | 7.9 ± 9.6 | |||
| Drug type | Ranibizumab (n = 67) | 116 ± 203 | .788 | 6.2 ± 10.1 | .170 |
| Aflibercept (n = 48) | 126 ± 164 | 8.7 ± 8.3 | |||
| Angiographic type of CNV | Occult (n = 71) | 109 ± 211 | .467 | 6.4 ± 9.2 | .025 |
| Miniclassic (n = 22) | 157 ± 157 | 4.6 ± 9.4 | |||
| Classic (n = 5) | 14 ± 187 | 11.0 ± 2.8 | |||
| RAP (n = 17) | 146 ± 99 | 13.2 ± 9.7 | |||
| Vascularization on ICGA | Serous (n = 52) | 161 ± 202 | .035 | 8.1 ± 9.4 | .333 |
| Vascularized (n = 63) | 86 ± 169 | 6.4 ± 9.5 | |||
| OCT reflectivity under PED | Empty or patchy (n = 72) | 146 ± 213 | .041 | 6.8 ± 9.6 | .551 |
| Dense (n = 43) | 78 ± 130 | 7.9 ± 9.3 | |||
| Foveal cyst on baseline OCT | Yes (n = 56) | 138 ± 150 | .345 | 6.9 ± 10 | .745 |
| No (n = 59) | 105 ± 215 | 7.5 ± 8.6 | |||
| Foveal SRF on baseline OCT | Yes (n = 104) | 120 ± 196 | .972 | 7.6 ± 9.5 | .155 |
| No (n = 11) | 118 ± 101 | 3.4 ± 7.9 | |||
| Continuous Variables | Maximum PED Height Change Correlation | BCVA Change Correlation | ||
|---|---|---|---|---|
| Correlation Coefficient r | P Value | Correlation Coefficient r | P Value | |
| Age | −0.09 | .367 | 0.04 | .678 |
| Size of CNV (disc areas) | 0.23 | .017 | 0.08 | .405 |
| Size of PED (disc areas) | 0.30 | .002 | 0.15 | .111 |
| Central macular thickness | −0.16 | .087 | 0.072 | .453 |
| Maximum PED height at baseline | 0.61 | .001 | −0.09 | .347 |
| Subfoveal PED height at baseline | 0.40 | .001 | −0.09 | .361 |
| Baseline BCVA (ETDRS letters) | 0.02 | .811 | −0.37 | .001 |
For the multivariate analysis of predictors of change in BCVA at 3 months, the variables included in the model were age, sex, drug type, angiographic type of CNV, size of PED, presence of subretinal fluid (SRF) on OCT at baseline, and baseline BCVA. The overall regression was significant, and the R 2 value was 0.28. Table 2 summarizes the last step of the regression model. BCVA change was significantly associated with baseline BCVA, presence of SRF on baseline OCT, and RAP type of CNV. If all other variables were kept constant, eyes with RAP type of CNV would have a BCVA change of 10.5 letters greater than other CNV lesions ( P = .001); while eyes with SRF on baseline OCT would have a BCVA change of 9 letters greater than eyes with no SRF ( P = .001). Baseline BCVA was negatively correlated with change in BCVA (r = −0.22, P = .001).
| Variables | Nonstandardized Coefficient | Standard Error | P Value |
|---|---|---|---|
| Baseline BCVA | −0.22 | 0.05 | .001 |
| SRF in baseline OCT | 9.1 | 2.4 | .001 |
| CNV type (RAP) | 10.5 | 2.8 | .001 |
We also assessed baseline characteristics associated with maximum PED height decrease after the loading dose using univariate analysis. We found that eyes with predominantly serous PED, and eyes with empty/patchy OCT reflectivity under the PED, were associated with higher maximum PED decrease after 3 months ( P = .035 and P = .041, respectively). The maximum PED height decrease was also correlated with maximum PED height at baseline (r = 0.61, P = .001), size of CNV (r = 0.23, P = .017), and size of PED (r = 0.30, P = .002) ( Table 1 ).
For the multivariate analysis of predictors of maximum PED height decrease, the variables included in the model were age, sex, drug type, PED vascularization on ICGA, OCT reflectivity under PED, size of CNV, size of PED, central foveal thickness, and maximum and subfoveal PED height. The overall regression was significant, and the R 2 value was 0.45. Table 3 summarizes the last step of the regression model. Decrease of maximum PED height was significantly associated with drug type, type of PED vascularization on ICGA, and baseline maximum PED height. If all other variables were kept constant, eyes with serous PED vascularization on ICGA would have a reduction of the maximum PED height that would be 83 μm greater than eyes with vascularized PED ( P = .003); and eyes treated with aflibercept would have a reduction of the maximum PED height that would be 66 μm greater than the ranibizumab-treated eyes ( P = .022). The maximum PED height at baseline was strongly and positively correlated with maximum PED height decrease; for every 10 μm of maximum PED height at baseline, the reduction of PED height after 3 months was approximately 6 μm greater ( P = .001) ( Table 3 ).
