To evaluate the incidence and associated risk factors of retinal pigment epithelium (RPE) tears after intravitreal ranibizumab injection treatment for retinal angiomatous proliferation (RAP).
Retrospective, observational case series.
Ninety-eight treatment-naïve RAP eyes (86 patients) that received intravitreal ranibizumab injections were included. All patients were treated with an initial series of 3 monthly loading injections, followed by further injections as required. Baseline characteristics and features were evaluated as potential risk factors for RPE tearing. The visual and anatomic outcomes after treatment were evaluated at 12 months from baseline.
RPE tears had developed in 8.2% (8) of the eyes by 12 months. Of these, all had pigment epithelium detachment (PED) at baseline, and the majority (7) had developed an RPE tear within 3 months. Multiple regression analysis showed that higher PED (odds ratio [OR] = 1.411, 95% confidence interval [CI] = 1.199–1.888, P = .011) and thinner choroid (OR = 0.815, 95% CI = 0.719–0.912, P = .023) were associated with a higher risk of RPE tearing. The mean best-corrected visual acuity of the patients with RPE tearing (0.56 ± 0.49 logarithm of the minimal angle of resolution [logMAR]) was significantly worse at 12 months than that of patients without RPE tearing (0.74 ± 0.55 logMAR, P = .009) after treatment.
RPE tears developed in 8.2% of eyes with RAP during the 12 months following ranibizumab injections. Higher PED height and thinner subfoveal choroidal thickness were associated with the development of RPE tears after ranibizumab treatment for RAP.
Retinal angiomatous proliferation (RAP) is a clinical entity generally categorized as a subtype of neovascular age-related macular degeneration (nAMD). RAP involves the proliferation of deep retinal capillaries; at later stages, this neovascularization progresses from subretinal to choroidal, with pigment epithelial detachment (PED). However, the origin of the neovascular process (intraretinal vs choroidal) remains controversial. The prevalence of RAP in newly diagnosed nAMD has been reported to be 5%–15%.
A tear to the retinal pigment epithelium (RPE) is one of the more serious complications of nAMD. Although RPE tears can occur during the natural disease progression of nAMD, they are also seen after therapeutic interventions such as photodynamic therapy (PDT) or intravitreal injections with anti–vascular endothelial growth factor (VEGF) substances such as pegaptanib, bevacizumab, aflibercept, and ranibizumab. RPE tears after ranibizumab treatment have been reported in 2%–6% of eyes with nAMD, and in as much as 12%–25% of eyes with vascularized PED. When the fovea is involved in the RPE tear, visual prognosis is markedly poorer.
In recent times, intravitreal injection of anti-VEGF drugs has been used extensively for the treatment of nAMD. Because RAP is considered to be a distinct subtype of nAMD, anti-VEGF injections are commonly used. However, owing to its relative infrequency, not much is known about RPE tearing following intravitreal anti-VEGF treatment for RAP. The purpose of this study was to identify the characteristics of RPE tearing after ranibizumab injections in RAP patients, the associated risk factors, and the prognosis.
We performed a computerized search and medical record review for patients who had been diagnosed with RAP and treated with anti-VEGF injections (ranibizumab [Lucentis; Novartis Pharma AG, Basel, Switzerland]) between June 2012 and June 2014. All patients had been examined and treated at the Retina Center of Kim’s Eye Hospital at Konyang University College of Medicine. This study was approved by the Institutional Review Board of Kim’s Eye Hospital, Konyang University College of Medicine. The study adhered to the tenets of the Declaration of Helsinki (IRB No A-2015-012).
The following inclusion criteria were used: (1) age >50 years; (2) confirmation of RAP using funduscopy, spectral-domain optical coherence tomography (SD OCT [Spectralis; Heidelberg Engineering, Heidelberg, Germany; and Spectral OCT/SLO; OTI Ophthalmic Technologies Inc, Miami, Florida, USA]), fluorescein angiography (FA), and indocyanine green angiography (ICGA) performed using a confocal laser scanning system (Spectralis HRA+OCT; Heidelberg Engineering) at the first visit; (3) treatment naïve; (4) treatment with ranibizumab; and (5) a minimum follow-up period of 12 months.
The diagnosis of RAP was based on the characteristic features of the condition, which include intraretinal hemorrhage, intraretinal vascular anastomoses, and PED with overlying cystic retinal edema, observed using OCT. In addition, the presence of retinal anastomotic feeder and drainage vessels on the FA or ICGA images, or of a “hot spot” corresponding to the neovascular lesion on the ICGA image, was also required in order for RAP to be diagnosed. Patients with RAP were classified as stage 1, 2, or 3 according to established grading criteria. The diagnosis and staging of RAP were evaluated by 2 independent investigators (H.J.C. and H.S.K.). When the evaluation was inconsistent, a senior investigator (T.G.L.) made the final decision.
Exclusion criteria were as follows: (1) the presence of RPE tearing at baseline; (2) treatment using another anti-VEGF agent (bevacizumab [Avastin; Genentech Inc, South San Francisco, California, USA]) or photodynamic therapy; (3) concomitant ocular diseases such as diabetic retinopathy, high myopia (more than 6 diopter spherical equivalent), vein or artery occlusion, or epiretinal membrane; and (4) trauma either during the study or in the contralateral eye, aphakia, or previous vitreoretinal surgery.
Assessment and Outcome Measures
RPE tears were identified using 4 imaging modalities: color fundus photograph, FA and/or ICGA images, SD OCT, and autofluorescence images. FA and/or ICGA images revealed hypofluorescence of the bare choroid, and all suspicious RPE tear-like lesions were confirmed by SD OCT consisting of either 19 or 31 horizontal lines (6 × 6 mm area); OCT showed interruption of the hyperreflective RPE layer, with elevation or scrolling of the torn RPE flap. Increased depth of signals, corresponding to the bare choroid underlying the torn RPE, was frequently seen on OCT.
When the patient had PED at baseline, PED height was defined as the vertical distance from the hyperreflective line of the Bruch membrane to the inner margin of the hyperreflective line of the RPE layer. All OCT scans were taken horizontally over the PED (consisting of either 19 or 31 horizontal lines [6 × 6 mm area]), such that the highest point of the PED was assessed. Subfoveal choroidal thickness was defined as the vertical distance from the hyperreflective line of the Bruch membrane to the hyperreflective line of the inner surface of the sclera on fovea-centered SD OCT images. When the inner surface of the sclera was not visible on the conventional image, the OCT image was taken using an enhanced depth imaging (EDI) technique, a feature of the Spectralis SD OCT machine, as previously reported. Each section was obtained using eye tracking, and 100 scans were averaged in order to improve the signal-to-noise ratio. The lesion was defined as choroidal neovascularization (CNV) lesion included CNV and 1 or more of the following in or adjacent to the location of CNV: PED, blocked fluorescence, or hemorrhage. Hemorrhage was considered to be part of the lesion only when it was contiguous with the total neovascular lesion and the hemorrhage extended beyond the fluorescence of the underlying CNV on FA. The lesion size was measured manually by outlining the lesion on the FA image. All measurements were estimated using the Heidelberg Eye Explorer software (v. 22.214.171.124; Heidelberg Engineering) by 2 retinal specialists (H.J.C. and H.S.K.) who had been masked to patient information.
As a measurement of functional outcome after treatment, the best-corrected visual acuity (BCVA) 12 months after treatment of patients with RPE tearing was compared with that of patients without RPE tearing. In addition, anatomic outcome, measured as central foveal thickness (defined as the vertical distance from the internal limiting membrane [ILM] to Bruch membrane on SD OCT images) after treatment was also compared between groups. In addition, the percentage of patients who had gained or lost more than 3 lines of vision since baseline was also assessed.
Intravitreal Ranibizumab Treatment
All patients had been given a loading dose of 3 intravitreal ranibizumab injections (0.5 mg/0.05 mL; Lucentis; Novartis Pharma AG, Basel, Switzerland) with a 1-month interval between treatments. Subsequent additional injections were given if any of the following conditions were observed: (1) visual deterioration of more than 2 lines (>0.2 logarithm of the minimal angle of resolution [logMAR]); (2) OCT evidence of persistent fluid or hemorrhage involving the macula for at least 1 month following the previous injection; (3) an increase in central foveal thickness of at least 100 μm on OCT; (4) evidence of an active RAP lesion found on FA, ICGA, or OCT. All patients were followed up on a monthly basis for 12 months after the initial treatment; the follow-up consisted of examinations such as fundus photography and SD OCT. Additional FA, ICGA, autofluorescence, and SD OCT examinations were performed whenever RAP recurred, or when RPE tears were suspected. The treatment was continued after occurrence of an RPE tear according to the retreatment criteria except when the patient strongly refused treatment.
SPSS software (version 13.0; SPSS, Inc, Chicago, Illinois, USA) was used for all statistical analyses. Frequencies were compared between treatment groups using either the χ 2 or Fisher exact test. Comparative statistical analyses were performed using unpaired t tests. A multivariate logistic regression analysis was performed using various baseline characteristics to identify risk factors for RPE tearing after treatment. Logistic regression models were tested using the Hosmer–Lemeshow test to ascertain the goodness of fit and forward and backward stepwise regression were performed using the likelihood-ratio model. The change in the likelihood ratio was used for variable selection, which was based on the maximum partial likelihood estimates for the covariate. All tests were 2-sided, and a P value <.05 was considered statistically significant.
A total of 109 eyes had been diagnosed with RAP and treated with intravitreal ranibizumab injections at our institution during the study period. Among these, 11 were excluded owing to follow-up loss. Finally, 98 eyes with RAP in 86 patients (45 men, 41 women; all South Korean) met the inclusion criteria and were enrolled in our study. The mean age of the subjects at their initial visit was 74.6 ± 7.9 years, and the mean number of injections administered was 4.96 ± 1.81 over 12 months. Table 1 outlines the clinical details of the patients included in the study.
|Total Eyes (n = 98)||RPE Tear During Treatment (n = 8)||No RPE Tear During Treatment (n = 90)||P|
|Age (y ± SD)||74.6 ± 7.9 (52–85)||77.1 ± 4.2||74.3 ± 6.6||.221 a|
|Male, n (%)||52 (53.1%)||3 (37.5%)||49 (54.4%)|
|Female, n (%)||46 (46.9%)||5 (62.5%)||41 (45.6%)||.113 b|
|Mean of baseline BCVA (logMAR; Snellen equivalent)||0.75 ± 0.52 (20/112)||0.76 ± 0.55 (20/115)||0.74 ± 0.49 (20/109)||.626 a|
|Baseline BCVA (logMAR; Snellen equivalent)|
|<0.54 (20/70)||20 (20.4%)||1 (12.5%)||19 (21.1%)|
|0.54 (20/70) to 1.0 (20/200)||47 (48.0%)||5 (62.5%)||42 (46.7%)|
|>1.0 (20/200)||31 (31.6%)||2 (25.0%)||29 (32.2%)||.352 c|
|Mean central foveal thickness ± SD (μm)||327 ± 166||344 ± 139||325 ± 199||.717 a|
|Mean subfoveal choroidal thickness ± SD (μm)||141 ± 55 (67–285)||101 ± 33||147 ± 61||.021 a|
|Lesion location, n (%)|
|Subfoveal||13 (13.3%)||1 (12.5%)||12 (13.3%)|
|Juxtafoveal||63 (64.2%)||4 (50.0%)||59 (65.6%)|
|Extrafoveal||22 (22.5%)||3 (37.5%)||19 (21.1%)||.511 c|
|Mean lesion size ± SD (mm 2 )||2.4 ± 1.1||2.9 ± 1.3||2.2 ± 1.2||.031 a|
|Stage of RAP, n (%)|
|Stage 1||9 (9.2%)||0 (0%)||9 (10.0%)|
|Stage 2||62 (63.3%)||4 (50.0%)||58 (64.4%)|
|Stage 3||27 (27.5%)||4 (50.0%)||23 (25.6%)||.323 c|
|Presence of PED, n (%)||59 (60.2%)||8 (100%)||51 (56.7%)||.016 b|
|Mean PED height ± SD (μm)||221 ± 144||408 ± 142||197 ± 112||.003 a|
|Reticular pseudodrusen, n (%)||60 (61.2%)||7 (87.5%)||53 (55.6%)||.083 b|
|Mean number of injections ± SD||4.96 ± 1.81||4.52 ± 1.43||5.01 ± 1.96||.550 a|
RPE tears had developed in 8 eyes (8.2%) after treatment during the 12-month follow-up period. All 8 of these eyes had shown PED at baseline. The majority of them (7 eyes) had developed the RPE tear during the period of the 3 monthly loading injections (within 3 months from baseline). RPE tears had developed in 2 eyes after the first injection, in 4 eyes after 2 injections, in 1 eye after 3 injections, and in 1 eye after 5 injections ( Table 2 ). The Supplemental Figure (available at AJO.com ) shows images taken from 1 representative case that had developed RPE tearing after intravitreal ranibizumab injections ( Table 2 , Patient 7).
|Patient||Age/Sex||RAP Stage||Lesion Size (mm 2 )||Lesion Location||Baseline BCVA (Snellen)||BCVA at 12 Months||Number of Injections Before RPE Tear||Baseline Central Foveal Thickness (μm)||Baseline Subfoveal Choroidal Thickness (μm)||Baseline PED Height (μm)||Foveal Involvement of RPE Tear||Reticular Pseudodrusen|