To evaluate the long-term visual outcome after combination therapy of photodynamic therapy (PDT) with intravitreal bevacizumab injections for polypoidal choroidal vasculopathy (PCV).
Retrospective observational study.
The medical records of 34 eyes (34 patients) with naïve PCV who were treated with combination therapy were analyzed. All patients completed at least 3 years of follow-up. All clinical data, including age, best-corrected visual acuity (BCVA, logarithm of the minimal angle of resolution [logMAR]), imaging data of fluorescein angiography, indocyanine green angiography, and optical coherence tomography, were investigated.
During a mean follow-up period of 46.8 ± 5.2 months, a mean of 1.4 ± 0.71 times of PDT and 9.2 ± 6.6 intravitreal bevacizumab injections were performed. During follow-up, 21 eyes (61.8%) showed at least 1 recurrence. Mean BCVA was 0.59 ± 0.35 logMAR (20/77 Snellen equivalent) at baseline and 0.39 ± 0.34 logMAR (20/49 Snellen equivalent) at 3 years ( P = .004). At 3 years, 14 patients (41.2%) gained 0.3 logMAR or more BCVA and 4 patients (11.8%) lost 0.3 logMAR or more BCVA than baseline. Baseline polyp size (β = .551; P = .005) and location of polyps (β = −.400; P =.033) were significantly correlated with long-term visual outcome after combination therapy for PCV.
Combination therapy of PDT with intravitreal bevacizumab injections showed favorable visual outcomes, and significant visual improvement was maintained in PCV patients. A total of 88.2% of patients avoided visual loss at 3 years after treatments. Largest polyp size at baseline and location of polypoidal lesions were prognostic factors for long-term visual outcomes in these patients.
Polypoidal choroidal vasculopathy (PCV) is a unique subset of type 1 choroidal neovascularization (CNV) with abnormal branching vascular network and terminal polypoidal structures. PCV is more prevalent among Asians, and clinical characteristics, such as location of polypoidal lesions and sexual predilection, seem to be different in Asian vs European patients. Since being introduced as a clinical entity in 1990, PCV is now an important diagnosis because of its relative resistance to anti–vascular endothelial growth factor (VEGF) therapy among patients with presumed exudative age-related macular degeneration (AMD).
Photodynamic therapy (PDT), alone or combined with anti-VEGF therapy, is currently recommended for treating patients with PCV. Several studies have demonstrated favorable visual outcomes and 71%-99% regression of polypoidal lesions after PDT. Long-term follow-up results of PDT also showed favorable visual outcomes during more than 5 years, avoiding significant visual loss in 88.1% of the patients. Combination of PDT with anti-VEGF therapy has been reported to be a good therapeutic option for patients with PCV. With combination therapy, anti-VEGF therapy seems to help rapid resolution of exudative fluid, in addition to the polyp regression effect of PDT, resulting in a more favorable visual outcome. However, long-term validation of the therapeutic efficacy of combination therapy for PCV has not been determined.
Although several studies investigated visual outcomes after PDT, anti-VEGF therapy, or combination therapy, relatively few studies investigated visual prognostic factors after treatments for PCV. Several factors, such as baseline best-corrected visual acuity (BCVA), lesion size and location, age, baseline hemorrhage, and presence of serous macular detachment, are suggested as prognostic factors for visual outcomes after PDT or combination therapy for PCV. However, the results are rather inconsistent because these studies did not consider the same factors during analysis of prognostic factors.
There has been no report regarding long-term visual outcomes and visual prognostic factors after combination therapy of PDT with anti-VEGF therapy for PCV. In this study, we evaluated long-term visual outcomes after initial combination therapy of PDT with intravitreal bevacizumab injections for PCV. In addition, we investigated long-term prognostic factors for visual outcomes in these patients.
Enrollment of Study Subjects
We retrospectively reviewed 34 eyes of 34 Korean patients with treatment-naïve PCV who fulfilled the inclusion and exclusion criteria. The patients were treated at the Vitreoretinal Service Clinic of Yonsei University Medical Center between January 1, 2006 and December 31, 2012. All patients underwent combination therapy of PDT with intravitreal bevacizumab injections as an initial treatment for PCV and completed at least 3 years of follow-up after the first treatment. Written informed consent was obtained from all participants at the time of treatment. This retrospective study was approved by the Institutional Review Board of Yonsei University College of Medicine and was conducted in accordance with the tenets of the Declaration of Helsinki. Inclusion criteria were: (1) symptomatic macular PCV with subfoveal leakage on fluorescein angiography (FA); and (2) presence of branching vascular networks and polypoidal lesions on indocyanine green angiography (ICGA). Exclusion criteria were: (1) any treatment, including laser photocoagulation, transpupillary thermotherapy, and radiotherapy; (2) other concomitant ocular diseases, such as diabetic retinopathy, high myopia, vein or artery occlusion, and epiretinal membrane; and (3) patients with aphakia, absence of posterior capsule, or history of vitrectomy owing to reduced anti-VEGF retention.
All patients received a complete ocular examination, including BCVA testing using a decimal visual acuity chart, slit-lamp biomicroscopy, dilated fundus examination with indirect ophthalmoscopy, color fundus photography, digital FA, ICGA, and optical coherence tomography (OCT). OCT images were obtained by time-domain OCT (Stratus; Carl Zeiss Meditec, Dublin, California, USA) before 2009, and then by spectral-domain OCT (Spectralis; Heidelberg Engineering, Heidelberg, Germany). FA and ICGA were obtained using the Heidelberg Retina Angiograph system (HRA-2; Heidelberg Engineering) with a confocal scanning laser ophthalmoscope.
CNV type was determined by FA. Greatest linear dimension and area of the lesion were determined by ICGA. Polypoidal lesions were classified according to angiographic characteristics. The locations of PCV lesions were subdivided into 3 categories: subfoveal, juxtafoveal (within 1-199 μm of the foveal center), and extrafoveal (200 μm or more from the foveal center). In addition, formation of the PCV lesions was classified into 3 categories: single, cluster, and string. PCV were classified into 2 categories by the number of discrete areas of polypoidal lesions: single and multiple. Greatest linear dimension included the entire branching vascular network of vessels and polypoidal lesions at the early phase of ICGA, assessed by an electronic caliper using HRA-2 software. The PCV lesion area was also measured manually the same way. When measuring the greatest linear dimension and the PCV lesion area, pigment epithelial detachment (PED) was not counted if there was no underlying vascular component.
Baseline OCT characteristics were also investigated, including intraretinal fluid cyst, subretinal fluid (SRF), and PED at the fovea. Central foveal thickness (CFT) was defined as the distance between the internal limiting membrane and the inner surface of the retinal pigment epithelium (RPE), and was measured manually at the fovea. The SRF and intraretinal fluid cyst were included in the CFT measurements, whereas PED was not.
Patients were treated with combination therapy of 1 session of PDT and 3 monthly intravitreal injections of 0.5 mL bevacizumab (1.25 mg Avastin; Genentech, South San Francisco, California, USA). For PDT, all patients received a 6 mg/m 2 infusion of verteporfin (Visudyne; Novatis AG, Bulach, Switzerland) over a period of 10 minutes, followed by delivery of diode laser at 689 nm to the CNV 15 minutes after the commencement of infusion. A total light energy of 50 J/cm 2 and light dose rate of 600 mW/cm 2 for 83 seconds were used to cover the entire lesion. The PDT lesion included an additional 500 μm covering the borders on each side. Right after PDT, intravitreal bevacizuamab injections were performed on the same day. For intravitreal injection, topical anesthesia was applied and 10% povidone-iodine was used to scrub eyelids and lashes. Povidone-iodine eye drops (1.25%) were applied, and a sterile lid speculum was put between the eyelids. Bevacizumab was injected into the vitreous cavity through the superior sclera using a 30 gauge needle, at a position 3.5 mm posterior to the corneal limbus in phakic eyes and 3.0 mm posterior in pseudophakic eyes. Pressure was applied to the injection site using a sterile cotton swab for 1 minute to prevent leakage. After injections, patients were instructed to apply antibiotic eye drops for 1 week.
Follow-up and Reinjection Protocols
Follow-up visits were arranged 1 week after each baseline treatment and then 1, 3, 6, 12, 24, and 36 months afterward, with examination including BCVA, dilated fundus examination with indirect ophthalmoscopy, and OCT. Three months after the first combination therapy, FA and ICGA were additionally performed to evaluate treatment outcomes. Patients were asked to return to the clinic immediately in cases of visual loss and/or metamorphopsia recurrence. Additional FA, ICGA, OCT, and dilated fundus examinations were performed once a year after treatment and whenever physicians suspected PCV recurrence or visual loss and/or metamorphopsia recurrence. We defined PCV recurrence as the reappearance of active PCV lesions on ICGA with subfoveal leakage on FA after at least 6 months without treatment. Earlier detection of active PCV with leakage before 6 months was considered to be persistence of the original lesion, rather than recurrence.
Retreatment was considered if any of the following changes were observed during follow-up: (1) any loss of visual acuity lines, determined with a decimal visual acuity chart, with OCT evidence of fluid in the macula; (2) an increase in CFT by OCT of at least 100 μm; (3) new macular hemorrhage; or (4) evidence of persistent fluid on OCT.
Additional treatment was given in case of recurrent or residual PCV during the follow-up examination. If polyps persisted or recurred on ICGA, additional PDT with pro re nata intravitreal bevacizumab injections was done. If only abnormal branching vascular network persisted, monthly pro re nata intravitreal bevacizumab injections were done.
Patient characteristics were retrieved from their medical charts, including age at initial diagnosis, sex, and BCVA as determined using decimal visual acuity charts. BCVA results were converted to a logarithm of the minimal angle of resolution (logMAR) value for statistical analysis. After BCVA was converted to logMAR, an improvement of ≥0.3 in logMAR visual acuity was defined as visual gain and a decrease of ≥0.3 logMAR visual acuity was defined as visual loss. All examination data at baseline and at 1-, 3-, 6-, 12-, 24-, and 36-month follow-ups were interpreted retrospectively.
As for the possible prognostic factors for long-term visual outcome in these patients, several factors were considered. These included age at diagnosis, sex, BCVA at baseline, size of the largest polyp, greatest linear dimension of choroidal vascular networks on ICGA, presence of subretinal or intraretinal fluid, presence of PED (larger than the area of 1 optic disc diameter), type of PED (hemorrhagic or serous), baseline CFT, polyp location (subfoveal, juxtafoveal, or extrafoveal), type of polyps (single, cluster, or string), and area of discrete polyp lesions (single or multiple).
IBM SPSS 18.0 software for Windows (SPSS/IBM Corporation, Chicago, Illinois, USA) was used for statistical analyses. To compare the mean BCVA at each time point and baseline BCVA, a paired t test was used. The 2-sample t test was used for analysis of continuous variables, and the χ 2 test was used for categorical variables. For subgroup analysis, nonparametric analyses were performed, using the Mann-Whitney U test for continuous variables and Wilcoxon signed-rank test for categorical variables. Stepwise logistic regression was used for determining predictive factors for long-term visual prognosis after anti-VEGF therapy for PCV. Mauchly’s test of sphericity and Kolmogorov-Smirnov analysis were used to confirm statistical validity. Differences with P < .05 were considered statistically significant.
Overall Characteristics of the Patients
In total, 34 eyes of 34 patients who had completed at least 3 years of follow-up visits after the initial combination therapy were analyzed. Patients had a mean age of 64.2 ± 10.9 years at diagnosis and showed a male predominance (22 male patients, 64.7%). Detailed characteristics of the patients are listed in Table 1 . Baseline FA showed occult CNV in all eyes. During a mean follow-up of 46.8 ± 5.2 months, mean total numbers of PDT were 1.4 ± 0.7 (median 1.0), and those of intravitreal bevacizumab injections were 9.2 ± 6.6 (median 5.5). During the first year of treatment, mean 1.3 ± 0.5 times (median 2.0) of PDT and mean 4.3 ± 2.8 injections (median 3.5) of intravitreal bevacizumab were performed. During the second year, mean 0.3 ± 0.2 times of PDT (median 0) and mean 4.1 ± 3.8 intravitreal bevacizumab injections (median 3.0) were done. During the third year, mean 0.2 ± 0.1 times of PDT (median 0) and 2.9 ± 3.5 intravitreal bevacizumab injections (median 2.5) were done.
|Age of onset (y), mean ± SD||64.2 ± 10.9|
|Male/female, n (%)||22 (64.7%)/12 (35.3%)|
|Presence of drusen, n (%)||4 (11.8%)|
|Greatest linear dimension (mm), mean ± SD||3.1 ± 1.5|
|Largest polyp size (mm), mean ± SD||0.5 ± 0.3|
|Central foveal thickness (μm), mean ± SD||362.0 ± 99.1|
|Baseline OCT characteristics|
|Subretinal fluid||33 (97.1%)|
|Intraretinal cyst||9 (26.5%)|
|Discrete polyp area|
Visual Outcomes After Combination Therapy for Polypoidal Choroidal Vasculopathy
The mean BCVA was 0.59 ± 0.35 logMAR (20/77 Snellen equivalent) at baseline and 0.39 ± 0.34 logMAR (20/49 Snellen equivalent) at 3 years. Changes in mean BCVA from baseline during the follow-up periods are shown in Figure 1 . Compared with baseline BCVA, the mean BCVA improved 1 month after initial treatment ( P = .083), and significant visual improvement was achieved at 3 months ( P = .002). Significant visual improvement was maintained throughout 3 years ( P = .007 at 6 months, P < .001 at 1 year, and P = .001 at 2 years).
As previously defined, the patients were analyzed and classified as having visual gain or visual loss. At 1 year, 15 of 34 patients (44.1%) showed visual gain, whereas 1 patient (2.9%) showed visual loss. Fourteen patients (41.2%) had visual gain at 2 and 3 years, whereas 3 patients (8.8%) at 2 years and 4 patients (11.8%) at 3 years showed visual loss ( Figure 2 ).
Changes in Lesions After Combination Therapy for Polypoidal Choroidal Vasculopathy
The mean CFT significantly decreased from 362.0 ± 99.1 μm at baseline to 269.4 ± 46.0 μm at 3 years ( P < .001). During the follow-up period, mean CFT showed a decreasing tendency. The changes of mean CFT at each time point are shown in Figure 3 .
We evaluated polyp regression rates determined by ICGA at 3 months after initial treatment. Thirty-two of 34 eyes (94.1%) showed polyp regression at 3 months. By additional ICGA, 2 eyes showed persistent polyps and 3 eyes showed recurrence of polyps at 1 year. At 2 years, 1 eye still showed persistent polyp and 4 eyes showed recurrence of polyps. At 3 years, 3 eyes showed recurrence of polyps and no eyes showed persistent polyps. Polyp regression was 88.2% at 1 year, 85.3% at 2 years, and 91.2% at 3 years.
Recurrence of Polypoidal Choroidal Vasculopathy
Recurrences of PCV were noted in 21 out of 34 eyes (61.8%). The mean recurrence interval was 12.8 ± 5.7 months, and the mean number of recurrences was 1.5 ± 0.6 times during the follow-up. Among 21 eyes, 10 eyes (47.6%) showed reappearance of new polyps; the others showed only persistent exudation from abnormal branching vascular networks. The mean number of intravitreal bevacizumab injections was 12.4 ± 6.5 in the eyes with recurrent PCV vs 3.9 ± 1.2 injections in eyes without recurrence ( P < .001). The mean number of PDT was 1.7 ± 0.8 in the eyes with PCV recurrence and 1.1 ± 0.3 in the eyes without PCV recurrence ( P = .027).
When comparing the eyes with and without recurrence, polyp size ( P = .024) at baseline was significantly smaller in the eyes without recurrence. The comparison between the eyes with and without recurrence is shown in Table 2 .
|Characteristic||Recurrence (N = 21)||No Recurrence (N = 13)||P Value|
|Age of onset (y), mean ± SD||66.1 ± 9.0||61.0 ± 5.1||.427 a|
|Greatest linear dimension (mm), mean ± SD||3.2 ± 1.4||2.8 ± 1.5||.529 a|
|Largest polyp size (mm), mean ± SD||0.6 ± 0.4||0.3 ± 0.2||.024 a|
|Presence of drusen, n (%)||3 (14.3%)||1 (7.7%)||.502 b|
|Central foveal thickness (μm), mean ± SD||351.0 ± 89.5||380.0 ± 114.5||.292 a|
|Lesion location||.362 b|
|Subfoveal||13 (61.9%)||7 (53.8%)|
|Juxtafoveal||6 (28.6%)||6 (46.2%)|
|Extrafoveal||2 (9.5%)||0 (0%)|
|Baseline OCT characteristics|
|Subretinal fluid||20 (95.2)||13 (100.0%)||.618 b|
|Intraretinal cyst||5 (23.8)||4 (30.8)||.475 b|
|PED||12 (57.1%)||6 (46.2%)||.393 b|
|Polyp type||.876 b|
|Single||8 (38.1%)||4 (30.8%)|
|Cluster||12 (57.1%)||8 (61.5%)|
|String||1 (4.8%)||1 (7.7%)|
|Discrete polyp area||.498 b|
|Single||19 (90.5%)||11 (84.6%)|
|Multiple||2 (9.5%)||2 (15.4%)|