Purpose
To compare the long-term visual and anatomic outcome of treatment with photodynamic therapy (PDT) or intravitreal bevacizumab (IVB; Avastin; Genentech Inc, South San Francisco, California, USA) for choroidal neovascularization attributable to pathologic myopia (mCNV).
Design
An open-label, interventional case series.
Methods
setting: Multi-institutional. patients: Thirty-one eyes of Japanese women who received either PDT or IVB for mCNV. Inclusion criteria were age 50 years or older, greatest linear dimension (GLD) 1200 to 3000 μm, and baseline best-corrected visual acuity (BCVA) 20/200 to 20/40. intervention procedures: Patients received either PDT or IVB (1 mg/40 μL) throughout the study, with re-treatment when necessary. main outcome measures: BCVA and visual gain/loss at 3, 6, 12, 18, and 24 months after the initial treatment.
Results
Age, BCVA, location of CNV, refractive error, and symptom duration at baseline did not differ significantly between groups. BCVA was significantly improved at 3 to 12 months ( P < .05); however, the significance was lost at 18 and 24 months in the IVB group. The PDT group showed no significant improvement within the first year, and vision slowly worsened after 12 months, becoming significantly worse at 18 and 24 months compared to baseline ( P < .01). BCVA was significantly higher in the IVB group at 6 months ( P < .05), and 12 months or further ( P < .01). Visual gain was significantly greater in the IVB group at 6, 12, 18, and 24 months ( P < .05 for 6, 18, and 24 months and P < .01 for 12 months).
Conclusions
These findings indicate that the effects of PDT and IVB have a different time course, and that IVB provides a significantly better BCVA than PDT for mCNV over the long-term.
High myopia, generally defined as a refractive error (RE) of −6 diopters (D) or greater, is very common in Europe, in Asia, and in some ethnic groups in the United States. Myopia is a major cause of visual impairment in many countries. Choroidal neovascularization (CNV) is a major myopia-related vision-threatening disease attributable to progressive and irreversible central chorioretinal atrophy. A recent prospective study revealed that more than 90% of patients with myopia-related CNV (mCNV) have vision worse than 20/200 5 to 10 years after disease onset, indicating the importance of developing effective treatment options for mCNV.
Photocoagulation with thermal burn was first introduced as a treatment for mCNV, especially for extrafoveal cases. The coagulated scar usually expands over time, and the resulting progressive central scotoma is a major concern. Thereafter, photodynamic therapy (PDT) using verteporfin was introduced. A prospective, randomized clinical trial, the VIP Study, reported that the visual outcome following PDT treatment was significantly better than that of placebo group at both 3 months and 1 year. The significant benefit was lost after 2 years, suggesting that PDT has only a short-term benefit.
Vascular endothelial growth factor (VEGF) has a critical role in CNV development, because VEGF transgenic mice expressing photoreceptors that are driven by a rhodopsin promoter develop CNV. This finding led us to hypothesize that anti-VEGF treatment would be useful to shrink CNV and subsequently improve vision. Bevacizumab (Avastin; Genentech Inc, South San Francisco, California, USA) is a monoclonal mouse antibody for human VEGF that is currently used for various retinal diseases, such as diabetic retinopathy, idiopathic CNV, age-related macular degeneration, and retinal vaso-occlusive diseases. Intravitreal bevacizumab (IVB), first introduced in 2007, for mCNV, has become a major treatment option for the disease. There are several reports of the 1-year results of treatment with IVB. All of these reports, however, are case series studies that lack controls, making it difficult to evaluate the potential therapeutic effect of IVB as a treatment for mCNV.
Here, we report the long-term (2-year) results of a retrospective interventional multicenter study evaluating the treatment outcomes between patients who received IVB or PDT.
Patients and Methods
Patient Selection
Thirty-one eyes of 31 consecutive female patients who received PDT or IVB for newly developed mCNV in the Department of Ophthalmology of Osaka University Hospital (Suita, Japan), Kansai Medical University (Hirakata Japan), Osaka Dental University (Osaka, Japan), or Osaka General Medical Center (Osaka, Japan) were included in the study. Patient records were retrospectively reviewed for information such as age, gender, symptom duration, refractive error, axial length, treatment choice, history, pre- and posttreatment best-corrected visual acuity (BCVA), and results of color fundus photography, fluorescein angiography (FA), and all other necessary information. Patient eligibility criteria were as follows: 1) active subfoveal or juxtafoveal CNV associated with pathologic myopia, defined as a RE of −6 D or greater and an axial length of 26.5 mm or longer, and treated with either PDT or IVB; 2) no history of previous treatment for mCNV, such as photocoagulation, PDT, or IVB; 3) no triamcinolone acetonide sub-Tenon treatment within the previous 6 months; 4) baseline BCVA between 20/200 and 20/40; 5) baseline greatest linear dimension (GLD) between 1200 and 3000 μm; and 6) women at least 50 years of age. Exclusion criteria included: 1) history of vitrectomy or intraocular surgery other than cataract surgery; 2) presence of macular hole, retinal detachment, foveoschisis; 3) severe cataract obscuring the fundus; 4) symptom duration of more than 24 months; and 5) significant glaucoma detected by visual field loss.
Treatment Choice
Treatment was selected following discussion with the patients. All relevant explanations regarding the benefits, risks, and complications were described, and agreement of patients and doctors, as well as informed consent, was obtained at this time. Retinal specialists performed all the diagnostic and interventional procedures. Because IVB is off-label, the approval of the Institutional Review Board was obtained from each of the hospitals. In Japan, PDT became available in May 2004 and IVB became available in October 2005. The initial treatment period was between May 2004 and October 2006 for the PDT group, and between October 2005 and October 2006 for the IVB group in the present study. Therefore, only PDT was available from May 2004 to September 2005, and both PDT and IVB were available after October 2005.
Treatment Procedures
Standard PDT was applied according to the VIP Study with a verteporfin (Visudyne; Novartis AG, Basel, Switzerland) infusion. Spot size was set at +1000 μm greater than the GLD. IVB was performed as previously described. For the IVB injection, 1 mg of bevacizumab (Avastin) was injected via the pars plana using aseptic techniques. Antibiotic eye drops were applied for 1 week after the bevacizumab injection.
Treatment Schedule, Follow-Up, and Retreatment
Photodynamic Therapy
After the initial PDT, patients were examined every month for the first 3 months, and thereafter FA was performed to confirm the CNV regression. When the CNV persisted, the PDT was repeated, but the interval between each treatment was at least 3 months. Once the CNV was no longer active, patients were examined every 1 to 3 months. BCVA was measured and OCT examination was performed at every visit. FA was performed if recurrent mCNV was suspected based on the appearance of subretinal or intraretinal fluid, new hemorrhage, or significant visual loss (2 lines or more). If persistent or recurrent dye leakage was observed in FA, patients were re-treated.
Intravitreal Bevacizumab
The IVB regimen was described previously and was similar to the PDT regimen. After the initial IVB, the patients were examined every month for the first 3 months. If persistent subretinal fluid was observed in OCT images, IVB was repeated until it resolved. Thereafter, FA was performed to confirm the CNV regression 3 months after the initial IVB. When the CNV persisted, the IVB treatment was repeated. Once the CNV was no longer active, patients were examined every 1 to 3 months. BCVA was measured and OCT examination was performed at every visit. FA was performed if recurrent mCNV was suspected based on the appearance of subretinal or intraretinal fluid, new hemorrhage, or significant visual loss (2 lines or more). If persistent or recurrent dye leakage was observed, patients were retreated with IVB. None of the patients switched treatment modalities during the study.
Examinations
BCVA and intraocular pressure was measured, and optical coherence tomography (OCT; Stratus OCT or Cirrus HD-OCT; Carl Zeiss Meditec, La Jolla, California, USA) was performed at every visit by a technician masked to the patient’s treatment. The presence/absence of subretinal or intraretinal fluid was evaluated by horizontal B scan. GLD was measured with Imagenet System software (Topcon, Tokyo, Japan) from the digitized baseline FA images. Changes in the FA dye pattern of leakage from the CNV were graded as 1 of 3 types: vanished, diminished, and unchanged/increased. FA was graded by 3 of the authors (Y.I., Y.N., and S.M.). Myopic chorioretinal atrophy was graded as previously described by Avila and associates and evaluated by 3 of the authors (Y.I., Y.N., and S.M.), similar to FA.
Endpoints
The efficacy of IVB and PDT was evaluated based on BCVA at 1, 3, 6, 12, 18, and 24 months after the initial treatment, as well as BCVA gain at each time point as the primary outcome. The treatment number, changes in FA findings after treatment, and rate of serious adverse events were also compared as secondary outcome measures.
Statistical Analysis
BCVA was converted to the logarithm of minimal angle of resolution (logMAR) scale. Statistical evaluation was performed using JMP software version 7.0 (SAS Inc, Cary, North Carolina, USA). A stepwise regression analysis was performed to investigate the factors predictive of BCVA as well as visual improvement after IVB at 24 months. Visual outcome is affected by baseline visual acuity (VA). Therefore, the difference in visual outcome between the 2 groups was also compared using an analysis of covariance (ANCOVA) after adjusting for the baseline BCVA. A P value of less than .05 was considered significant.
Results
Patient Characteristics
Thirty-one eyes (18 right eyes, 13 left eyes) of 31 female patients were eligible for the study. Twenty eyes were treated with PDT (PDT group), and 11 were treated with IVB (IVB group). Mean ± standard deviation (SD) age was 67.0 ± 7.0 years, ranging from 53 to 79 years. Mean symptom duration was 5.7 ± 4.8 months, ranging from 1 to 20 months. Twenty were phakic, 11 were pseudophakic, and there were no aphakic eyes. The mean ± SD spherical equivalent RE was −9.8 ± 5.0 D, ranging from −1.5 to −17.5 D, including pseudophakic eyes. Mean ± SD GLD was 2054 ± 456 μm, ranging from 1327 to 2984 μm. The degree of myopic atrophy according to Avila’s classification was as follows: M1 in 4 eyes, M2 in 13 eyes, M3 in 4 eyes, M4 in 8 eyes, and M5 in 2 eyes.
Baseline Comparison
Baseline characteristics of patients in both groups are shown in Table 1 . There were no significant differences in age, symptom duration, RE, lens status, CNV size, myopic chorioretinal atrophy grading, CNV location, or pretreatment BCVA between the 2 groups.
| Factors | IVB Group (n = 11) | PDT group (n = 20) | P value |
|---|---|---|---|
| Age (years), mean ± SD | 67.8 ± 6.2 | 66.6 ± 7.5 | .63 a |
| Symptom duration (months), mean ± SD | 5.5 ± 4.3 | 5.9 ± 5.2 | .81 a |
| Refractive error (D), mean ± SD | −7.8 ± 4.5 | −10.9 ± 5.0 | .10 a |
| Lens status (n) | |||
| Phakic | 6 | 14 | |
| Pseudophakic | 5 | 6 | .32 c |
| CNV size (GLD, μm), mean ± SD | 1875 ± 446 | 2153 ± 442 | .10 a |
| Myopic atrophy score (0 to 5), mean ± SD | 2.9 ± 1.1 | 2.3 ± 1.1 | .10 a |
| CNV location (sub/juxta/extra) | |||
| Subfoveal | 8 | 18 | |
| Juxtafoveal | 3 | 2 | .32 b |
| Pretreatment logMAR value, mean ± SD | 0.68 ± 0.29 | 0.74 ± 0.20 | .32 a |
Visual Outcome
Baseline Comparison
The mean logMAR values of both groups at each time point are shown in Table 2 . BCVA was significantly better ( P < .05) in the IVB group than in the PDT group after 6 to 24 months. In the IVB group, BCVA was significantly better than baseline at 1, 3, 6, and 12 months after treatment ( P < .05 by paired t test), but not at 18 months ( P = .29) or 24 months ( P = .38). In the PDT group, there was no visual improvement at 1 and 3 months ( P = .65 and P = .61, respectively). Thereafter, the BCVA began to slowly decrease. The mean logMAR value was not different from baseline at 6 and 12 months ( P = .41 and P = .09, respectively). BCVA continued to decrease throughout the second year, and at 18 and 24 months was significantly worse than at baseline ( P < .01 for both).
| Time Point | Mean logMAR Value | P value | |
|---|---|---|---|
| IVB Group (n = 11) | PDT Group (N = 20) | ||
| Baseline | 0.68 ± 0.29 | 0.74 ± 0.20 | .49 |
| 1 month | 0.56 ± 0.31 | 0.76 ± 0.25 | .054 |
| 3 months | 0.54 ± 0.34 | 0.70 ± 0.34 | .19 |
| 6 months | 0.50 ± 0.34 | 0.79 ± 0.31 | .02 a |
| 12 months | 0.49 ± 0.29 | 0.90 ± 0.36 | <.01 a |
| 18 months | 0.56 ± 0.35 | 0.89 ± 0.24 | <.01 a |
| 24 months | 0.56 ± 0.34 | 0.92 ± 0.24 | <.01 a |
Comparison Between Groups
The logMAR change from baseline in both groups is shown in Figure 1 . Mean logMAR improved by 0.13 at 1 month after IVB treatment, and this improvement was maintained until the final follow-up at 24 months. BCVA tended to improve for 3 months after PDT treatment, but then began to decrease after 3 months, and this trend continued until the end of the follow-up period. The mean logMAR gain/loss was similar between groups at 1 ( P = .08) and 3 ( P = .29) months after treatment; the gain was significantly greater in the IVB group at 6 ( P < .05), 12 ( P < .01), 18 ( P < .05), and 24 months ( P < .05).
