Article Text
PDF
Abstract
Aims To assess the long-term efficacy of intravitreal bevacizumab for recurrent leakage owing to the residual branching vascular networks in polypoidal choroidal vasculopathy after photodynamic therapy.
Methods Forty-five eyes with exudative branching vascular networks were treated with intravitreal bevacizumab and followed for at least 24 months. Original polypoidal lesions had been treated successfully with previous photodynamic therapy in all eyes. The best-corrected visual acuity and retinal morphological changes were assessed retrospectively.
Results Exudative branching vascular networks were characterised as occult choroidal neovascularisation (38 eyes) or classic choroidal neovascularisation (7 eyes) on fluorescein angiography. Intravitreal bevacizumab maintained or improved vision in 38 eyes (84%) over 12 months and in 36 eyes (80%) over 24 months, although the mean visual acuity at 12 and 24 months did not differ significantly compared with baseline. Complete resolution of macular fluid was achieved continuously in 26 eyes (58%) during 24 months. Sixteen eyes (36%) responded once to treatment but became unresponsive to additional injections for recurrent exudation. Three eyes (7%) were refractory to treatment throughout follow-up. Cystoid macular oedema eventually developed in 10 eyes and was a poor prognostic sign for visual outcome.
Conclusion Intravitreal bevacizumab improved the retinal morphology and maintained vision over 1 year in most eyes with recurrent fluid owing to persistent abnormal vascular networks in polypoidal choroidal vasculopathy. The therapeutic response, however, may decrease during the second year.
- Polypoidal choroidal vasculopathy
- exudative branching vascular networks
- intravitreal bevacizumab
- choroid
- retina
- macula
- neovascularisation
- treatment medical
Statistics from Altmetric.com
- Polypoidal choroidal vasculopathy
- exudative branching vascular networks
- intravitreal bevacizumab
- choroid
- retina
- macula
- neovascularisation
- treatment medical
Polypoidal choroidal vasculopathy (PCV), characterised by a complex of branching vascular networks terminating in polypoidal lesions,1–5 accounts for 23–54% of neovascular age-related macular degeneration (AMD) in Asian populations.6 7 Photodynamic therapy (PDT) maintains or improves vision by resolving the polypoidal lesions and accompanying fluid beneath the neurosensory retina.8–10 However, the branching vascular networks usually remain even after PDT,11–13 and may enlarge further over time, resulting in persistent exudation appearing as choroidal neovascularisation (CNV) secondary to AMD in some cases.14 Therefore, stabilisation of the branching vascular networks may be crucial for long-term management of PCV.
Recent studies have reported the efficacy and safety of bevacizumab (Avastin, Genentech, South San Francisco, California) for stabilising neovascular activity and maintaining vision in patients with neovascular AMD.15 Those results promoted us to consider the drug as the treatment of choice for CNV-like branching vascular networks of PCV after PDT.
The purpose of this retrospective study is to assess the long-term efficacy of intravitreal bevacizumab in the management of exudative branching vascular networks in eyes with PCV.
Methods
This study was a retrospective, consecutive, interventional case series conducted at Osaka University Hospital. Patients who received intravitreal bevacizumab for the treatment of recurrent exudation associated with branching vascular networks between December 2006 and April 2008 and followed for at least 24 months were initially enrolled. All eyes had a previous history of successful PDT for PCV with resolution of polypoidal lesions. Recurrent exudation was determined by the presence of subretinal fluid (SRF) or macular oedema on optical coherence tomography (OCT) with CNV-like fluorescein leakage caused by branching vascular networks. Patients were excluded if they had SRF or macular oedema caused by recurrent polypoidal lesions, follow-up <24 months after intravitreal bevacizumab or clinically relevant media opacity.
All patients underwent a comprehensive ocular examination, including measurement of the best-corrected visual acuity (BCVA), intraocular pressure, binocular indirect ophthalmoscopy and contact lens slit-lamp biomicroscopy, colour fundus photography, OCT, fluorescein angiography (FA) and indocyanine green angiography (ICGA). The patients were examined after a detailed explanation of the study was provided, and they provided informed consent. This study was approved by the institutional review board committee of Osaka University Hospital.
ICGA analysis
ICGA was performed at baseline before intravitreal bevacizumab and 3–6 months and 12–24 months after intravitreal bevacizumab. A confocal scanning laser ophthalmoscope (Heidelberg Retina Angiograph 2 (HRA2), Heidelberg Engineering GmbH, Dossenheim, Germany) was used as previously described.14 The planimetric size of the branching vascular networks was measured from the early-phase ICGA at baseline to the final visits using software included in the HRA2.14 The individual readings of two investigators were averaged. A change in lesion size was recorded when the lesion increased or decreased more than 50% in the corresponding area.
OCT analysis
OCT images were obtained by Stratus OCT (Carl Zeiss Meditec, Dublin, California) or the Cirrus HD-OCT (Carl Zeiss Meditec). The central retinal thickness (CRT), defined as the distance between the internal limiting membrane and the inner surface of the retinal pigment epithelium (RPE), was measured manually at the fovea.16 The SRF and intraretinal fluid were included in the CRT measurements, whereas sub-RPE fluid was not included. The fluid in the macula was identified as intraretinal fluid (macular oedema) and SRF, and a fluid-free macula was defined by the absence of macular oedema and SRF as determined by OCT.
Follow-up and reinjection protocols
Intravitreal bevacizumab (1.25 mg) was injected during an outpatient procedure under strict aseptic conditions. All patients were followed monthly for more than 24 months. The VA and OCT were examined at every visit. Re-treatment with intravitreal bevacizumab was considered if there was OCT evidence of macular fluid with at least one-line loss of VA, new macular haemorrhage or newly developed fibrinous changes. A treatment response on OCT was defined as complete if there was no macular fluid, partial if there was no macular fluid initially but the lesions become refractory to treatment after recurrence of exudation, or no response if there was not an absence of macular fluid. When recurrence was suspected, FA and ICGA were performed at the discretion of the physician.
Data collection and statistical analysis
The main outcome measures were the changes in exudative fluid and the BCVA during 24 months after the initial injection. The changes in the CRT and the size of the branching vascular networks before and after injection were also evaluated. Statistical analyses were performed with SAS software, version 9.1 (SAS Institute). P-values <0.05 were considered significant.
Results
Forty-five eyes of 45 patients met the criteria for data analysis. The patient demographics are shown in table 1. Original polypoidal lesions had resolved in all eyes on ICGA with complete absence of fluid after previous PDT, but the branching vascular network remained in all eyes. The development of subsequent exudative changes associated with residual branching vascular networks without recurrent polypoidal lesions was seen 9.3±8.4 months (range 1–36) after the previous PDT (figure 1).
Patient baseline characteristics
Exudation associated with branching vascular networks in eyes with polypoidal choroidal vasculopathy. The exudation patterns are characterised as occult choroidal neovascularisation (CNV) without apparent pigment epithelial detachment (PED) (A–D), occult CNV with fibrovascular PED (E–H) or classic CNV (I–L). (A) Indocyanine green angiography (ICGA) before previous photodynamic therapy (PDT) showing branching vascular networks terminating in small polypoidal lesions in a 64-year-old man. The polypoidal lesions resolved completely after PDT. However, the branching vascular networks remained (B) and subsequently developed exudative changes resembling occult CNV on fluorescein angiography (FA) 6 months after the previous PDT (C). (D) Optical coherence tomography (OCT) showing subretinal fluid (SRF) with limited retinal pigment epithelium elevation. (E) ICGA before previous PDT showing branching vascular networks terminating in polypoidal lesions with a serous PED in a 77-year-old man. The polypoidal lesions have resolved completely after PDT. However, the branching vascular networks have enlarged (F) and subsequently developed exudative changes resembling occult CNV with a fibrovascular PED on FA 3 months after previous PDT (G). (H) OCT showing a large PED with SRF. (I) ICGA before previous PDT showing small branching vascular networks terminating in polypoidal lesions in a 57-year-old woman. The polypoidal lesions have resolved completely after PDT. Without recurrence of the polypoidal lesion, the branching vascular networks show exudative changes at the macula (J) resembling classic CNV on FA 6 months after previous PDT (K). (L) The OCT shows fibrin accumulation and macular oedema.
Angiographic and OCT characteristics of exudative branching vascular networks
FA showed leakage mimicking occult CNV in 38 eyes (84%) (figure 1C,G) and classic CNV in seven eyes (16%) (figure 1K). Of the 38 eyes with occult CNV-like lesions, ICGA showed thin branching choroidal vessels that depicted relatively well-delineated plaque in the late phase. Of the seven eyes with classic CNV-like lesions, subretinal fibrinous exudation was seen in all eyes. The exudation associated with the branching vascular networks was characterised by OCT as SRF in 43 eyes (96%) and macular oedema in 10 eyes (22%). Apparent PED was seen in 24 eyes (53%), and the limited RPE elevation was detected in 14 eyes (31%).
Visual acuity, ICG angiography and OCT outcomes
The mean number of injections in the 45 eyes during the first 12 months was 2.9±1.8 (range 1–8) and 1.9±1.8 (range 0–6) during the second year.
The mean baseline BCVA was 0.45±0.30 (table 1). The mean BCVA values were 0.43±0.33, 0.43±0.30, 0.48±0.32 and 0.51±0.38 at 3, 6, 12 and 24 months, respectively (p=0.118, p=0.428, p=0.523 and p=0.206, respectively) (figure 2A). The BCVA at the 12-month follow-up visit improved by three or more lines in four eyes (9%), was unchanged within three lines in 34 eyes (76%) and worsened in seven eyes (16%). At 24 months, BCVA improved by three or more lines in six eyes (13%), was unchanged in 30 eyes (67%) and worsened in nine eyes (20%).
(A) Mean change in the best-corrected visual acuity (BCVA) (mean±SD) after intravitreal bevacizumab for exudative branching vascular networks through 24 months. The visual acuity is expressed as the logarithm of the minimum angle of resolution (logMAR). There are no statistical differences between the baseline and post-treatment BCVA values at any time point. (B) Mean change in the central retinal thickness (CRT) (mean±SD) after intravitreal bevacizumab for exudative branching vascular networks through 24 months. The mean CRT has decreased significantly at 3, 6 and 12 months after initial intravitreal bevacizumab (*p<0.05 compared with baseline). However, there are no statistical differences in CRT between baseline and 24 months. M, months.
ICGA images showed persistent branching vascular networks in all 45 eyes during follow-up. The size of the branching vascular network increased in 20 eyes (44%), remained unchanged in 21 eyes (47%) and decreased in four eyes (9%) at the final ICGA examination compared with baseline. In six eyes, polypoidal lesions reappeared at the site connected to the branching vascular networks. The period between the initial intravitreal bevacizumab and detection of newly developed polypoidal lesions was 3 months in one eye, 6 months in one eye and 24 months in four eyes.
The mean baseline CRT measured on OCT was 222±71 μm (range 87–434). The CRT decreased an average of 31.0 μm from baseline by 3 months (p<0.001) and an average of 40.0 μm by 6 months (p<0.001) (figure 2B). Those initial decreases lessened over time after initial intravitreal bevacizumab, with an average reduction of 25.8 μm by 12 months (p=0.050) and 1.8 μm by 24 months (p=0.296). There tended to be a correlation between the CRT at 24 months and the BCVA at 24 months (r=0.292, p=0.051).
A fluid-free macula was achieved during the first 12 months in 42 eyes (93%). The mean number of injections required to achieve a fluid-free macula was 1.9±1.6 (range 1–9). Subsequently, 36 eyes developed recurrent macular fluid and received additional injections. Overall, 26 eyes (58%) were regarded on OCT as complete responders because intravitreal bevacizumab was effective throughout the 24 months of follow-up in completely resolving the exudation. One initial injection was effective to keep the macula dry for 24 months in two eyes. Sixteen eyes (36%) were considered to be partial responders. Those eyes once had no macular fluid after injection during the first 12 months; however, they became unresponsive to treatment despite repeated injections for recurrent exudation (figure 3). Three eyes (7%) were considered non-responders because the macular fluid persisted during 24 months despite repeated injections. Eight of 16 eyes with partial responses and two of three eyes with no response eventually developed cystoid macular oedema (CMO), and six of these lost three lines or more at 24 months compared with baseline.
Indocyanine green angiography (ICGA) and optical coherence tomography (OCT) images before and after intravitreal bevacizumab in an 83-year-old man. (A) ICGA showing branching vascular networks terminating in polypoidal lesions with a serous pigment epithelial detachment (PED) before previous photodynamic therapy (PDT). (B) OCT image confirming the PED. The polypoidal lesions have resolved completely after PDT. However, the branching vascular networks remain with subsequently developing exudation 6 months after previous PDT (C). (D) OCT showing slight subretinal fluid (SRF), macular oedema and a fibrovascular PED. The visual acuity (VA) is 0.2. Intravitreal bevacizumab was administered. After intravitreal bevacizumab, the SRF and macular oedema have resolved on the OCT image (E). The VA is 0.3. However, OCT shows recurrent macular oedema 3 months after the first injection (F). The VA is 0.2. Additional intravitreal bevacizumab was administered, and the SRF and macular oedema resolved again. The macular oedema recurred once again and was successfully treated with repeated injections during the first year. Fifteen months after the initial injection, the branching vascular networks have increased in size (G). Despite repeated intravitreal bevacizumab for recurrent exudation, the SRF and macular oedema that had once responded to treatment during the first year persistently remained during the second year with deterioration of VA to 0.1, indicating loss of the treatment efficacy (H).
The differences between complete responders and partial or non-responders with respect to BCVA and OCT findings are shown in table 2. Although the baseline BCVA, the leakage pattern on FA, CRT and the lesion size did not differ significantly between the groups, the mean BCVA in eyes with a complete response was significantly better than in eyes with a partial or no response at 24 months (p=0.041). However, there was no significant difference in the BCVA between eyes with a complete response and those with partial responses or no response at 24 months when the 10 eyes with CMO were excluded from the analysis (0.41±0.34 vs 0.34±0.30, p=0.675). The CRT also differed significantly between the groups at 24 months (p=0.001). The branching vascular networks were significantly larger in eyes with a partial response or no response compared with those who were complete responders at 24 months (p=0.049).
Differences between complete responders and partial or non-responders in best-corrected visual acuity and optical coherence tomography findings
No adverse systemic and local complications related to intravitreal bevacizumab were observed during the study period. One eye developed an RPE tear.
Discussion
In the current study, we focused on identifying the characteristics of the exudative features associated with branching vascular networks and assessed the potential efficacy of bevacizumab to treat those lesions. We found two angiographic patterns of exudative branching vascular networks at an average of 9.3±8.4 months (range 1–20) after previous PDT, that is, occult CNV-like lesions with or without fibrovascular PED (84%) and classic CNV-like leaky vascular lesions (16%).
In eyes with occult CNV-like leakage, the complex of abnormal vessels was clearly seen on early-phase ICGA, with hyperfluorescent plaques on late-stage ICGA. On OCT, the lesion was also identifiable as elevated RPE, indicating invasion of the networks beneath the RPE. Thus, PCV in this stage was clinically and angiographically indistinguishable from type 1 neovascularisation. Imamura and associates also indicated that it would be difficult to differenciate between PCV and CNV unless typical polypoidal lesion exists in the network vessels.17 Serous and haemorrhagic PEDs associated with PCV before PDT seemed to be potentially predictive of the development and progression of occult CNV-like exudative vessels, because in those eyes, the rate of PED at the time of PDT was 63% and was higher than that reported previously in PCV (19–44%).7 8 18 In eyes with classic CNV-like lesions, OCT showed subretinal fibrin formation, indicating extensive exudation. Both the occult and classic CNV-like lesions in our study did not seem to have the same characteristics as the original branching vascular networks of PCV, which had previously been considered as long-term stable abnormal choroidal vessels.18 19 We speculate that pre-existing branching vascular networks acquired hyperpermeable properties mimicking CNV or the CNV newly emerged from residual branching vascular networks over time after PDT.
Intravitreal bevacizumab for exudative branching vascular networks maintained or improved vision in 38 eyes (84%) over 12 months, although the mean VA at 12 months did not differ significantly compared with baseline. The significant decrease in retinal thickness was achieved during the first year due to complete resolution of macular fluid in 42 (93%) eyes, indicating the potent effect of bevacizumab on antipermeability in most eyes over 12 months.
Improved or stabilised VA was continuously achieved in 36 eyes (80%) over 24 months. According to a previous study of patients with PCV treated only by PDT for 3 years, VA deteriorated three lines or more in 37%, owing to the enlargement of abnormal networks with neovascular changes and recurrent polyps.20 In contrast to these reports, the current results using bevacizumab are encouraging because VA deterioration was seen in 20%, despite a mean follow-up period of 33 months after initial PDT.
Complete resolution of macular fluid was maintained throughout the study in 58% of eyes (complete responders). However, 36% of eyes that responded to treatment during the first 12 months showed less treatment effect, despite repeated injections for recurrent exudation (partial responders). Three eyes were considered non-responders on OCT, and the macular fluid increased further. As a result, the decrease in the mean retinal thickness during the first year lessened during the second year and became non-significant (p=0.296) at 24 months compared with baseline.
The loss of the therapeutic response to bevacizumab during 2 years may be explained by a potential change in the sensitivity to bevacizumab. In the current study, the vascular networks persisted on ICGA in all eyes after intravitreal bevacizumab and even enlarged in 20 eyes (44%) despite repeated injections. Continuous growth of the CNV despite repeated anti-vascular endothelial growth factor therapy also had been reported in eyes with AMD.21 The persistence and expansion of the networks may lead to more mature and less vascular-endothelial-growth-factor-dependent vessels with increased treatment resistance. In addition, a tachyphylactic response, that is, a progressive decrease in the bioefficacy of bevacizumab, after repeated injections has been reported in AMD.22 Those responses also may be responsible for the loss of treatment efficacy in the current study eyes.
Fortunately, half of the eyes that were partial and non-responders maintained the VA through 24 months despite persistent macular fluid. However, 10 eyes eventually developed significant CMO with poor visual outcomes. The CMO has been reported to be associated with all forms of neovascular AMD, including classic CNV, occult CNV, PED and disciform scars,23 24 or with PCV.20 We speculated that the persistent exudative vascular networks beneath the RPE may cause RPE decompensation, resulting in severe damage to the neurosensory retina, as seen in progressed AMD.
The limitations of the current study were its retrospective nature and the absence of a control group. Because the individual responses vary among the patients, we did not re-treat patients at fixed intervals but did so with the criteria based on their responses under monthly monitoring. Some patients with a lower morphological response despite consecutive injections were not always re-treated unless VA declined, as reported in the treatment of AMD.25 To confirm whether the current strategy is optimal for long-term follow-up, a prospective, randomised, comparative study should be considered comparing different injection strategies for exudative branching vascular networks in PCV.
In summary, the intravitreal bevacizumab to treat exudative branching vascular networks in patients with PCV improves the retinal morphology and maintains vision over 1 year. However, the network vessels persist, and the therapeutic response may be lost during the second year. The visual prognosis is poor in eyes with CMO. Further studies may elucidate the appropriate use of intravitreal bevacizumab or other treatment modalities for better management of branching vascular networks in PCV.
References
Footnotes
Competing interests None.
Patient consent Obtained.
Ethics approval Ethics approval was provided by the institutional review board committee of Osaka University Hospital.
Provenance and peer review Not commissioned; externally peer reviewed.