Article Text

Download PDFPDF

Efficacy of intravitreal bevacizumab for polypoidal choroidal vasculopathy
  1. F Gomi,
  2. M Sawa,
  3. H Sakaguchi,
  4. M Tsujikawa,
  5. Y Oshima,
  6. M Kamei,
  7. Y Tano
  1. Department of Ophthalmology, Osaka University Medical School, Osaka, Japan
  1. Fumi Gomi, Department of Ophthalmology, Osaka University Medical School, 2-2 Yamada-oka, Suita, Osaka 565-0871, Japan; fgomi{at}


Aims: The aim of the study was to assess the short-term efficacy of intravitreal injections of bevacizumab for polypoidal choroidal vasculopathy (PCV).

Methods: Intravitreal bevacizumab (1 mg) was injected into 11 eyes of 11 patients with PCV in this retrospective, interventional case series. The main outcome measure was the change in the polypoidal vessels on indocyanine green angiography (IA) 3 months after injection. The foveal height determined by optical coherence tomography and the best-corrected visual acuity (BCVA) also were evaluated before and after treatment.

Results: At baseline, subretinal fluid was observed in five eyes and a pigment epithelial detachment in eight eyes. The foveal height 1 month after injection decreased significantly (p = 0.023), but at 3 months, no significant decrease was observed, although an additional injection was administrated in five of 11 eyes. The IA at 3 months showed resolution of polyps in one eye but residual or enlarged lesions in the other ten eyes. The BCVA did not improve significantly, although the subjects had relatively good BCVA at baseline (mean 0.45).

Conclusion: Intravitreal injection of bevacizumab may reduce the fluid from PCV but seems to be ineffective for diminishing its choroidal vascular changes.

View Full Text

Statistics from

Polypoidal choroidal vasculopathy (PCV) has a distinctive choroidal appearance characterised by an inner choroidal vascular network of vessels with polyp-like structures.15 The prevalence of PCV is high in non-White populations and has been reported to be 23–40% in Japanese patients diagnosed with presumed age-related macular degeneration (AMD).68 The pathogenesis of PCV is not fully understood; some reports have suggested that PCV is a variant of AMD,912 and others have suggested that it represents a different abnormality of the inner choroidal vasculature.1315

The natural course of PCV is reported to be more favourable compared with exudative AMD; however, Uyama et al reported that one-half of the study eyes had decreased vision due to a massive haemorrhage or severe retinal pigment epithelial (RPE) atrophy after a longer follow-up period.12 Photodynamic therapy (PDT) with verteporfin has shown good results for PCV, but extensive subretinal haemorrhages are an unavoidable side effect of PDT in some eyes with PCV.8 16 We have reported that recurrence of a PCV lesion is another complication that reduces vision after PDT.17

Recently, patients with PCV were reported to have increased concentrations of vascular endothelial growth factor (VEGF) in the aqueous humor compared with those with AMD.18 Histopathological examination also suggested an association between VEGF and PCV.19

Bevacizumab (Avastin; Genentech, San Francisco, CA, USA), a humanised monoclonal antibody that inhibits all VEGF isoforms, has shown promising results against choroidal neovascularisation (CNV) secondary to AMD.2024 If VEGF is related to the pathogenesis of PCV, bevacizumab also may be efficacious against PCV. We offered off-label administration of bevacizumab to patients with PCV and evaluated the efficacy and safety of the drug.


Eleven eyes of 11 patients with PCV who had progressing visual symptoms and had not undergone any previous treatment received a 1 mg intravitreal injection of bevacizumab at the Osaka University Hospital after the approval of the institutional review board was received. Informed consent was obtained from all patients. The diagnosis of PCV was established by the finding of polyp-like choroidal vessel dilatation with or without a branching vascular network on indocyanine green angiography (IA) using a confocal scanning laser ophthalmoscope (Heidelberg Retina Angiograph 2 (HRA2), Heidelberg Engineering, Heidelberg, Germany). Fluorescein angiography also was performed, and eyes suspected of also having CNV were excluded. An intravitreal injection of bevacizumab was administered in the same manner as reported previously.25

To assess the efficacy of intravitreal bevacizumab, the vascular findings on IA and the findings on optical coherence tomography (OCT) were compared before and after injection. The appearance, size and number of polypoidal lesions and abnormal vessels on IA were compared before and 3 months after treatment. The foveal architecture was evaluated using the OCT 3000 (Zeiss Humphrey Instruments, Dublin, CA, USA) with the cross-hair mode default setting (5.65 mm) before and 1 and 3 months after injection. Retinal thickness analysis on OCT 3000 was used to measure the foveal height including the retina and the subretinal fluid, but in the eyes with a pigment epithelial detachment (PED), the OCT images were uploaded to ImageNet (Topcon, Tokyo, Japan) and the foveal height from the presumed line of the RPE was measured by the combination of retinal thickness analysis and ImageNet software. A greater than 10% decrease from the baseline foveal height was defined as a reduction and a greater than 10% increase was defined as an increase compared with the height before treatment. The best-corrected visual acuity (BCVA) was examined using Landolt C charts and converted to the logarithm of the minimal angle of resolution (logMAR) equivalents to compare the values before and after treatment. The foveal heights and BCVA values from baseline and after injection were compared using the paired t test.


The clinical details of the patients are summarised in Table 1. Eleven patients (nine men, two women; mean age 65.4 years (range 55–78 years; mean BCVA 0.45) were included in this case series between January 2006 and December 2006. The mean follow-up time was 9.4±4.4 months.

Table 1 Patient characteristics and clinical data before and after intravitreal injection of bevacizumab for PCV

At baseline, the mean lesion size on IA was 2.95 disc areas and four of 11 eyes had a relatively large network of abnormal vessels besides polyps. Subretinal fluid was seen in five eyes and a serous or haemorrhagic PED in eight eyes. The mean foveal height was 553 μm.

One month after treatment, OCT showed decreased or resolved subretinal fluid in four of five eyes (fig 1E) and a PED that had decreased in height in six of eight eyes (fig 2E). The mean foveal height decreased significantly (p = 0.023) to 449 μm.

Figure 1 Case 10. (A) Baseline fundus photograph shows orange nodular lesion at the nasal area from the fovea and (B) IA shows polypoidal dilatation of choroidal vessels at the corresponding region of the fundus with a branching network under the fovea. (C) Three months after the first injection, the vascular findings on IA have not changed. (D) A baseline horizontal OCT image shows protruding RPE due to the polypoidal vessels and subretinal fluid and (E) 1 month after injection, the subretinal fluid reduced; however, at 3 months, the subretinal fluid recurred (F).
Figure 2 Case 1. (A) Baseline fundus photograph shows yellowish PED at the macula and the orange nodular lesion in the peripapillary area. (B) Baseline IA shows polypoidal dilatation in the peripapillary area and dark area at the macula corresponding to the PED and (C) those abnormal findings remain at 3 months after injection. (D) A baseline horizontal OCT image shows PED and at 1 month (E), the foveal height has decreased from 839 to 703 μm; however, at 3 months, the PED recurred (F).

A second bevacizumab injection was administered in five of 11 eyes 1 or 2 months after the first injection at the discretion of the treating physician. OCT at month 3 showed complete resolution of subretinal fluid in one eye (case 4) (fig 3F) but recurrent fluid in three of four eyes (fig 1F). The PED also recurred in five eyes (fig 2F) and remained in all eight eyes. The mean foveal height increased to 526 µm which was not a significant decrease from baseline. IA at month 3 showed that the polypoidal lesions had resolved in one eye (case 4) (fig 3C, D) but no apparent reduction of the lesion size was observed in the other ten eyes. To the contrary, the number of polypoidal lesions increased in three eyes (figs 4, 5). The abnormal vascular networks did not show any apparent changes in any eyes. The BCVA 1 and 3 months after injection did not change significantly from baseline (p = 0.310, p = 0.124, respectively), but three eyes had a decrease in the BCVA of three or more lines. The eye with complete resolution of polyps (case 4) had a decrease in vision due to RPE atrophy. No adverse events such as RPE tears or subretinal haemorrhages occurred.

Figure 3 Case 4. (A) The baseline fundus photograph showed subretinal fluid including the macula from the inferonasal orange nodular lesion but (B) at 3 months, subretinal fluid resolves accompanying exudates. (C) Baseline IA showed the abnormal submacular vascular network with an inferonasal polypoidal lesion. (D) Three months after injection, the polyps resolve completely while the abnormal branching vascular network remains. The subretinal fluid and the protruding RPE due to the polyps on baseline horizontal OCT image (E) are not observed at 3 months (F).
Figure 4 (A) IA findings at baseline and (B) 3 months after injection in case 11. The number of the polypoidal lesions have increased at 3 months compared with baseline and the area of the PED has enlarged.
Figure 5 (A) IA findings at baseline and (B) 3 months after injection in case 9. The upper lesions resembling a cluster of grapes have not regressed even after the second bevacizumab injection.


The pathogenesis of PCV has not been clarified. Two theories have been presented, i.e. that PCV is a feature of CNV912 or of abnormally dilated choroidal vessels.1315 An association between VEGF and PCV has been suggested by histopathological examination19 and the increased concentration of VEGF in the aqueous humor in patients with PCV.18 However, the clinical and histopathological findings would be complicated because the vascular abnormalities of PCV might change spontaneously in appearance during follow-up12 15 26 and PCV might accompany true CNV.8 12

The current study was planned to determine the short-term effect of bevacizumab for treating PCV and did not require the monthly injection. Although this study was limited by its small sample size and retrospective nature, the results suggested that one intravitreal injection of bevacizumab is ineffective for treating the choroidal vascular abnormalities of PCV. The volume of associated fluid decreased, but the effect seemed temporary. The efficacy of additional injection was unconfirmed in this small case series.

Surprisingly, three of 11 eyes had lesion progression after bevacizumab was injected. These findings implied that the choroidal vascular abnormalities of PCV have a different pathology from CNV, which is induced by and progresses as a result of VEGF.

In contrast, we previously observed remarkably better responses to PDT of PCV compared with AMD.8 Eyes with PCV had significantly better visual results, a higher rate of cessation of fluorescein leakage and required fewer additional treatments than eyes with AMD 1 year after application of PDT. This also suggests a possible difference in pathogenesis between PCV and CNV.

The reason that bevacizumab might be less efficacious for PCV, however, might be due to the limited retinal penetration of the drug, because bevacizumab is a full-length antibody and a relatively large molecule.20 21 The lack of retinal penetration of bevacizumab beyond the inner limiting membrane had been postulated previously,27 but recent findings have suggested that bevacizumab might penetrate the full retinal thickness based on an electrophysiological study.28 Our report, which confirmed the efficacy of bevacizumab for treating idiopathic CNV in young patients with a healthier retina, also suggested the drug achieved sufficient retinal penetration.29 However, the PCV lesion is located under the RPE,10 11 13 15 and sufficient drug may not reach the sub-RPE lesion. One eye with complete resolution of polyps after bevacizumab injection had an atrophic RPE; therefore, bevacizumab might have reached to the sub-RPE area and treated the polyps, although it might have been the result of the natural course of PCV.12

Another possibility for the resolution of the lesion is that the choroidal vascular abnormalities of PCV may have CNV-like characteristics during the disease course or among the individual subjects. Okubo et al reported a case of PCV that was successfully treated with a sub-Tenon’s injection of triamcinolone, which also supports this hypothesis.30

The current study showed that the single injection of bevacizumab is insufficient for the treatment of PCV. However, regular injection might maintain vision over the long term because of the drug’s anti-leakage effect on wet changes associated with PCV. The efficacy of different anti-VEGF drugs such as ranibizumab (Lucentis, Genentech), which has a smaller molecular mass and higher affinity for VEGF, should be evaluated for PCV.27 31 Further studies are needed to determine the role of anti-VEGF therapy for treating PCV.


View Abstract


  • Funding: The authors have no proprietary interest in any aspect of this report.

  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.