Anti-Vascular Endothelial Growth Factor Therapy Regimens for Polypoidal Choroidal Vasculopathy: A Systematic Review

Polypoidal choroidal vasculopathy (PCV) is a subtype of type 1 choroidal neovascularization that is characterized by dilated polypoidal vascular lesions at the end of a network of large choroidal vessels [1]. PCV is most common in Asian patients, though there is evidence that it may also be common in Caucasian patients and underdiagnosed [2]. PCV is characterized by serous macular detachments, serous or hemorrhagic pigment epithelial detachments, subretinal haemorrhage, and occasionally nodular lesions [3]. PCV can mimic neovascular age-related macular degeneration (nAMD), making an accurate diagnosis and treatment difficult [4]. Thermal laser photocoagulation, photodynamic therapy (PDT), anti-VEGF monotherapy, and combination therapy with anti-VEGF and PDT are common therapeutic options for PCV [5]. Several clinical trials have shown that anti-VEGF agents are effective in PCV [4, 6]. The transition from PDT to anti-VEGF therapy in PCV began with the EVEREST randomized clinical trial comparing ranibizumab monotherapy to ranibizumab in combination with verteporfin PDT [7]. EVEREST II found that both treatments were safe and efficacious in PCV, with combination therapy achieving superior gains in best-corrected visual acuity (BCVA) compared to ranibizumab monotherapy [8]. The landmark PLANET trial reported that aflibercept monotherapy was non-inferior to aflibercept with rescue PDT treatment for visual and anatomical outcomes in PCV [9]. A randomized trial of combination PDT and anti-VEGF agents found that eyes treated with treat-and-extend (T&E) achieved better visual acuity than a pro re nata (PRN) treatment regimen after initial PDT [10].

A long-term retrospective review found that anti-VEGF monotherapy was an effective treatment in PCV patients after 6 years [11]. Options for anti-VEGF monotherapy in PCV include monthly [12], bimonthly [13], T&E [14], and PRN [15] regimens. Currently, it is not well understood which is the optimal intravitreal anti-VEGF agent and treatment regimen for PCV.

There has been no guideline that has aimed to inform clinicians on the optimal anti-VEGF monotherapy regimen for patients with PCV. To the best of our knowledge, this represents the first systematic review to assess the comparative safety and efficacy of different treatment regimens of anti-VEGF monotherapy agents for PCV.

A systematic literature search was conducted on Ovid MEDLINE, Embase, and the Cochrane Library from January 2000 to May 2023. Our search strategy was used for a multipurpose study and is detailed in online supplementary eTable 1 (for all online suppl. material, see https://doi.org/10.1159/000533529). We included peer-reviewed comparative articles, both randomized controlled trials and observational studies, reporting on different treatment regimens of anti-vascular endothelial growth factor (anti-VEGF) agents at the same dose for the management of PCV. Studies must have reported on at least 10 eyes to justify inclusion. We excluded studies with eyes that received intravitreal corticosteroids and laser treatment for PCV, or studies in which the use of PDT was reserved to one treatment arm. We also excluded editorials, conference abstracts, letters to the editor, and non-comparative studies. We excluded follow-up data from patients who switched treatment regimens during studies. Our systematic review was registered in the International Prospective Register or Systematic Reviews (PROSPERO) with the ID number CRD42022345465. Our study adhered to the Declaration of Helsinki, and we did not seek Research Ethics Board approval due to the study design.

Three independent authors (A.M., A.H., N.S.P.) performed title and abstract screening, full-text screening, and data collection. A fourth impartial reviewer (M.M.P.) was consulted for conflict resolution. Our primary outcomes were BCVA at last study observation and the change in BCVA from baseline. The secondary outcomes were retinal thickness (RT) at last study observation, the change in RT from baseline, the rate of polyp closure, and the incidence of adverse events. Studies were screened in Covidence (Veritas Health Innovation, Melbourne, Australia), and data were collected using Microsoft Excel^® (Microsoft Corporation, Redmond, WA, USA).

The risk of bias for randomized controlled trials was assessed using the ROB-2 tool (online suppl. eTable 2). The following parameters were assessed for randomized trials: random sequence generation, allocation concealment, masking of participants and personnel, masking of outcome assessment, incomplete outcome data, selective reporting, conflicts of interest, industry sponsorship, and other bias. Observational studies were assessed for risk of bias using the ROBINS-1 tool (online suppl. eTable 3). The following parameters were assessed for non-randomized studies: confounding, selection of participants, classification of interventions, deviations from intended interventions, missing data, measurement of outcomes, and selection of reported results. Risk of bias evaluation was conducted by three independent reviewers (A.M., A.H., N.S.P.).

A total of 10,784 studies underwent title and abstract screening, 133 studies underwent full-text review, and seven studies reporting on 636 eyes with PCV were included in this systematic review (Fig. 1) [10, 16‒21]. Baseline demographic and patient characteristics are summarized in Table 1, and study outcomes are summarized in Table 2.

The outcomes described herein were reported across at least two included studies over a follow-up period of 1 year. The mean BCVA at 12 months was 0.39 ± 0.41 logMAR (∼20/50 Snellen, n = 43 eyes; 3 studies) across study arms administering anti-VEGF agents using a PRN regimen [10, 19, 21], 0.41 ± 0.30 logMAR (∼20/50 Snellen, n = 144 eyes, 3 studies) across arms using a T&E regimen [10, 17, 18], and 0.17 ± 0.28 logMAR (∼20/32 Snellen, n = 61 eyes, 3 studies) across arms using a fixed bimonthly regimen [17, 19, 21]. The mean improvement in BCVA from baseline was 0.08 ± 0.25 logMAR (mean follow-up = 12 months, n = 33 eyes, 2 studies) across study arms administering anti-VEGF agents using a PRN regimen [10, 19], 0.15 ± 0.26 logMAR (mean follow-up = 11.5 months, n = 270 eyes, 4 studies) across arms using a T&E regimen [10, 17, 18, 20], and 0.15 ± 0.21 logMAR (mean follow-up = 12 months, n = 38 eyes, 2 studies) across arms using a fixed bimonthly regimen [17, 19]. The mean RT at 12 months was 253.93 ± 110.71 μm (n = 27 eyes, 2 studies) across study arms administering anti-VEGF agents using a PRN regimen [19, 21], 244.05 ± 83.48 μm (n = 130 eyes, 2 studies) across arms using a T&E regimen [17, 18], and 238.60 ± 74.89 μm (n = 61 eyes, 3 studies) across study arms using a fixed bimonthly regimen [17, 19, 21]. The mean improvement in RT from baseline was 137.35 ± 158.81 μm (mean follow-up = 11.5 months, n = 256 eyes, 3 studies) across study arms administering anti-VEGF agents using a T&E regimen [17, 18, 20]. Overall, the complete polyp closure rate was 44.6% (n = 112/251 eyes, 3 studies) across study arms administering anti-VEGF agents using a T&E regimen [17, 18, 20] and 45.9% (n = 17/37 eyes, 2 studies) across arms using a fixed bimonthly dosing regimen [17, 19].

The DRAGON study by Li et al. [16] was a phase 4 Chinese randomized controlled trial of 139 eyes with PCV treated with 0.5 mg ranibizumab using a monthly or PRN treatment regimen. In the monthly treatment regimen cohort, 70 eyes received monthly injections up to month 11, followed by PRN injections up to month 23. In the PRN group, 69 eyes received monthly injections until patients achieved a stable visual acuity for three consecutive visits, followed by PRN injections up to month 23. No statistical comparison was conducted between groups for baseline characteristics. Data beyond month 12 were not analysed as both treatment arms received PRN. At 12 months, 63 eyes remained in the monthly group and 65 in the PRN group. Up to month 11, the mean number of injections was 11.2 in the monthly group and 8.4 in the PRN group. At 12 months, the mean change in BCVA was +12.7 letters for the monthly ranibizumab group and +9.4 letters in the PRN ranibizumab group. The proportion of patients achieving a BCVA of ≥69 ETDRS letters was 48.6% in the monthly group and 47.8% in the PRN group. The mean change in CSFT was −180.0 μm in the monthly group and −152.9 μm in the PRN group. The proportion of patients achieving a dry retina was 50.0% in the monthly group and 30.4% in the PRN group. Safety outcomes were not stratified by presence of PCV and may have also included eyes with nAMD.

The ALTAIR study by Okada et al. (2022) was a phase 4 Japanese randomized controlled trial of 246 eyes with exudative age-related macular degeneration treated with 2 mg aflibercept using two different T&E treatment regimens [18]. The included data from the ALTAIR study were a subgroup of 90 PCV eyes from an overall sample of 246 eyes in the trial [22]. Patients received three initial monthly doses of aflibercept. 46 eyes received aflibercept extended by 2 weeks and 44 eyes continued to receive aflibercept extended by 4 weeks if extension criteria were met. Treatment intervals were shortened with new or persistent fluid, loss of ≥5 ETDRS letters, increase in CRT of ≥100 μm, new onset of neovascularization, macular haemorrhage, or fluid [22]. Treatment intervals were maintained if shortening criteria were not met and residual fluid decreased, and treatment intervals were extended if there was no fluid on OCT [22]. No statistical analysis was conducted between groups for baseline characteristics. Up to week 52, the mean number of injections was 7.1 ± 1.0 in the 2-week group and 6.7 ± 1.1 in the 4-week group. Up to week 96, the mean number of injections was 10.0 ± 2.9 in the 2-week group and 9.9 ± 2.5 in the 4-week group. The mean BCVA at week 52 was 60.1 ± 18.5 letters (∼20/63 Snellen) in the 2-week group and 65.6 ± 13.0 letters (∼20/50 Snellen) in the 4-week group. The mean BCVA at 96 weeks was 56.3 ± 21.1 letters (∼20/80 Snellen) in the 2-week group and 62.3 ± 18.9 letters (∼20/63 Snellen) in the 4-week group. The mean change in BCVA from baseline to week 52 was +7.5 ± 14.7 letters in the 2-week group and +8.2 ± 11.6 letters in the 4-week group. The mean change in BCVA from baseline at 96 weeks was +3.7 ± 18.5 letters in the 2-week group and +4.9 ± 17.7 letters in the 4-week group. The mean CRT at 52 weeks was 251 ± 73 μm in the 2-week group and 260 ± 114 μm in the 4-week group. The mean CRT at 96 weeks was 251 ± 67 μm in the 2-week group and 263 ± 91 μm in the 4-week group. The mean change in CRT from baseline at 52 weeks was −153 ± 177 μm in the 2-week group and −112 ± 132 μm in the 4-week group. The mean change in CRT from baseline at 96 weeks was −153 ± 179 μm in the 2-week group and −110 ± 112 μm in the 4-week group. The proportion of patients with no fluid on OCT was 60.9% in both the 2-week and the 4-week group at 52 weeks. The proportion of patients with no fluid on OCT was 60.9% in the 2-week group and 70.5% in the 4-week group at 96 weeks. The proportion of patients with complete polyp regression at 52 weeks was 47.8% in the 2-week group and 52.3% in the 4-week group. The proportion of patients with complete polyp regression at 96 weeks was 56.5% in the 2-week group and 50.0% in the 4-week group. 3 patients in the 2-week group (6.5%) and 4 patients in the 4-week group (9.1%) developed cataract. 1 patient in the 2-week group developed glaucoma (2.2%) and 1 patient in the 4-week group developed ocular hypertension (2.3%). One patient in the 2-week group (2.2%) and 4 patients in the 4-week group (9.1%) developed subconjunctival haemorrhage.

The study by Teo et al. [17] was a phase 4 randomized controlled trial in Singapore of 53 eyes with PCV treated with 2 mg aflibercept using a personalised treatment regimen of T&E with a 6-month loading phase (n = 40) or a fixed treatment regimen (n = 13). All patients received three injections of aflibercept at week 0, 4, and 8. Patients in the personalised treatment regimen received monthly aflibercept until week 24 and began T&E afterwards. Patients in the fixed treatment regimen received aflibercept every 8 weeks. No statistical analysis was conducted between groups for baseline characteristics. The mean number of injections was 8.2 ± 0.9 in the personalised group and 8.0 ± 0 in the fixed group (p = 0.51). At the final follow-up of 52 weeks, the mean BCVA was 67.5 ± 13.6 letters (∼20/40 Snellen) in the personalised group and 71.5 ± 16.7 letters (∼20/40 Snellen) in the fixed group. The two treatment regimens were similar in final BCVA (p = 0.53). The mean change in BCVA from baseline to 52 weeks was +8.9 ± 13.4 letters in the personalised group and +6.0 ± 13.9 letters in the fixed group. The two treatment regimens were similar in the change in BCVA (p = 0.51). The proportion of patients that gained ≥15 letters (p = 1.0) or lost ≥15 letters (p = 1.0) at 52 weeks of follow-up were similar between the two treatment regimens. Moreover, the proportion of patients with a BCVA of ≥70 letters (p = 0.46) or ≤35 letters (p = 1.0) at 52 weeks of follow-up were similar between the two treatment regimens. The mean CSFT at 52 weeks was similar between the two treatment regimens (p = 0.08), with 218.5 ± 39.5 μm for the personalised group and 285.9 ± 97.2 μm for the fixed group. The mean change in CSFT from baseline to week 52 was significantly greater in the personalised group (−248.8 ± 169.9 μm) than the fixed group (−164.8 ± 148.9 μm, p = 0.03). The proportion of patients with subretinal fluid at last study observation was similar between the two treatment regimens (18.4% in the personalised group and 16.6% in the fixed group, p = 0.89). The proportion of patients with intraretinal fluid at last study observation was similar between the two treatment regimens, with 10.0% in the personalised group and 16.6% in the fixed group (p = 0.63). The proportion of eyes with complete regression of polypoidal lesions at last study observation was similar between the two treatment regimens (55.2% in the personalised group and 41.6% in the fixed group, p = 0.41). Nevertheless, there was a significantly greater number of polypoidal lesions closed per eye in the personalised group (1.7 ± 0.9 polyps closed) than in the fixed group (1.1 ± 0.5 polyps closed, p = 0.02).

The study by Maruyama-Inoue et al. [21] was a Japanese retrospective review of 33 eyes at baseline treated with 2.0 mg aflibercept using a fixed bimonthly dosing regimen (n = 23) or a PRN treatment regimen (n = 10). The following baseline characteristics were similar between patients in the fixed-dosing group and PRN group: mean age (p = 0.30), proportion of males (p = 0.4), mean BCVA (p = 0.55), mean CSFT (p = 0.77), and presence of subretinal fluid (p = 1.0). In the PRN group, there was no significant difference between the BCVA at baseline and at month 12 (0.16 ± 0.27 logMAR or ∼20/32 Snellen, p > 0.05). BCVA was similar between the fixed-dosing and PRN groups at 12 months (p = 0.35). The mean CSFT in the fixed-dosing group was significantly lower compared to baseline at month 12 (231 ± 40 μm, p < 0.001). The mean CSFT in the PRN group was significantly lower compared to baseline at month 12 (230 ± 66 μm, p = 0.01). Statistical comparisons were not performed between groups for CSFT. The presence of subretinal fluid at month 12 was 13.0% in the fixed-dosing group and 30.0% in the PRN group.

The article by Inoue et al. [19] from Japan prospectively recruited 42 eyes treated with 2.0 mg aflibercept using a fixed-dosing regimen of injections every 8 weeks (n = 25) or a PRN treatment regimen (n = 17). All patients received 3 initial monthly injections of aflibercept. The following baseline characteristics were similar between patients in the fixed-dosing group and PRN group: mean age (p = 0.95), proportion of males (p = 0.97), mean BCVA (p = 0.78), mean CFT (p = 0.78), and presence of subretinal fluid (p = 0.60). The PRN group (5.0 ± 2.9 injections) had a significantly lower mean number of injections compared to the fixed group (7 injections) (p < 0.01). The mean BCVA at 12 months (0.15 ± 0.27 logMAR or ∼20/25 Snellen, p < 0.01) in the 8-week group was significantly better compared to baseline. The mean BCVA at 12 months (0.24 ± 0.34 logMAR or ∼20/32 Snellen, p = 0.03) in the PRN group was significantly better compared to baseline. The mean change in BCVA was similar between the two treatment regimens at month 12 (p = 0.24). The mean CSFT was 221 ± 79 μm in the 8-week group and 268 ± 130 μm in the PRN group at 12 months. The mean CSFT at 12 months was significantly lower compared to baseline in both groups (p < 0.01); however, statistical analysis was not conducted between groups. The proportion of eyes with complete resolution of polypoidal lesions was similar between the two treatment regimens (48.0% in the 8-week group and 52.9% in the PRN group, p = 0.50).

The STAR study by Li et al. [20] was a phase 4 randomized controlled trial in China of 249 eyes with PCV treated with 0.5 mg conbercept using a fixed-dosing regimen every 12 weeks (n = 123) or a T&E treatment regimen (n = 126). All patients received 3 initial monthly injections of conbercept. The following baseline characteristics were similar between patients in the fixed-dosing group and T&E group: mean age (p = 0.85), proportion of males (p = 0.45), BCVA (p = 0.62), CRT (p = 0.55), presence of intraretinal fluid (p = 0.78), and number of polyps (p = 0.71). There was a significantly greater proportion of eyes in the fixed-dosing group (96.7%) with subretinal fluid at baseline than that in the T&E group (90.5%, p = 0.04). Through 48 weeks of follow-up, the mean number of injections in the fixed-dosing group was 6.6 ± 0.9 and in the T&E group was 9.4 ± 2.0. 14 eyes in the fixed-dosing group and 9 eyes in the T&E group went on to receive rescue PDT by week 48. The mean change in BCVA from baseline to week 48 was similar between the fixed-dosing group (5.18 ± 16.24 letters) and T&E group (6.29 ± 13.23 letters, p = 0.42). The mean change in CRT from baseline to week 48 was also similar between the fixed-dosing group (−94.4 ± 151.5 μm) and T&E group (−105.1 ± 143.9 μm, p = 0.82). Compared to baseline, the average number of polypoidal lesions at week 48 was significantly reduced in the fixed-dosing and T&E groups (by 42% and 47%, respectively; p < 0.001). The regression of polypoidal lesions was similar between the two treatment regimens (p = 0.81), with 43.8% of polypoidal lesions completely regressing in the fixed-dosing group and 41.8% in the T&E group. The incidence of agent-related adverse events was similar between treatment regimens, involving 7 patients in the fixed-dosing group and 8 patients in the T&E group.

The article by Rouvas et al. [10] from Greece prospectively recruited 30 eyes treated with 2.0 mg aflibercept using a T&E regimen (n = 14) or a PRN treatment regimen (n = 16). All patients received 3 initial monthly injections of aflibercept and one PDT session. The following baseline characteristics were similar between patients in the fixed-dosing group and T&E group: mean age (p = 0.14), proportion of males (p = 0.96), and BCVA (p = 0.34). The mean number of injections in the T&E group was 6 ± 0 in the maintenance phase, which was significantly greater than the mean number of injections in the PRN group (5.13 ± 1.08, p = 0.006). The mean BCVA in the T&E group at 12 months (0.41 ± 0.15 logMAR) was significantly improved compared to baseline (p = 0.04). In the PRN group, mean BCVA worsened throughout the study period and at 12 months (0.70 ± 0.36 logMAR) was similar to baseline (p = 0.61). The T&E group (−0.16 ± 0.13 logMAR) achieved a significantly better change in BCVA from baseline than the PRN group (+0.05 ± 0.31 logMAR, p = 0.02). The PRN group experienced a significantly greater incidence of polypoidal lesion recurrence than the T&E group (1.37 ± 0.5 vs. 0, p < 0.001).

This systematic review investigated the comparative safety and efficacy of different treatment regimens of monotherapy with anti-VEGF agents in eyes with PCV. PDT has been shown to be an effective treatment for PCV; however, its questionable utility, the increasing difficulty in maintaining Visudyne supply, laser maintenance, and logistics required to administer PDT has shifted PCV treatment towards anti-VEGF monotherapy. Data from clinical trials suggest that anti-VEGF agents are effective in the management of PCV; however, the optimal treatment regimen remains unclear [5]. T&E regimens are becoming common in PCV, and guidelines for retreatment are typically based on OCT features adopted from nAMD [5]. Due to clinical similarities between PCV and nAMD, accurate diagnosis and treatment remains challenging [23]. In general, our systematic review found that visual and anatomical outcomes may be similar between different treatment regimens of anti-VEGF agents, currently used in PCV management, yet stronger evidence is needed for more nuanced treatment decision making. Hence, there is a lack of clarity as to which anti-VEGF regimen is most advantageous in PCV.

T&E regimens with anti-VEGF agents are a preferred treatment modality in nAMD, given the ability to have reduced patient burden from injections and clinic visits while maintaining high visual acuity gains [24, 25]. The CANTREAT and TREND trials found that ranibizumab T&E in nAMD achieved comparable visual outcomes to fixed monthly dosing with fewer injections [26, 27]. T&E regimens of aflibercept and brolucizumab have also shown a high level of efficacy [28, 29]. T&E regimens of anti-VEGF agents are increasing in popularity in PCV and show favourable outcomes [5, 14]. Although there is currently a lack of guidelines on the use of T&E in PCV, the Asia-Pacific Vitreo-Retina Society recommends T&E in PCV [25]. Furthermore, anti-Ang-2/VEGF therapies are becoming increasingly popular in nAMD management, as they may reduce treatment burden by extending the time between treatments. For instance, the TENAYA and LUCERNE trials found that fixed dosing of faricimab in nAMD achieved comparable visual outcomes to bimonthly aflibercept with fewer injections [30].

The EVEREST trial in 2012 found that complete polyp regression rates at 6 months were significantly better when eyes with PCV were treated with verteporfin PDT alone (71.4%) or in combination with ranibizumab (77.8%) compared to ranibizumab monotherapy (28.6%) [7]. Similar results were found in the EVEREST II trial in 2017, where the complete polyp regression rates at 6 months were 69.7% for eyes treated with verteporfin PDT in combination with ranibizumab and 33.8% for eyes treated with ranibizumab monotherapy [31]. Across study arms using different treatment regimens of anti-VEGF monotherapy published from 2016 to 2022, our systematic review found a complete polyp closure rate ranging from 41.6% to 56.5%.

The limited number of studies reporting on different treatment regimens in our systematic review prevented meta-analysis. Statistical analysis was variably performed across studies. Inconsistent PRN and T&E treatment regimens were used between studies. Most included studies reported on aflibercept, one reported on ranibizumab, and one on conbercept. Two studies allowed for patients to receive PDT across both treatment arms [10, 20]. Our included studies varied in the length of last study observation, and there was a paucity of data on adverse events. Our analysis may be limited by shorter follow-up periods of included studies, and it is possible that some recurrences of PCV were not captured in the presented data. The lens status of patients was not reported in included studies. Conclusions of this systematic review should be interpreted at the level of the cohort and may not be applicable to the patient level.

Our systematic review of 636 eyes comparing different anti-VEGF treatment regimens in PCV suggests that current treatment regimens are similarly efficacious, although T&E may achieve superior outcomes in some patients. Nevertheless, further trials are needed for a stronger understanding of different regimens, which may allow for improved patient outcomes and a more individualized approach to managing PCV.

An ethics statement is not applicable because this study is exclusively based on a published literature. The study is in accordance with the World Medical Association Declaration of Helsinki.

A.M., A.H., N.S.P.: none. M.M.P.: financial support (to institution) – PSI Foundation, Fighting Blindness Canada. R.H.M.: Advisory Board – Alcon, Bausch + Lomb, Bayer, Novartis, Allergan, Roche; financial support (to institution) – Bayer, Novartis, Roche. P.J.K.: Advisory Board – Novartis, Alcon, Bayer, Roche, Allergan, Novelty Nobility; financial support (to institution) – Bayer, Roche, Novartis; financial support – Novartis, Bayer, Roche, Boehringer Ingelheim, Pfizer, Zeiss; equity owner – ArcticDx. D.T.W.: consultant – AbbVie, Alcon, Apellis, Bayer, Bausch Health, Biogen, Boehringer Ingelheim, Novartis, Ripple Therapeutics, Roche, Topcon, Zeiss. Research grants from Novartis, Roche.

No specific grants were received from any funding agency for this research.

Andrew Mihalache, Amin Hatamnejad, and Nikhil S. Patil: conception of the study, data acquisition, statistical analysis, and drafting of manuscript; Marko M. Popovic, Rajeev H. Muni, Peter J. Kertes, and David T. Wong: conception of the study, revision for intellectual content, and final approval.

All generated data are included in this paper and its supplementary material files. Further enquiries are to be directed to the corresponding author.