Dexamethasone implant improves anatomic response to anti-VEGF therapy in treatment-resistant polypoidal choroidal vasculopathy

A significant proportion of eyes with polypoidal choroidal vasculopathy (PCV) can be resistant to anti-vascular endothelial growth factor (VEGF) injections. We evaluated the efficacy of a combination of dexamethasone intravitreal implant (DXI) and anti-VEGF therapy in eyes resistant to anti-VEGF monotherapy. In this retrospective study, patients with PCV resistant to anti-VEGF injections were additionally injected with a DXI along with an anti-VEGF agent. Best-corrected visual acuity (BCVA), slit-lamp examination, fundus evaluation, and optical coherence tomography (OCT) data were analyzed. Anatomical response on OCT was the primary outcome measure. Change in visual acuity and injection-free interval after DXI were evaluated as secondary outcome measures. Twelve eyes of 11 patients were included in the study. Mean age of patients at presentation was 64.7 ± 9.5 years (range, 49–78.8 years), and there were seven females (63.6%). Median number of anti-VEGF injections prior to DXI was 4 (interquartile range IQR, 3–7). Median follow-up duration after DXI was 32.2 months (IQR, 6.6–41.6 months). Median logMAR BCVA immediately prior to DXI was 0.41 (IQR, 0.30–0.88) and after injection was 0.40 (IQR, 0.30–1.05), which was not significantly different (p = 0.85). Median Central Retinal Thickness (CRT) after DXI was 305.5 µm (IQR, 249–409 µm), which was significantly (p = 0.003) lesser than pre-injection thickness of 547 µm (IQR, 431–771 µm). Median injection-free interval in these eyes after DXI was 5 months (IQR, 2.8–6.4 months). Kaplan–Meier estimates of first injection after DXI were 27.3% at 3 months, 67.3% at 6 months, and 89.1% at 12 months. Dexamethasone implant combined with anti-VEGF treatment can prolong the treatment-free interval in eyes with PCV resistant to anti-VEGF injection while maintaining visual acuity.


Abstract
Background A significant proportion of eyes with polypoidal choroidal vasculopathy (PCV) can be resistant to anti-vascular endothelial growth factor (VEGF) injections. We evaluated the efficacy of a combination of dexamethasone intravitreal implant (DXI) and anti-VEGF therapy in eyes resistant to anti-VEGF monotherapy. Methods In this retrospective study, patients with PCV resistant to anti-VEGF injections were additionally injected with a DXI along with an anti-VEGF agent. Best-corrected visual acuity (BCVA), slit-lamp examination, fundus evaluation, and optical coherence tomography (OCT) data were analyzed. Anatomical response on OCT was the primary outcome measure. Change in visual acuity and injection-free interval after DXI were evaluated as secondary outcome measures. Results Twelve eyes of 11 patients were included in the study. Mean age of patients at presentation was 64.7 ± 9.5 years (range, 49-78.8 years), and there were seven females (63.6%). Median number of anti-VEGF injections prior to DXI was 4 (interquartile range IQR, [3][4][5][6][7]. Median follow-up duration after DXI was 32.2 months (IQR, 6.6-41.6 months). Median logMAR BCVA immediately prior to DXI was 0.41 (IQR, 0.30-0.88) and after injection was 0.40 (IQR, 0.30-1.05), which was not significantly different (p = 0.85). Median Central Retinal Thickness (CRT) after DXI was 305.5 lm (IQR, 249-409 lm), which was significantly (p = 0.003) lesser than pre-injection thickness of 547 lm (IQR, 431-771 lm). Median injection-free interval in these eyes after DXI was 5 months (IQR, 2.8-6.4 months). Kaplan-Meier estimates of first injection after DXI were 27.3% at 3 months, 67.3% at 6 months, and 89.1% at 12 months. Conclusions Dexamethasone implant combined with anti-VEGF treatment can prolong the treatment-free interval in eyes with PCV resistant to anti-VEGF injection while maintaining visual acuity.

Introduction
Intravitreal injection of anti-vascular endothelial growth factor (VEGF) is the standard of care for choroidal neovascular membrane (CNVM) due to neovascular age-related macular degeneration (nAMD) and polypoidal choroidal vasculopathy (PCV) [1][2][3][4]. However, a significant portion of eyes treated with anti-VEGF show little or no response to the therapy [5,6]. Estimates of the prevalence of this anti-VEGF resistance among eyes with CNVM vary significantly in the literature based on anti-VEGF drug of choice, specific sub-type diagnosis such as nAMD or PCV, treatment style (as-needed vs treat-andextend vs monthly regimen), and the ethnic population [6][7][8][9].
Various inflammatory cytokines such as macrophage-derived chemokine (MDC), interleukins, and monocyte chemotactic protein levels were significantly higher in eyes with PCV [10]. Pharmacologic therapies which reduce inflammation in the eye may improve the outcomes in eyes with PCV. We hypothesize that the dexamethasone implant (Ozurdex (DXI); Allergan, Irvine, California, USA), combined with anti-VEGF therapy, could improve the response to anti-VEGF agents in treatment-resistant PCV. A study by Kuppermann et al. combining DXI with anti-VEGF agents in patients with nAMD demonstrated that the injection-free interval in patients receiving DXI along with anti-VEGF therapy was significantly longer than in patients receiving placebo with anti-VEGF therapy [11]. However, the LuceDex trial did not show any difference in the number of injections between combination therapy and anti-VEGF monotherapy, although there was more visual gain in the combination therapy arm [12]. A few other studies have evaluated the efficacy of a combination treatment of DXI and anti-VEGF in nAMD patients resistant to anti-VEGF, but with contrasting outcomes [13][14][15]. None of these patients included PCV. The purpose of this study was to evaluate visual outcomes and retreatment intervals after combination of Dexamethasone intravitreal implant (DXI) and anti-VEGF therapy in eyes resistant to anti-VEGF therapy in PCV.

Materials and methods
The study protocol adhered to the tenets of the Declaration of Helsinki and was approved by the local Institutional Review Board (LEC 05-18-089) at L.V. Prasad Eye Institute. All study participants gave a written informed consent before enrollment. Patients were recruited from March 2017 through November 2019. previous 6 months; (4) previous vitreoretinal surgery in the study eye or anticipated surgery within 12 months of enrollment.

Ocular evaluation
All participants had a comprehensive ocular examination including best-corrected visual acuity (BCVA) testing, intraocular pressure (IOP) assessment, dilated fundus examination with slit-lamp biomicroscopy, color fundus photography, fundus fluorescein (FFA) and indocyanine green angiography (ICGA, Heidelberg Engineering GmbH, Heidelberg, Germany), and optical coherence tomography (OCT, Triton DRI, Topcon, Japan). Patients underwent OCT at subsequent visits. All eyes with PCV were confirmed by ICGA.

Intravitreal injections
Intravitreal injections were administered using a strict aseptic technique under topical anesthesia in an outpatient procedure room. Intravitreal injections were performed with 29-gauge needle inserted through the inferotemporal pars plana, 4 mm posterior to the limbus in phakic eyes, and 3.5 mm in pseudophakic eyes. In eyes receiving combination treatment, DXI was injected after the anti-VEGF injection in the same sitting.

Outcome measures
Since this study was performed in resistant PCV, anatomical response on OCT was the primary outcome measure. Change in visual acuity and injectionfree interval after the dexamethasone implant were evaluated as secondary outcome measures.

Statistical analysis
Data were analyzed using STATA v14.2 (StataCorp, College Station, Texas, USA). The continuous data were checked for the normality of distribution by Shapiro-Wilk test and described in mean ± standard deviation (with range) if normally distributed and median with interquartile range (IQR) if otherwise. Categorical data were described in proportions. Preand post-injection BCVA and central retinal thickness (CRT) were compared by a mixed effects model analysis of longitudinal data. Relationships between continuous data were assessed by Spearman correlation. PED, IRF, and SRF were compared pre-and postinjection by McNemar test. Kaplan-Meier survival analysis was used to estimate the probability of first

Results
Twelve eyes of 11 patients were included in the study. Mean age of patients at presentation was 64.7 ± 9.5 years (range, 49-78.8 years), and there were seven females (63.6%). The median duration of symptoms was 45 days (IQR, 15-90 days). The median duration from presentation to DXI was 1.4 years (IQR, 0.6-4.7 years). The median number of anti-VEGF injections prior to DXI was 4 (IQR, 3-7). The median duration from first prior anti-VEGF injection to DXI was 1 year (IQR, 0.7-2.4 years). Prior to DXI, four eyes (33.3%) had a history of photodynamic therapy and two eyes (16.7%) had intravitreal triamcinolone acetate. Lens status was not documented in one eye. In the remaining 11 eyes, lens was clear in 1 (9.1%), cataractous in 5 (45.45%), and pseudophakic in 5 (45.45%) prior to DXI.
The median follow-up duration after DXI was 32.2 months (IQR, 6.6-41.6 months). Following DXI, the eye with clear lens had the same lens status until the last visit. Among the five eyes with cataract, three eyes had cataract until last visit and two eyes underwent phacoemulsification with intraocular lens implantation during the follow-up. Mean IOP before DXI was 14.7 ± 2.5 mm Hg (range, 10-18 mm Hg), and one month after the injection was 14.3 ± 2.6 mm Hg (range, 10-18 mm Hg) (p = 0.44).

Retinal fluid compartments
The proportion of eyes with PCV having pigment epithelial detachment (PED) before injection was 76.9%, which reduced to 53.8% and 30.8% after 1 st month and final follow-up, respectively. At the last visit, the proportion of eyes with PED was 7.7% and a statistical significance (p = 0.03) was noted in the reduction of PED in eyes with PCV. The proportion of eyes having intraretinal fluid and subretinal fluid was also reduced at the last visit which was significant (p = 0.01). (Fig. 3). Figures 4 and 5 show representative OCT images from two patients who initially showed no response to anti-VEGF monotherapy, yet showed good response and reduction in CRT after combination treatment with DXI and anti-VEGF.
The median injection-free interval in these eyes after DXI was 5 months (IQR, 2.8-6.4 months). The injection-free interval was not correlated with number of injections (p = 0.73 and pre-injection CRT (p = 0.61). The survival analysis estimates of first injection after DXI were 27.3% at 3 months, 67.3% at Fig. 4 The patient had received two monthly ranibizumab injections with no response (3a and 3b show OCT images after each injection). Figure 3c and d are OCT images taken after the combination of ranibizumab and DXI injections, which showed good response with no intraretinal or subretinal fluid three weeks and two months after the injection, respectively 6 months, and 89.1% at 12 months. (Fig. 6) The median number of anti-VEGF injections post-DXI was 2.5 (IQR, 1-6). Post-DXI, two eyes (16.7%) received intravitreal triamcinolone acetate four times each, while three eyes (25%) received DXI implant. Post-DXI, three eyes (25%) had photodynamic therapy (Table 1).

Discussion
Intravitreal steroid injections in resistant nAMD have been studied with variable results [13][14][15][16][17][18][19]. Inflammation is known to play a key role in the pathogenesis of PCV and nAMD [10,20]. Oxidative stress has been shown to initiate the assembly of inflammasome complexes in the retinal pigment epithelium [21]. Complement components C3a and C5a have been shown to promote CNVM [22,23]. Corticosteroids counteract macrophages and related cytokines involved in inflammation and neovascularization Fig. 5 The patient had received four bevacizumab and one ranibizumab injections with poor response. (Fig. 4a shows OCT images after these injections). Figure 4b and c show complete resolution of macular fluid after ranibizumab and DXI implant, one and six months later, respectively [24,25]. There is also evidence that a combination of intravitreal corticosteroids and anti-VEGF agents can decrease tachyphylaxis, or the decline in potency of these medications after a number of injections [26].
Though intravitreal dexamethasone could have substantial impact, it is cleared from the eye quickly with a half-life of less than 3.5 h. [27] In contrast, the dexamethasone intravitreal implant used in this study had biological activity in the eye for 4-6 months [28]. A small retrospective study on anti-VEGF-resistant eyes also found that CRT declined significantly, which was held up until 3 months post-implantation [13]. This study also reported a total absence of any fluid 3 months after dexamethasone injection in 71.4% of their patients.
The stability of BCVA after DXI in these anti-VEGF-resistant PCV patients in our study is understandable, given the fact that DXI was given after anti-VEGF therapy had been started. It is possible that the usual decline in BCVA for these patients associated with the normal course of PCV on anti-VEGF therapy has been slowed by DXI. If this is the case, vision stabilization in these patients with a reduced burden of injections would indeed be a desirable outcome.
Given that the frequency of anti-VEGF injections poses a huge burden on patients, reducing the burden of injections by increasing the time between injections must be a goal of future therapy [29]. However, for the significant portion of patients who fail to respond to anti-VEGF therapy, combination with steroids could potentially have a role in reducing the resistance to anti-VEGF therapy. Our study in eyes with PCV demonstrates that the burden of frequent injections can be reduced with the addition of a steroid extendedrelease implant by significantly increasing the injection-free interval to more than 5 months. This is a substantially longer duration compared to other studies in nAMD and promises to reduce the treatment burden significantly. Our study shows that intravitreal steroid implants can also lead to a significant reduction of macular edema. In all but two patients in this study, CRT declined during the month after the implant was placed. Significantly fewer patients had any fluid in any compartment on OCT after the implant.
The limitations of our study include the relatively small sample size as well as the fact that it was a retrospective study. A larger, randomized, prospective study of PCV would be more clearly able to evaluate the effects of DXI implant used in combination with anti-VEGF therapy as well as determine what biomarkers or trends are unique to patients who still fail to respond significantly to this dual therapy. Our

Conclusion
This study shows that the addition of intravitreal dexamethasone implant in PCV, resistant to anti-VEGF, could stabilize visual acuity and partially restore foveal anatomy. Further studies with larger sample size should also look into biomarkers of eyes resistant to anti-VEGF.