Aqueous Humor Cytokine Analysis in Age-Related Macular Degeneration After Switching From Aflibercept to Faricimab

Abstract

Purpose: To examine the changes in aqueous humor cytokine levels and clinical outcomes of switching from aflibercept to faricimab in eyes with neovascular age-related macular degeneration (nAMD).

Methods: Fifty-four eyes of 54 patients with AMD undergoing treatment with aflibercept under a treat-and-extend (TAE) regimen were switched to faricimab and studied prospectively. Best-corrected visual acuity (BCVA; in logarithm of the minimum angle of resolution), central retinal thickness (CRT), central choroidal thickness (CCT), and exudative status were analyzed using optical coherence tomography. Aqueous humor was collected before and after the switch, and angiopoietin-2 (Ang-2), placental growth factor (PlGF), and vascular endothelial growth factor (VEGF) A levels were measured.

Results: After switching from aflibercept to faricimab, exudative changes improved in 28 eyes (52%), remained stable in eight eyes (15%), and worsened in 18 eyes (33%). BCVA changed from 0.27 ± 0.31 to 0.26 ± 0.29 (P = 0.46), CRT decreased from 306.2 ± 147.5 µm to 278.6 ± 100.4 µm (P = 0.11), and CCT changed from 189.5 ± 92.8 µm to 186.8 ± 93.9 µm (P = 0.21). VEGF-A levels were below the detection sensitivity in many cases throughout the pre- and post-switching periods. Ang-2 significantly decreased from 23.8 ± 23.5 pg/mL to 16.4 ± 21.9 pg/mL (P < 0.001), and PlGF significantly increased from 0.86 ± 0.85 pg/mL to 1.72 ± 1.39 pg/mL (P < 0.001).

Conclusions: Switching from aflibercept to faricimab in patients with nAMD may not only suppress VEGF-A but also Ang-2 and reduce exudative changes.

Neovascular age-related macular degeneration (nAMD) is a vision-threatening disease characterized by pathological macular neovascularization. Anti-vascular endothelial growth factor (VEGF) agents are the gold standard for nAMD treatment. The anti-VEGF agents, ranibizumab and aflibercept, have been widely used for more than a decade; however, treatment resistance and the need for frequent injections remain problematic.^1,2 When administering aflibercept under a treat-and-extend (TAE) regimen, the mean dosing interval is approximately 10 weeks.^2,3 Brolucizumab, which primarily targets VEGF-A, has the smallest molecules among the available anti-VEGFs, is administered at higher concentrations, and is more effective in controlling exudative changes than aflibercept.^4,5 However, it should be used with caution because brolucizumab administration is associated with a higher incidence of intraocular inflammation (IOI) than aflibercept.⁶ 

Faricimab, which was approved by the Food and Drug Administration in February 2022, is the first bispecific antibody for intravitreal use that targets the independent inhibition of both angiopoietin-2 (Ang-2) and VEGF-A pathways by binding and neutralizing both factors.⁷ In the context of nAMD, VEGF has been recognized as a driver of neovascularization and increased vascular permeability, making it a primary therapeutic target. However, recent evidence has highlighted an additional role of Ang-2 in vascular instability. Angiopoietin-1 (Ang-1) binds to Tie-2 receptors on vascular endothelial cells, strengthening tight junctions, thereby promoting vascular stabilization. Conversely, Ang-2 competes with Ang-1 for Tie-2 receptor binding, inhibiting the effects of Ang-1 and resulting in weakened endothelial tight junctions and pericyte detachment, ultimately leading to increased vascular permeability.^8–10 The combined action of VEGF and Ang-2 exacerbate vascular leakage, potentially involving the progression of nAMD. The clinical outcomes of faricimab in treatment-naive nAMD were investigated in the TENAYA and LUCERNE study, a global Phase III clinical trial that compared the efficacy of faricimab with aflibercept.¹¹ In this study, faricimab was administered at fixed intervals of eight, 12, or 16 weeks, determined by disease activity after four initial induction treatments, and this regimen was continued for a duration of 60 weeks. The dosing intervals over the course of one year were 45% at 16 weeks, 33% at 12 weeks, and 21% at eight weeks. The mean change in the best-corrected visual acuity (BCVA) score from baseline was comparable to that in the aflibercept group at an eight-week fixed interval. 

Cytokines, such as VEGF and Ang-2, contribute to increased vascular permeability, inflammation, and growth of pathological neovascularization in AMD.^12,13 Elevated aqueous humor concentrations of VEGF are associated with reactivation of macular neovascularization or worsening central retinal thickness.^14–17 Different AMD types and the presence/absence of drusen and pachychoroid show distinct cytokine profiles.^18,19 Thus aqueous humor concentrations of cytokines may be associated with clinical pathophysiology and disease activity. This study aimed to compare the changes in aqueous humor cytokine concentrations and clinical outcomes between two different anti-VEGF agents, aflibercept and faricimab, in patients with nAMD. To the best of our knowledge, there have been no previous reports comparing the effects of aflibercept and faricimab on cytokines in the aqueous humor. 

Methods

Ethics Statement

This prospective observational study was conducted in accordance with the tenets of the Declaration of Helsinki, approved by the Institutional Review Board of Nagoya University Graduate School of Medicine (2021-0474), and registered with the University Hospital Medical Information Network (UMIN000047339). Patients who met the inclusion criteria were informed of the study, and written informed consent was obtained before enrollment. All patient data were anonymized before the analysis. 

Participants

This prospective observational study screened consecutive patients with nAMD treated at the Nagoya University Hospital between June and November 2022, in whom the treatment was switched from aflibercept to faricimab. The decision to switch from aflibercept to faricimab was made by the physician when patients met the criteria of having been treated with aflibercept for more than six months after a TAE regimen and were unable to extend the treatment interval to 13 weeks or more. The treatment interval before and after switching was kept constant, and optical coherence tomography (OCT) and aqueous humor collection were performed at the time of switching (point affected by aflibercept) and after switching (point affected by faricimab). Aqueous humor sampling was performed simultaneously with the vitreous injection. Patients with a history of diabetic retinopathy, cerebral or myocardial infarction, vitrectomy, or vitreous injection administration in the contralateral eye within two months were excluded from the study. These inclusion and exclusion criteria were determined based on the TENAYA and LUCERNE study,¹¹ a global Phase III clinical trial using faricimab, and few other reports on aqueous humor.^20,21 

Clinical Measurements

Data on age, nAMD subtypes, BCVA before and after switching to faricimab, and treatment history with anti-VEGF agents were obtained from medical records. Spectral-domain OCT (Spectralis HRA + OCT; Heidelberg Engineering, Heidelberg, Germany) was performed to evaluate exudative changes in the intraretinal fluid (IRF), subretinal fluid (SRF), and subretinal pigment epithelium (sub-RPE) fluid. The scan protocol for OCT included enhanced depth imaging (EDI) of six radial scans of 100 frames (30° centered at the fovea) and 31 single lines of nine frames (30° × 25° volume scan centered at the fovea), each using a follow-up mode. Central retinal thickness (CRT) and central choroidal thickness (CCT) were measured using the caliper function of the built-in software of the OCT device. CRT was defined as the distance from the inner limiting membrane to Bruch's membrane, and CCT was defined as the distance between Bruch's membrane and choroid-sclera boundary at the fovea on EDI-OCT images. 

Analysis in Aqueous Humor

Aqueous humor samples were collected using a 27G needle and stored at −80°C until analysis. Aqueous concentrations of VEGF-A, placental growth factor (PlGF), and Ang-2 were analyzed using Luminex xMAP technology (Bio-Rad Laboratories Inc., Hercules, CA, USA) with Milliplex MAP Human Angiogenesis/Growth factor Magnetic Bead Panel 1 assay (EMD Millipore, Darmstadt, Germany). Samples (25 µL) were diluted twofold and analyzed according to the manufacturer's protocol. A standard control was performed and the lower limits of detection for VEGF-A, PlGF, and Ang-2 were 13.7, 1.4, and 13.7 pg/mL, respectively. 

Statistical Analyses

All statistical analyses were performed using the SPSS (ver. 28.0.0.0; IBM Corp., Armonk, NY, USA), and the Shapiro–Wilk test was used to test data normality. For comparisons between the two groups, Student's t-test, Mann–Whitney U test, and Wilcoxon signed-rank test were used depending on the correspondence and whether the distribution was normal. Fisher's exact test was used to analyze the contingency tables. Multiple logistic regression analysis was also performed. Data are presented as the mean ± standard deviation. Statistical significance was set at P < 0.05. 

Results

Of the 74 patients with nAMD, 20 eyes were excluded for the following reasons: nine had cerebral or myocardial infarction, nine had received vitreous injection within two months in the contralateral eye, one had diabetic retinopathy, and one was post-vitrectomy. Therefore this study included 54 eyes of 54 patients (44 males and 10 females) with nAMD who switched from aflibercept to faricimab during the study period. The baseline characteristics of the patients are summarized in Table 1, and the mean age was 75.4 ± 9.9 (range 44–93) years. The mean aflibercept treatment interval directly before switching to faricimab was 6.6 ± 2.5 weeks: four- or five-week intervals in 25 eyes (46.3%); six- or seven-week intervals in eight eyes (14.8%); eight-, nine-, or 10-week intervals in 16 eyes (29.6%); and 11- or 12-week intervals in five eyes (9.3%). There were no ocular or systemic complications. 

Table 1.

 

Baseline Characteristics When Switching to Faricimab

 

All the patients underwent BCVA measurements and OCT imaging after the first faricimab injection at the same interval as the aflibercept injection interval before the switch (4–12 weeks). The mean BCVA and CCT did not change significantly after switching. The mean CRT decreased from 306.2 ± 147.5 µm to 278.6 ± 100.4 µm; however, it was not statistically significant (Table 2). 

Table 2.

 

Changes in Clinical Findings After the First Faricimab Injection

 

We compared exudative changes, including SRF/IRF and sub-RPE fluid, on OCT images before and after switching to faricimab. The following results were observed: 48 eyes had fluid, including IRF/SRF or sub-RPE fluid, before switching to faricimab; 11 (22.9 %) showed full absorption, 14 (29.2%) showed reduction, eight (16.7%) showed no changes, and 15 (31.3%) showed worsened exudative changes after switching. Of the remaining six eyes that did not show exudative changes before the switch, three eyes remained without exudative changes, and three eyes developed exudative changes after the switch (Fig. 1). 

Figure 1.

 

Changes in the OCT findings after switching to faricimab. The graph shows the number of eyes in which the fluid shown on the Y-axis is fully absorbed (fully absorbed), reduced (reduced), not changed (no change), or worsened (worsened) compared to OCT findings at baseline. sub-RPE fluid, subretinal pigment epithelium fluid. Fluid presence includes IRF, SRF, and sub-RPE fluid.

Analysis of aqueous humor cytokine levels showed no significant change in VEGF-A, which decreased below the detection sensitivity in most eyes, both before and after switching (Fig. 2A). However, the mean concentration of PlGF increased significantly from 0.86 ± 0.85 pg/mL (interquartile range [IQR] = 0.3–1.1) to 1.72 ± 1.39 pg/mL (IQR = 0.9–2.2) (P < 0.001, Fig. 2B), and the mean concentration of Ang-2 decreased significantly from 23.8 ± 23.5 pg/mL (IQR = 8.7–30.5) to 16.4 ± 21.9 pg/mL (IQR = 3.7–19.9) (P < 0.001, Fig. 2C). There were no significant differences in aqueous humor cytokine levels before switching because of nAMD subtype, lens status, or posterior vitreous detachment. 

Figure 2.

 

Changes in aqueous humor cytokine levels after switching to faricimab. (A) VEGF-A, (B) PlGF, and (C) Ang-2. Data are indicated as mean ± standard deviation unless otherwise indicated. ^*Wilcoxon signed-rank test.

All 54 eyes were classified into two groups according to exudative changes after switching to faricimab: Response group with reduction or complete resolution of exudative changes, and No Response group with no change or worsening of exudative changes. Of the 54 eyes, 28 were in the response group and 26 were in the no response group. Regarding the clinical status, no significant differences were observed in baseline BCVA, treatment intervals before switching to faricimab, nAMD subtype, or CRT at switching between the two groups (Figs. 3B–E). Furthermore, there were no significant differences between the two groups in terms of sex, lens status or posterior vitreous detachment (Figs. 3G–I). The mean age at baseline was significantly higher in the response group (P = 0.003, Fig. 3A), and CCT was significantly thinner in the response group than that in the no response group (P = 0.02, Fig. 3F). However, multiple logistic regression analysis, including these clinical items, showed that only age was significantly associated with treatment outcomes in the response and no response groups (OR = 0.90; 95% CI, 0.82, 0.99; P = 0.041). 

Figure 3.

 

Comparison between the response and no response groups. (A) Mean age at baseline, (B) BCVA at the point of switching to faricimab, (C) treatment intervals before switching, (D) subtype of nAMD, (E, F) CRT and CCT at the switch, (G) sex, (H) lens status, and (I) posterior vitreous detachment. Error bars indicate standard deviation. ^*Student t-test; ^†Mann–Whitney U test; ^‡Fisher's exact test.

When comparing aqueous humor cytokine levels at baseline, there were no significant differences in baseline VEGF-A, PlGF, and Ang-2 between the response and no response groups (P = 0.63, 0.44, and 0.18, respectively; Fig. 4). Furthermore, all 54 eyes were subdivided into two groups based on the interval before switching: (i) 4–6 weeks (n = 26) and (ii) 7–12 weeks (n = 28) and additional analysis was performed. The Ang-2 concentrations before and after switching significantly decreased in both groups, from 23.6 to 15.8 pg/mL in the 4–6 weeks group (P = 0.02, Fig. 5A), and from 23.9 to 16.9 pg/mL in the 7–12 weeks group (P < 0.001, Fig. 5B). In the 4–6 weeks group, the mean concentration of Ang-2 at baseline were 21.9 pg/mL (13 eyes) in the response group and 25.3 pg/mL (13 eyes) in the no response group. This difference was not statistically significant (P = 0.61, Fig. 5C). In the 7–12 weeks group, the mean concentration of Ang-2 at baseline were 31.7 pg/mL (15 eyes) in the response group and 14.9 pg/mL (13 eyes) in the no response group, thus being significantly higher in eyes included in the response group (P = 0.02, Fig. 5D). 

Figure 4.

 

Comparisons of aqueous humor cytokine levels at switching between the response and no response groups. (A) VEGF-A, (B) PlGF, and (C) Ang-2. Error bars indicate standard deviation. ^*Mann–Whitney U test.

Figure 5.

 

Changes in aqueous humor Ang-2 levels after switching to faricimab in the 4–6 and 7–12 weeks groups, and comparison of Ang-2 levels at baseline between the response and no response groups within each group. (A, C) The 4–6 weeks group, and (B, D) the 7–12 weeks group. Error bars indicate standard deviation. ^*Wilcoxon signed-rank test; ^†Mann–Whitney U test.

Discussion

This study evaluated the short-term outcomes of switching from aflibercept to faricimab in patients with nAMD. To the best of our knowledge, this is the first study in clinical practice showing that VEGF was strongly suppressed below the detection sensitivity before and after switching from aflibercept to faricimab in patients with nAMD. Our results suggest that faricimab is effective in reducing exudative changes by suppressing Ang-2 expression after switching. 

VEGF-A has pathological effects such as angiogenesis, increased vascular permeability, and inflammation and is the primary target for nAMD therapy. Aflibercept inhibits PlGF and VEGF-A. However, faricimab suppressed the expression of VEGF-A and Ang-2. In this study, VEGF was suppressed below the detection sensitivity of the aqueous humor cytokine analysis throughout the period before and after switching from aflibercept to faricimab. This indicates that both aflibercept and faricimab effectively inhibit VEGF-A. Further, PlGF significantly increased and Ang-2 significantly decreased after switching, consistent with the expected effects of switching agents. PlGF is a member of the VEGF family and exclusively binds VEGFR1, whereas VEGF-A binds to both VEGFR1 and VEGFR2.²² PlGF induces angiogenesis and inflammation, and controlling it may prove beneficial in treating nAMD.²³ However, despite multiple randomized clinical trials comparing anti-VEGF-A therapy with anti-VEGF-A/anti-PlGF therapy, it remains unclear whether there is a clear difference between the therapies regarding visual improvement or anatomical benefit.^23–25 Regarding Ang-2, preclinical studies suggest that simultaneously targeting the Ang/Tie pathway and VEGF-A result in better outcomes than the anti-VEGF monotherapy alone,^26,27 thereby positioning the Ang/Tie-2 pathway as a promising therapeutic target for nAMD. The results of the current study indicate that faricimab significantly suppresses Ang-2 expression. The Ang-2 concentration values obtained from this analysis do not differ significantly from previously reported values, and the data obtained are considered valid.^17,18 Additionally, 25 of the 48 eyes (52.1%) experienced a reduction in disease activity after switching to faricimab. These results suggest that the use of faricimab may further reduce exudative changes by suppressing Ang-2 and VEGF-A. 

However, the suppression of Ang-2 by faricimab treatment was relatively limited compared with the suppression of VEGF-A which was below the detection sensitivity in most cases. Conventional anti-VEGF agents, such as aflibercept, ranibizumab, and bevacizumab, suppress VEGF-A in the aqueous humor or plasma while increasing Ang-2.^28,29 Since our study involved patients with nAMD who were already receiving ongoing treatment with aflibercept, Ang-2 levels may have been elevated at baseline and could not be fully suppressed by a single dose of faricimab. A recent report in oncology found an upregulation of Ang-2 and vascular Tie-2 after anti-VEGF therapy in therapy-resistant or recurrent tumors.^30,31 In addition, the high expression of Ang-2 in pathologic neovascularization suggests that Ang-2 may be strongly involved in the “escape mechanism” of anti-VEGF therapy, i.e., a state in which the disease activity becomes less dependent on VEGF.^30,32,33 If this is also the case for nAMD, then continuous treatment with faricimab to further suppress Ang-2 may be effective in nAMD resistant to conventional anti-VEGF agents. Some patients in this study had very long treatment periods, many had previous injections, and some of these cases may have been in escape situations. In fact, in the group of eyes where the dosing interval before switching was seven to 12 weeks, Ang-2 levels before switching were higher in eyes with a good response to faricimab. It is possible that these eyes were in an escape situation that reduced their dependence on VEGF and increased their dependence on Ang-2; the anti-Ang-2 effect of faricimab may have improved their exudative changes. However, a similar trend was not seen in the group with a shorter dosing interval of four to six weeks before switching. Therefore there may be a mechanism different from the escape mechanism in aflibercept-refractory cases, which requires further investigation. 

The results of multiple studies have reported the short-term efficacy of faricimab in eyes that have undergone prior treatment.^34–37 Kishi et al.³⁵ evaluated 55 eyes with AMD that received intravitreal aflibercept injections with treatment intervals of less than eight weeks. After switching to faricimab and receiving at least three injections, they reported a significant decrease in CRT. Moreover, IRF and SRF rates significantly decreased (from 20% to 16% and 60% to 40%, respectively). Leung et al.³⁷ reported the clinical outcomes of 190 eyes with AMD that were switched from other anti-VEGF agents to faricimab. Participants received at least three intravitreal injections of faricimab. This study found significant improvements in the BCVA and CRT. The dosing interval for faricimab between the last two injections (7.64 ± 6.2 weeks) was significantly longer than that of ranibizumab (5.16 ± 2.0 weeks) or aflibercept (5.57 ± 3.6 weeks). The results of these studies demonstrate that a certain percentage of patients achieve functional and morphological improvements when switching from current anti-VEGF agents to faricimab in real-world clinical practice, although the study designs and results vary among these reports. In our study, 25 of 48 eyes (52.1%) that showed exudative changes before switching to faricimab achieved complete resolution or reduced exudative changes, resulting in decreased disease activity. These findings are in line with previous reports, and switching to faricimab may prolong the treatment interval in patients with nAMD who are refractory to conventional anti-VEGF agents. 

Identifying the prognostic factors associated with switching to faricimab is critical for determining cases that require such a switch. Further research is necessary to confirm these findings and identify other potential factors. In this study, we compared baseline parameters between the group that responded positively to faricimab treatment and non-responders. Our findings imply that older age and a thin CCT may indicate a good prognosis. However, there were no significant differences in the baseline VEGF, Ang-2, and PlGF levels between the two groups, and they cannot be considered as prognostic factors. In the subgroup analysis, a higher pre-switch Ang-2 level was associated with improved efficacy of switching to faricimab, when limited to the 7–12 weeks group as described above. Therefore, pre-switch Ang-2 value may be a useful prognostic factor for faricimab. Nevertheless, this study only evaluated the results of a single dose of faricimab. Hence, it is desirable to extend the analysis period in the future. 

This prospective study had certain limitations owing to its relatively small sample size and short follow-up period. This may be the reason for the lower improvement in BCVA or CRT in this study compared to previous reports. Moreover, cytokine changes in the anterior aqueous humor may be less pronounced because of the subretinal pigment epithelium being the predominant location for nAMD lesions. VEGF-A levels were below the detection sensitivity in most cases throughout the study. The inability to accurately measure changes in VEGF-A levels may have affected the results of the analysis. In addition, the present sample was sex biased (44 males and 10 females). In Japan, AMD is observed in males three times more than in females,³⁸ and this difference may have influenced the results. Therefore, although this study confirmed that faricimab alters cytokines in the aqueous humor, a more extensive observational period with a larger patient cohort is necessary to clarify the relationship between changes in cytokines in the aqueous humor and clinical outcomes. 

To summarize, we assessed the real-world clinical outcomes of switching to faricimab in patients who received aflibercept for nAMD at intervals of 12 weeks or less. The results highlighted cytokine modifications in the aqueous humor after intravitreal injection of faricimab. These results indicate that VEGF was adequately suppressed after aflibercept injection, and that Ang-2 was also suppressed. Moreover, in more than half of the cases, exudative changes improved after switching to faricimab. However, further longitudinal studies are necessary to evaluate the efficacy of faricimab in the long-term treatment of nAMD. Measuring the cytokine concentrations in the aqueous humor of patients with nAMD can aid in understanding the pathophysiology of each patient and the characteristics of each anti-VEGF agent. This approach can also help identify optimal therapeutic targets and select the most appropriate anti-VEGF agents, potentially serving as a new therapeutic strategy.¹⁹ 

Acknowledgments

Supported by Syneos Health Clinical K.K and Chugai Pharmaceutical Co., Ltd. 

Disclosure: T. Todoroki, Chugai (R); J. Takeuchi, the AMO (F), Novartis (R), Santen (R), Bayer (R), Senju (R), Chugai (R); H. Ota, Novartis (R); Y. Nakano, None; A.F. Sajiki, None; K. Nakamura, None; H. Kaneko, Santen (R), Bayer (R), Chugai (R); K.M. Nishiguchi, Bayer (F, R), Novartis (C, F, R), Santen (F, R), Senju (C, F, R), Chugai (C, R)