Turnover of Microaneurysms After Intravitreal Injections of Faricimab for Diabetic Macular Edema

Abstract

Purpose: Microaneurysm (MA) plays an important role in the pathogenesis of diabetic macular edema (DME) progression and response to anti-vascular endothelial growth factor (VEGF) therapy. This study aimed to investigate the effect of faricimab, a bispecific antibody against angiopoietin-2 and VEGF, on the number of MAs and their turnover in the treatment of DME.

Methods: We included that patients with DME who underwent three monthly injections of faricimab in one eye, with the other eye as control. We examined central retinal thickness (CRT) based on optical coherence tomography (OCT) and best-corrected visual acuity. Turnover, including loss and newly formed MAs, and the total number of MAs were counted based on merged images of the OCT map and fluorescein angiography.

Results: We enrolled 28 patients with DME. After 3 monthly injections of faricimab, CRT significantly improved, 66.0 ± 16.2% of MAs disappeared, and 6.71 ± 5.6% of new MAs were generated, resulting in total reduction to 40.7 ± 15.2%. In the treated eyes, MA disappearance (P < 0.0001) and turnover (P = 0.007) were significantly greater, and new formation was smaller (P < 0.0001) than in non-treated eyes. The size of the retained MAs decreased after treatment. Microaneurysm turnover was not significantly different between areas with and without edema before treatment.

Conclusions: In the process of improving edema in DME with faricimab, MAs shrink and disappear, and formation of MAs are inhibited, resulting in decreased total number of MAs. Intravitreal administration of faricimab suppresses vascular permeability and improves vascular structure.

Diabetic macular edema (DME) is a main cause of central vision loss among people with diabetes mellitus and can occur at any stage of diabetic retinopathy (DR).¹ Long termed hyperglycemia leads to hypoxia and chronic microvascular damage in retina. The permeability of the retinal vessels is dramatically enhanced by angiogenic mediators, such as vascular endothelial growth factor (VEGF) and angiopoietin (Ang)-2.² Microaneurysms (MAs) and impaired vessels with the breakdown of the blood retinal barrier are the primary sources of focal and diffuse leakage, respectively.³ Aberrant leakage results in the accumulation of intraretinal fluid and causes structural and functional abnormalities of the macular area. 

Recently, intravitreal injection of anti-VEGF agents has become the gold standard and first-line treatment for DME.^1,4 Many clinical studies have demonstrated that anti-VEGF therapy is effective in improving macular swelling and visual acuity.^1,5–7 Frequent injections are required to sustain its therapeutic effect. Visual recovery with anti-VEGF therapy contributes to relief anxiety in patients and improves glycemic control.⁸ 

Several studies have shown that MAs are resistant to anti-VEGF therapy. Hirano et al. showed that a higher injection number of anti-VEGF agents is required to obtain stable improvement of edema when MAs are present beside the foveal avascular zone (FAZ).⁹ Lee et al. reported that small vascular flow density in the deep capillary plexuses, a large FAZ area, and a large number of adjacent MAs are characteristic of cases with poor response to anti-VEGF therapy.¹⁰ To overcome the weakness of anti-VEGF therapy, new pharmacological agents with mechanisms that inhibit new production of MAs and promote their disappearance in DME are required. 

Currently, four types of VEGF inhibitors are approved for treatment of DME in Japan, including ranibizumab, aflibercept, brolucizumab, and faricimab.¹¹ Ranibizumab and brolucizumab inhibit VEGF-A, whereas aflibercept inhibits VEGF-A, VEGF-B, and placental growth factor.¹² Faricimab is a bispecific antibody targeting VEGF-A and Ang-2.¹³ The concentration of Ang-2 in the vitreous body of patients with DR is higher than that of controls (macular hole).¹⁴ The YOSEMITE and RHINE trials demonstrated that anatomic and visual improvements with faricimab were achieved and had a potential to extend the dosing interval up to 16 weeks for patients with DME.¹³ In normal vessels composed of pericyte and endothelial cells (ECs), vascular stability and homeostasis are maintained by Ang-1 and tyrosine-protein kinase receptor (Tie2) signaling.¹⁵ However, in pathological vessels, Ang-2 inhibits Ang1/Tie2 signaling and destabilizes vessels, weakening junctions of ECs and promoting vascular leakage.¹⁵ Pericyte loss further destabilizes vessels and induces MA formation and neovascularization. Therefore, faricimab may affect MA dynamics during DME improvement. In this study, we analyzed changes in the total number and turnover of MA, including their production and disappearance, before and after three monthly injections of faricimab for DME. 

Patients and Methods

This prospective study was carried out according to the tenets of the Declaration of Helsinki. The study was approved by the Institutional Review Board of the University of Fukui, and was registered with the University Hospital Medical Information Network Clinical Trials Registry of Japan (UMIN000049521, registered on November 16, 2022). We obtained written informed consent from all patients after fully explaining the aims and procedure of the study. We included patients with DME requiring anti-VEGF treatment in only one eye, and the non-treated eye served as the control. 

Inclusion and Exclusion Criteria

The inclusion criteria were as follows: (1) age > 20 years; (2) a diagnosis of type 2 diabetes mellitus (DM) with DME; and (3) use of faricimab as anti-VEGF therapy for diffuse DME with central retinal thickness (CRT) more than 300 µm in one eye. The following exclusion criteria were: (1) history of injection of anti-VEGF drugs and steroids or retinal photocoagulation within 6 months before the initial injection of faricimab; (2) active intraocular inflammation or infection; (3) uncontrolled glaucoma in either eye; (4) other retinal diseases, such as retinal vein occlusion or retinal detachment; (5) history of stroke; (6) systolic blood pressure (BP) > 180 mm Hg, diastolic BP > 100 mm Hg or untreated hypertension, and (7) excessively severe medial opacity that precluded fundal evaluation (for example, severe cataract, corneal opacity, or vitreous hemorrhage). If DME developed or progressed in the untreated eye, anti-VEGF treatment was initiated, and the eye was excluded from the control group. In this study, we enrolled patients with diffuse and center-involved DME, but not those with focal DME, treated using faricimab. Diffuse DME was defined as (1) increased retinal thickness with center involvement on an optical coherence tomography (OCT) map and (2) fluorescein leakage starting from the early phase and continuously increasing to the late phase. Focal DME was defined as the location outside the foveal center and pin-point leakage in the early phase. 

Procedures

All patients underwent comprehensive ophthalmic examinations during their initial visits, including best-corrected visual acuity (BCVA) testing, slit-lamp microscopy, intraocular pressure measurements, and fundoscopy. The BCVA was converted to the logarithm of the minimum angle of resolution (logMAR) scale. We captured the color fundus photographs using a Kowa VX-10i fundus camera (Kowa Ltd., Nagoya, Japan), and the images of fluorescein angiography (FA) and the OCT map including the CRT measurement using a Spectralis Heidelberg Retinal Angiograph (Heidelberg Engineering, Heidelberg, Germany). Using OCT, we obtained automatically reconstructed false-color topographic images displaying the average thicknesses in each of the nine map sectors, as defined by the Early Treatment Diabetic Retinopathy Study criteria, to measure retinal thickness. The scanned areas measured 6 × 6 mm², with the fovea in the center, and were detected as hyperfluorescent dots during the early phase of FA imaging (within 1 minute after dye injection). All imaging tests were performed by experienced orthoptists who were blinded to the treatment status. We displayed highly thickened areas (>500 µm at the center of the edema) as white areas (WAs) and less thickened areas (approximately 400–500 µm area surrounding the center of the edema) as red areas (RAs) on the color OCT maps. We defined the WAs and RAs as areas of edema. We performed the FA within 1 week before the first injection of faricimab. 

Intravitreal injections were administered in a standardized manner by a trained ophthalmologist (author Y.Y.) using 0.4% oxybuprocaine hydrochloride (0.4% benoxyl ophthalmic solution; Santen Co. Ltd., Osaka, Japan) and 2% xylocaine as anesthetic and povidone iodine as antiseptic. An eyelid speculum was used to stabilize the eyelids. The injection volume of faricimab (VABYSMO; Cyugai pharmaceutical Co., Ltd. Tokyo, Japan) was 2 mg/0.05 mL. Three sessions of injections were administrated, once each month. We performed examination before the injection on the day of the first dose (day 0), 4 weeks, and 8 weeks for the second and third injections, respectively, and 4 weeks later at week 12. 

Merged Imaging Technique to Analyze MA Turnover

We merged FA images before and after treatment to analyze MA turnover using Adobe Photoshop Elements (Adobe Systems Inc., San Jose, CA, USA), as previously described.¹⁶ Initially, we marked the MAs in the FA images taken before (day 0) and after (12 weeks) with red and green colors, respectively. We reduced the transparency of one image by 50%, and overlapped the images with reference to the retinal vessels. The numbers of red, green, and yellow MAs were counted as disappeared MAs, newly developed MAs, and maintained MAs before and after injection, respectively, within a 6 × 6 mm² area. The MA turnover was defined as the sum of newly developed and disappeared MAs. The FA image was merged with the OCT map captured at the same time. The numbers of MAs lost, new MAs generated, and MAs maintained were expressed as percentages, with the number of MAs before injection being 100. We calculated reduction rate of MAs by dividing the difference in MAs before and after treatment by the number of MAs before treatment, expressed as a percentage. MAs were divided into three groups according to their diameters: small (less than 3 pixels), medium (3–7 pixels), and large (> 7 pixels).¹⁷ The unit of MA diameter was converted from pixels to µm based on the 200 µm scale displayed on the FA image. 

Statistical Analysis

A sample size of 22 subjects, accounting for an approximately 10% dropout rate, would have provided 80% power to prove at a 1-sided α-level of 0.025, were calculated by G*power. Statistical analyses were performed using the JMP software (SAS Institute Inc., Tokyo, Japan). Variables were expressed as means ± standard deviation (SD). Bartlett's test was used to examine equal variances across the samples. After confirming normal distribution of the data, the number of MAs was compared between time points using the Wilcoxon signed-rank test. Differences between the treated and untreated eyes were analyzed using the Mann–Whitney U test. Statistical significance was set at P < 0.05. 

Results

We enrolled 28 patients in this prospective study between November, 20, 2022, and August 20, 2023. A patient withdrew from the study due to discontinued visit. In one patient's untreated eye, DME occurred during the observation period, and received faricimab injection; therefore, this eye was excluded (Fig. 1). The Table shows the baseline characteristics of the patients enrolled at registration. None of the patients experienced adverse events after the injection, including retinal detachment, endophthalmitis, or vitreous hemorrhage. 

Figure 1.

 

Diagram showing recruitment of participants.

Table.

 

Baseline Characteristics at the Time of Registration

 

CRT significantly decreased at 8 and 12 weeks (P < 0.0001) after the first injection (Fig. 2A) and BCVA improved 12 weeks after the initial injection (P = 0.036; Fig. 2B). The untreated eyes showed no significant changes in BCVA and CRT. 

Figure 2.

 

Changes in the central retinal thickness (A) and best corrected visual acuity (BCVA) (B) eyes treated or untreated with faricimab injections. The BCVA is expressed as logMAR. Data are presented as mean ± standard deviation. *P < 0.05 (versus baseline).

We analyzed the changes in MAs and their turnover after three monthly injections of faricimab using merged images of the OCT map and FA. A representative example is shown in Figure 3. The total number of MAs decreased from 309 (see the red dots in Fig. 3A) to 119 (see the green dots in Fig. 3B) after treatment. The numbers of disappeared MAs (see the red dots in Fig. 3C) and newly developed MAs (see the green dots in Fig. 3C) were 206 and 16, respectively. The number of maintained MA (see the yellow dots in Fig. 3C) was 103. These analyses were performed in parallel for both treated and untreated eyes in all cases. As shown in Figure 4A, the number of MAs significantly decreased to 40.7 ± 15.2% in the treated eyes and showed an increasing trend to 121 ± 19.3% in the non-treated eyes. The reduction rate of MAs in the treated eyes was 59.3 ± 14.8%. In the treated eyes, 66.0 ± 16.2% of MAs disappeared and only 6.7 ± 5.6% of new MAs occurred. The number of MAs that remained unchanged in location before and after treatment was 85.8 ± 13.2% in the non-treated eyes compared to 35.3 ± 14.1% in the treated eyes. Significantly, more MAs disappeared, and few were newly generated and maintained in the treated eyes than in the non-treated eyes (P < 0.0001). Turnover, the sum of MAs disappeared and developed, was significantly greater in the treated eyes (72.8 ± 17.2%) than in non-treated eyes (50.5 ± 19.8%; P = 0.007). 

Figure 3.

 

A sample case showing microaneurysm (MA) turnover after faricimab treatment. Fluorescein angiography (FA) image was merged with OCT map in the eye treated with faricimab (A, B, C). MAs before (A) and after (B) faricimab treatment were marked with red and green, respectively, and the figures were merged (C). Images taken at same time in another untreated eye is shown in (E, F, G). The color bar indicates the number of MAs before (red), after (green), new developed (blue), disappeared (gray), and maintained (yellow) in the eyes treated (D) and untreated (H) with faricimab.

Figure 4.

 

The rate of MA number in total disappeared, newly formed, maintained, and turnover in the treated and untreated eyes with faricimab (A). Parallel measurement was carried out in the areas with and without edema before treatment (B). *P < 0.05, **P < 0.0001.

We examined whether there was a difference in the MA dynamics between areas with and without edema before treatment (Fig. 4B). There was no significant difference in the number of MAs that disappeared, were newly formed, or were maintained between areas with edema (as shown in red and white on the OCT map) and areas without edema. 

Next, we analyzed the changes in the size of the MAs that continuously existed during the observation period (Fig. 5). The size of the MAs decreased from 4.9 ± 0.5 pixels (53.9 ± 5.5 µm) to 4.3 ± 0.4 pixels (47.3 ± 4.4 µm), the difference was significant (P = 0.0005; see Fig. 5B). We divided the MAs into three groups according to size and analyzed the changes in size before and after treatment (see Fig. 5C). In the maintained MAs, 51.0 ± 27.7% and 8.3 ± 13.2% of MAs of large size became smaller to medium size and small size, respectively, and 37.1 ± 11.8% of MAs that were medium size became small size. As a result, the percentage of MAs of large size decreased from 18.0 ± 12.3% to 7.4 ± 4.2%, medium size MAs decreased from 58.0 ± 26.3% to 50.1 ± 21.5%, and small size MAs increased from 24.1 ± 14.8% to 42.5 ± 12.7%. On the other hand, MAs that disappeared after treatment were 4.9 ± 2.2% large size, 53.7 ± 21.8% medium size, and 41.2 ± 17.3% small size before treatment. 

Figure 5.

 

The change of the size of maintained MAs after treatment with faricimab. (A) In the FA images taken before (a) and after treatment (b), MAs were marked depending on the size of MAs (blue = large, yellow = medium, and red = small). (B) The change of MA size before and after faricimab injections. Data are presented as mean ± standard deviation (SD). *P < 0.05 (versus baseline). (C) Color bars indicate the distribution of large (blue), medium (yellow), and small (red) size MAs before and after faricimab treatment. Data represents mean (SD).

Discussion

In the present study, intravitreal injection of faricimab, a bispecific antibody targeting Ang-2 and VEGF, significantly eliminated previously existing MAs and reduced the formation of new MAs compared to untreated control eyes. The merged method showed that more MAs disappeared than newly produced MAs; consequently, the absolute number decreased. The reduction in MA formation and promotion of MA disappearance with faricimab treatment would play an important role in the pathogenesis of the DME recovery process by faricimab. Conversely, in untreated eyes, the rate of disappearance of MAs was lower, whereas newly formed MAs were greater than those in treated eyes. The conflicting dynamics of MA in treated and untreated eyes suggest that the temporal change in MA number in the treated eye was due to the pharmacological effect of faricimab and not the natural course of DME. Increased disappearance and decreased formation contributed to a reduction in the total number of MAs after treatment, and faricimab may effectively reduce the number of MAs via both mechanisms. 

Similar to our results, MA counts decrease after other anti-VEGF treatments, including ranibizumab and aflibercept.^18,19 Babiuch et al. reported that the reduction rate was 55.3% after 6 months with injections of aflibercept every 4 weeks.¹⁸ Mori et al. reported that the number of MAs in the interquartile range of retina decreased after anti-VEGF therapy from 6 in the early phase and 3 in the late phase of FA to 2 and 1, respectively.¹⁹ Sugimoto et al. reported that the reduction rate of MAs after 3 injections of aflibercept was 50.4 ± 21.2%.²⁰ On the other hand, that of our study using faricimab was 59.3 ± 18.3%. Leicht et al. reported that the total number of MAs decreased from 5.6 ± 0.7 to 4.0 ± 0.7 (approximately 28.6%) and MA turnover was higher in ranibizumab treated eyes than in control eyes.²¹ However, unlike our faricimab data, ranibizumab treatment resulted in more abundant number of newly produced MAs than the control. Faricimab, which suppresses Ang-2 and VEGF-A, might suppress MA production compared to ranibizumab or aflibercept, which suppresses VEGF-A only. 

Our data showed a size reduction in the remaining MAs after treatment. The ratio of large MAs decreased, whereas that of small MAs increased after treatment with faricimab. Furthermore, the merging method directly demonstrated that 59.3% of the large MAs and 37.1% of the medium-sized MAs became smaller. Thus, the reduction in MA size was probably involved in the disappearance of MAs following faricimab treatment. In fact, 93.3% of the MAs that disappeared were medium– or small-sized before treatment. An et al. reported that leaking MAs were smaller than MAs without leakage.¹⁷ MAs with leakage had the highest percentage at 70% in MAs less than 30 µm in diameter, and conversely, the lowest percentage at 30% in MAs larger than 70 µm. In our study, the sizes less than 30 µm and larger than 70 µm in diameter correspond to small and large sizes, respectively.¹⁷ The disappearance of many small MAs with leakage may contribute to the improvement of edema. 

MAs are characterized by pericyte loss and aberrant proliferation of ECs.³ Some cytokines, including VEGF, Ang-1, Ang-2, transforming growth factor-β, and platelet-derived growth factor-B, are associated with the pathogenesis of MA formation.^22,23 The eyes of monkeys injected with VEGF showed the formation of MAs, indicating the VEGF mediated pathology.²⁴ In the mouse retina, loss of pericytes causes perivascular macrophage infiltration, and macrophage derived VEGF activates VEGF-receptor-2 in ECs.²⁵ Therefore, it is not surprising that the number of MAs decreases after anti-VEGF treatment. In addition, several experimental studies have shown that Ang-2 is involved in MA formation. Epithelial cells without pericyte in MAs also cause the elevation of Ang-2.²⁵ In pericytes undergoing apoptosis induced by hyperglycemia, Ang-1 promotes cell survival, whereas Ang-2 promotes apoptosis.²⁶ Moreover, Park et al. reported that Ang-2 induces apoptosis of pericytes via the p53 pathway under hyperglycemia.²⁷ These findings suggest that blocking Ang-2/Tie-2 signaling using faricimab could be a potential therapeutic target to prevent pericyte loss. Ang-2 and VEGF interact to enhance vascular permeability. The analysis using porcine retinal ECs showed that VEGF and Ang-2 acting together increased permeability three times as much as VEGF alone.² Moreover, Benest et al. reported that Ang-2 deficient mice showed a remarkable reduction in VEGF-induced vascular leakage.¹⁵ Thus, treatment with faricimab designed to inhibit both VEGF and Ang-2/Tie-2 systems may accelerate their effects and reduce vascular permeability in DME. 

MA is a risk factor for refractory cases of anti-VEGF therapy for DME. Focal edema frequently remains after anti-VEGF treatment.²⁸ These areas have a low reduction rate of the retinal thickness,²⁹ and highly dense MAs are consistent with the residual edema area that is refractory to anti-VEGF therapy.²⁸ Direct photocoagulation of MAs in residual edema is an effective treatment tool.¹⁶ However, laser therapy must be minimally invasive to avoid damage of the photoreceptor cells and pigment epithelium. Alternatively, additional repeated injections of anti-VEGF agents for residual edema resulted in resolution of the center-involving DME in approximately 90% of cases.²⁸ If faricimab is particularly potent in reducing the number of MAs, it may decrease the frequency of residual edema after injections. 

Based on our data, there was no significant difference in the MA reduction rate between highly edematous areas and other areas at baseline. This finding suggests that the effect of reducing MAs by faricimab acts equally on the retina and is not restricted to edematous areas. When the rate of MA reduction is constant, areas with higher MA density before injection may have more MA remaining after injection than other areas. Nevertheless, multiple injections of faricimab, which strongly reduces the number of MAs, may decrease the risk of the formation of residual edema derived from MAs. Indeed, in our case series, residual edema was rarely observed after three monthly injections. 

Although OCT-angiography (OCTA) can visualize the MAs and non-perfusion areas without intravenous dye injection, it is reported that not all MAs identified with FA can be seen on OCTA.^30,31 Therefore, we analyzed the dynamics of MAs based on the FA images. Nevertheless, because OCTA has an advantage to show en face images of different retinal layers, further analysis using OCTA may reveal whether MAs in superficial or deep layers responded more to faricimab treatment. 

Gao et al. classified MAs into partially or fully perfused MAs and non-perfused MAs according to the differences in feature of colocalized MAs among the FA images, en face images of structural OCT and OCTA and evaluated flow status in MAs.³² They also demonstrated that OCT-identified MAs colocalized perfused flow on OCTA are associated with retinal fluid accumulation.³² In our study, it is not yet clear whether the disappearance of MAs after faricimab injection is due to closure of MA or reduced blood flow. This issue may be clarified by a comprehensive combined analysis of en face images of multiple modalities such as OCTA and structural OCT as well as FA. 

This study had several limitations. First, this study was based on a small number of cases because only monocular center-involved DME cases were included. In addition, the observational period was short; however, this was because the number of injections and their intervals differed among cases over a long period. Second, because we did not directly compare eyes treated with faricimab to those treated with other anti-VEGF agents, we were unable to analyze the differences in the effects of the drugs on MA kinetics. Further large-scale longitudinal studies are required to address these issues. 

Acknowledgments

Supported in part by grants-in aid for scientific research (J160000936) from the Japan Society for the Promotion of Science, Tokyo, Japan. 

Disclosure: Y. Takamura, None; Y. Yamada, None; M. Morioka, None; M. Gozawa, None; T. Matsumura, None; M. Inatani, None