Changes in Retinal Microcirculation After Intravitreal Ranibizumab Injection in Eyes With Macular Edema Secondary to Branch Retinal Vein Occlusion

Abstract

Purpose: To evaluate the effects of an intravitreal ranibizumab (IVR) injection on the retinal microcirculation of eyes with macular edema secondary to a branch retinal vein occlusion (BRVO).

Methods: Twenty-six eyes of 26 patients with macular edema due to a BRVO that had received a single IVR injection (0.5 mg/0.05 mL) were studied. The retinal microcirculation was assessed by laser speckle flowgraphy (LSFG) using the mean blur rate (MBR) and relative flow volume (RFV). The size of the retinal arteries and veins surrounding the optic nerve head were measured separately. All of the examinations were made before, and at 1 week, and 1 and 2 months after the IVR.

Results: The visual acuity improved significantly, and the mean central macular thickness decreased significantly during the follow-up period (both P < 0.001). The mean MBR_all and MBR_tissue decreased significantly at 1 week and 1 month after the IVR (both P < 0.001). The total RFV of the arteries and veins decreased significantly at 1 week and 1 month after the IVR injection in the occluded and nonoccluded quadrants (all P < 0.001). The width of the arteries and veins in the LSFG images decreased significantly at 1 week and 1 month after the IVR injection (P < 0.001).

Conclusions: An IVR injection leads to a transient vasoconstriction of the retinal arteries and veins and a reduction of the retinal blood flow and velocity in both the occluded and nonoccluded quadrants. The changes in retinal microcirculation might be related to the improvement of the macular edema and vision.

A branch retinal vein occlusion (BRVO) is a relatively common retinal vascular disorder in elderly patients that can lead to visual impairments.¹ Macular edema (ME) is the most frequent cause of the visual impairments in eyes with a BRVO.¹ There are several treatments for the ME secondary to a BRVO (e.g., laser photocoagulation,^2,3 intravitreal injection of steroids,⁴ subtenon injection of triamcinolone acetonide [STTA],⁵ and vitrectomy⁶). It has been reported that an intravitreal injection of an anti-vascular endothelial growth factor (VEGF) agent, such as bevacizumab,⁷ ranibizumab,⁸ or aflibercept,⁹ can significantly improve the best-corrected visual acuity (BCVA) and lead to greater benefits than the other treatments for patients with ME secondary to a BRVO.^7–9 

Ranibizumab is a 48 kDa antigen-binding fragment (Fab) of the bevacizumab molecule and was specifically developed for ocular use.¹⁰ Some studies have reported that an intravitreal bevacizumab (IVB) injection had some effect on the retinal blood flow and velocity.^11–13 On the other hand, it has been reported that an intravitreal ranibizumab (IVR) injection induces vasoconstriction, but there is no study that examined the retinal blood flow or velocity after an IVR injection. Thus, the retinal hemodynamic changes after IVR have not been determined for different retinal diseases including eyes with a BRVO. 

Different techniques have been used to measure the retinal blood flow including radioactive microspheres,¹⁴ hydrogen clearance,¹⁵ laser Doppler technique,^16,17 color Doppler ultrasonography,¹⁸ and the pulsatile technique.¹⁹ The time intensiveness and poor reproducibility have hampered the widespread use of these techniques. More recently, a Doppler optical coherence tomographic (OCT) technique,²⁰ OCT angiography,²¹ and optical microangiography (OMAG)^22,23 have been used to measure the retinal blood flow. However, these techniques still have limited clinical use due to the time intensiveness of the procedures. 

Laser speckle flowgraphy (LSFG) is a noninvasive, real-time method that has been used to measure the relative blood flow on the optic nerve head (ONH). The measurement duration is only 4 seconds and can be performed without the intravenous injection of any contrast agents.^24–26 Laser speckle flowgraphy detects the speckle contrast pattern produced by the interference of illuminating laser light that is scattered by the movement of erythrocytes in the blood vessels. The changes in the contrast pattern enables the device to calculate the relative blood flow in the vessels on the ONH and retina, which is expressed as the mean blur rate (MBR).^24–26 The relative flow volume (RFV) obtained by LSFG was recently reported to be an accurate and reliable index of the blood flow volume in the retinal vessels.^27,28 Laser speckle flowgraphy can measure the RFV of all of the major retinal arteries and veins surrounding the ONH. 

In the acute phase of BRVO, the occluded vein becomes narrower leading to difficulty in measuring the blood flow of the obstructed vein. The purpose of this study was to measure the RFV of the retinal arteries and veins separately by LSFG. The measurements were made for the occluded and the nonoccluded regions around the ONH. The changes in the retinal hemodynamics were investigated after IVR in BRVO patients with ME. 

Methods

Ethics Statement

This was a retrospective cross-sectional, single center study, and the procedures used were approved by the Ethics Committee of the Nagoya University Hospital (Nagoya, Japan). The procedures conformed to the tenets of the Declaration of Helsinki, and all patients signed a written informed consent. 

Subjects and Testing Protocol

Twenty-six eyes of 26 patients with a BRVO that received 1.25 mg/0.05 mL of IVR at the Nagoya University Hospital between July 2013 and July 2015 were studied. The criteria for receiving IVR were ME involving the fovea with a central macular thickness (CMT) >300 μm and a BCVA <20/30. All of the eyes were followed for at least 2 months after the IVR injection. The eyes were divided into two subgroups: whether ME recurred or not at 2 months after the IVR injection. We examined the differences in the values of the parameters determined by LSFG between the two groups. The unaffected fellow eyes were studied in the same way as controls. 

Exclusion Criteria

Patients who had had previous treatments such as an intravitreal injection of anti-VEGF agents and steroids, STTA, photocoagulation, and vitrectomy were excluded. In addition, subjects who had glaucoma, vitreous hemorrhage, corneal opacity, severe cataract, and history of ophthalmic or general disorders were excluded. Patients with diabetes were excluded because the retinal blood flow might be impaired even in patients with no retinopathy.²⁹ Patients with uncontrolled systemic hypertension (blood pressure [BP], 160/100 mm Hg) were also excluded. 

Examinations

The patients underwent comprehensive ophthalmologic examinations including measurements of the BCVA and intraocular pressure (IOP), slit-lamp biomicroscopy, color fundus photography, indirect ophthalmoscopy, optical coherence tomography (Spectralis, Heidelberg Engineering, Heidelberg, Germany), and LSFG before, and at 1 week, and 1, and 2 months after the IVR injection. Additionally, the BP and heart rate (HR) were measured with an automatic sphygmomanometer (CH-483C; Citizen, Tokyo, Japan). All examinations at each visit were performed in the sitting position from 10:00 to 12:00 hours to avoid diurnal variations.³⁰ All patients were asked to abstain from alcoholic and caffeinated beverages on the morning of the examination because the intake of alcohol and caffeine can influence the IOP³¹ and BP.³² Fluorescein angiography (FA) was performed at the initial examination (Optomap camera, Optos plc., Dunfermline, Scotland). A nonperfused area was measured using ImageJ software (ImageJ version 1.47; http://imagej.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA). and an area of more than five disc areas was classified as ischemic BRVO. 

Determination of RFV in Retinal Vessels With LSFG.

The laser speckle flowgraphy-NAVI was used to determine the retinal microcirculation. The principles of LSFG have been described in detail.^33–36 Briefly, this instrument consists of a fundus camera equipped with an 830-nm diode laser as the light source and a standard charge-coupled device sensor (750 width × 360 height pixels) as the detector. After switching on the laser, a speckle pattern appears due to the interference of the light scattered by the movements of the erythrocytes. The mean blur rate is a measure of the relative blood flow velocity, and it is determined by examining the pattern of the speckle contrast produced by the movements of the erythrocytes in the blood vessels. To evaluate the circulation on the ONH, a circular marker was set surrounding the ONH (Fig. 1A). The “vessel extraction” function of the software then identified the vessel and tissue areas on the ONH so that the MBR could be assessed separately as the total MBR (MBR_all) and tissue areas (MBR_tissue; Fig. 1B). The software in the instrument is able to track and compensate for eye movements during the measurement period. The laser speckle flowgraphy was measured three times at each time point in all eyes. The average of the variables derived from the LSFG device was calculated. 

Figure 1

 

Representative composite color maps of the MBRs of the ONH as measured by LSFG (A). The red color indicates a high MBR and the blue color indicates a low MBR. To measure the MBR on the ONH blood flow, a circle was set around the ONH (A). A binary format image for segmentation between the vessel (white area) and tissue (black area) areas on the ONH (B). In evaluating the RFV, we divided the ONH into four quadrants: a (white arrow), b, c, and d (black arrow) (C). In the representative case, the occluded vessel is included in area a, which is named the occluded area. The other areas, b, c, and d, are named the nonoccluded areas. Then, we calculated a sum of RFV on the vessels around the ONH in the occluded area and nonoccluded areas (D). A representative major BRVO (E, F). A representative macular BRVO (G, H).

The calculation of the RFV has been described in detail.²⁸ In evaluating the RFV, we divide the ONH into four quadrants “a,” “b,” “c,” and “d,” because it is often not easy to identify the occluded vein (Fig. 1C). In the representative case, the occluded vessel is included in area a, which is named the occluded area. The other areas, b, c, and d, are named the nonoccluded areas. Then, we calculate a sum of the RFV of the vessels around the ONH in the occluded area and nonoccluded areas (Fig. 1D). 

Intravitreal Injection of Ranibizumab and Measurement of VEGF Level in Aqueous by ELISA.

The eyes were anesthetized with topical 1% tetracaine, and the fornices were irrigated with 10% povidone-iodine. A mean volume of 0.1 mL of aqueous humor was extracted by anterior chamber paracentesis with a 27-gauge needle attached to a 1-mL syringe.³⁷ All aqueous humor samples were collected before the IVR injection, and they were stored at −80°C until use. The concentration of VEGF in the extracted aqueous humor was measured by enzyme-linked immunosorbent assay (ELISA) with a commercially available kit (Quantikine; R&D Systems, Minneapolis, MN, USA), which measures both human VEGF121 and VEGF165.³⁸ 

Statistical Analyses

All statistical analyses were performed with the IBM SPSS Statistics for Windows, Version 24.0 (IBM Corp., Armonk, NY, USA). Mixed linear model analyses were used to analyze the differences of the BCVA, CMT, MBR, RFV, and other systemic and ocular parameters. The Bonferroni's correction was applied to adjust for the multiple comparisons. Multiple regression analyses were used to determine the significance of the differences in the variables associated with the concentration of VEGF before the injection. The data are presented as means ± standard error of the means. A P value <0.05 was taken to be significant. 

Results

None of the patients was retreated during the follow-up period of 2 months after the initial IVR injection. The characteristics of the subjects at the baseline are shown in Table 1. There were 14 men and 12 women ages 66.0 ± 2.4 years. Fluorescein angiography revealed ischemic occlusion in 19 cases and nonischemic occlusion in seven cases. The mean BCVA was 0.59 ± 0.06 logMAR units, the CMT was 576.8 ± 40.1 μm, and the CT was 237.2 ± 15.6 μm before the IVR injection. Fourteen patients had hypertension, and none had diabetes. 

Table 1

 

Characteristics of Subjects

 

No endophthalmitis, retinal detachment, or any other severe procedure-related complications developed. The intraocular pressure did not exceed 21 mm Hg during follow-up in none of the eyes. No obvious ranibizumab-related ocular or systemic adverse events developed. None of the patients developed any neovascular complications or required peripheral scatter laser photocoagulation during the follow-up period. 

The mean BCVA improved significantly, and the mean CMT decreased significantly during the follow-up period after the IVR (both P < 0.001; Fig. 2). There was no significant change in the IOP, systolic BP, diastolic BP, and mean ocular perfusion pressure (MOPP) during the follow-up period (Table 2). 

Figure 2

 

Changes in the BCVA and CMT after an IVR injection. The mean BCVA improved, and the mean CMT decreased significantly after the IVR (all P < 0.001).

Table 2

 

Changes in Systemic and Ocular Parameters

 

The changes in the MBR_all (B) and MBR_tissue (C) with increasing follow-up times are shown in Figure 3. The mean MBR_all and MBR_tissue decreased significantly at 1 week and 1 month after the IVR injection (both P < 0.001; Fig. 3). 

Figure 3

 

Composite color map and changes in the MBR on the ONH in an eye with a BRVO treated with an IVR injection. The composite color map before IVR, 1 week, 1 month, and 2 months after the IVR (A). The mean MBR of the all area of the ONH is decreased significantly in eyes with a BRVO at 1 week and 1 month after the IVR (both P < 0.001; B). The mean MBR of the tissue area of the ONH was decreased significantly in eyes with a BRVO at 1 week and 1 month after the IVR (both P < 0.001; C).

The changes in the RFV with increasing follow-up times are shown in Figure 4. The total RFV surrounding the ONH decreased significantly at 1 week and 1 month after the IVR, but no significant changes were observed in the fellow eyes (Fig. 4A). The total RFV in the arteries and veins decreased significantly at 1 week and 1 month after the IVR injection compared to that before the injection (Fig. 4B). In both the occluded and nonoccluded quadrants, the RFV decreased significantly at 1 week and 1 month after the IVR injection (Fig. 4C). The width of both the arteries and veins recognized by LSFG decreased significantly at 1 week and 1 month after the IVR injection (Fig. 4D). 

Figure 4

 

Changes in the RFV and width of the retinal arteries and veins. The total of RFV surrounding the ONH is significantly decreased at 1 week and 1 month after IVR, whereas no change is observed in the fellow eye (A). The total of RFV for both the arteries and veins decreases significantly at 1 week and 1 month after the IVR injection compared to that before the IVR (B). In both the occluded and nonoccluded quadrants, the RFV decreased significantly at 1 week and 1 month after the IVR injection (C). The width of the arteries and veins recognized by LSFG significantly decreased at 1 week and 1 month after the IVR injection (D). ***P < 0.001, **P < 0.01.

Of the 26 eyes, 16 eyes (62%) had a recurrence of ME at 2 months after the IVR injection. There were no significant differences in any characteristics between the recurrent group and the nonrecurrent group (Table 3). In both the occluded and nonoccluded quadrants, there was no statistically significant difference in the RFV between the groups at any time points, and no difference in the trend after the IVR injection between the two groups (Figs. 5A, 5B). The ratio of the postinjection to that preinjection width of the arteries and veins in the nonrecurrent group was significantly lower than that in the recurrent group at all time points (Figs. 5C, 5D). 

Table 3

 

Characteristics of Subjects

 

Figure 5

 

The changes in the RFV and the ratio of width of the vessel at the base line in the retinal arteries and veins. In both the occluded and nonoccluded quadrants, there was no statistically significant difference in the RFV between the groups at all time points and no change in the trend after the IVR injection between the groups (A, B). The ratio of width of postinjection to that preinjection in the nonrecurrent group is significantly lower than that in the recurrent group in all time points in vein (C) and artery (D). **P < 0.01, *P < 0.05.

The mean aqueous concentration of VEGF before the injection was 355.1 ± 122.3 pg/mL. The nonperfused area was 1.05 ± 0.21 × 10⁶ pixels, which was positively correlated with the aqueous concentration of VEGF (r = 0.566, P = 0.029). Multiple regression analysis of the concentration of VEGF before injection and other parameters are shown in Table 4. The concentration of the VEGF before the IVR injection was correlated with the venous RFV in the occluded quadrant (β = −794, P = 0.042). 

Table 4

 

Result of Multiple Regression Analysis for Independence of Factors Contributing to VEGF Concentration

 

Discussion

The results showed that there was a transient and significant reduction of the RFV in the arteries and veins passing across the ONH after a single IVR injection in both the occluded or nonoccluded quadrants. In addition, the width of the retinal vessels determined by LSFG was transiently decreased, and the MBR_all and MBR_tissue were transiently reduced after the IVR injection. These results suggest that the IVR injection induced a temporary decrease of the retinal blood flow and velocity in addition to vasoconstriction of the arteries and veins. 

Sacu et al.³⁹ reported a vasoconstriction of the arteries and veins of the injected eyes measured by a retinal vessel analyzer in eyes with a BRVO after three IVR injections. It has also been reported that IVR induced a vasoconstriction of the retinal arterioles in patients with AMD.^40,41 These results indicate that an IVR injection can cause vasoconstriction of retinal vessels. However, these studies did not measure the retinal blood flow and velocity, so they could not evaluate the effect of ranibizumab on the retinal microcirculation comprehensively in these patients.^39–41 

Our results showed that the retinal blood flow and velocity were decreased. To the best of our knowledge, this is the first study that showed that IVR induces a transient reduction of retinal blood flow and velocity. On the other hand, Nitta et al.⁴² reported that there was no significant change in the MBR on the ONH after an IVB injection. In addition, Nagaoka et al.¹² did not observe any significant change in the vessel diameter, blood flow velocity, and blood flow in either the occluded or nonoccluded vein in the affected eye in response to single IVB injection after 3 months. Noma et al.¹³ reported that the venous RFV ratio increased significantly in the nonoccluded region of an eye with a BRVO at 1 month after IVB. These discrepancies in the effect of anti-VEGF agents were not determined, but one possibility is that the effect of these agents on the retinal microvessels might vary depending on the anti-VEGF agent because of the differences in their structure and molecular size.¹⁰ Another possibility is that the time of the measurements or the measuring instruments was different among the studies. 

Vascular endothelial growth factor can cause an increase of the retinal blood flow⁴³ probably by the production of nitric oxide.⁴⁴ Nitric oxide plays an important role in the vasodilation of the retinal vessels⁴⁵ and the increase in the retinal blood flow and velocity. Persistent ME might be associated with the dilation of the smaller vessels and/or capillaries in response to nitric oxide production induced by the increased VEGF and/or tissue hypoxia. Therefore, in contrast to VEGF, it is more likely that anti-VEGF agents can induce vasoconstriction of vessels and a decrease of retinal circulation. 

The ratio of the width of both the arteries and veins postinjection to that preinjection in the nonrecurrent group was significantly lower than that in the recurrent group. Im et al.⁴⁶ reported that the eyes that did not develop an ME recurrence after a single IVB injection showed a persistent (at least 6 months) decrease in the unaffected venular diameter in all quadrants. In contrast, eyes with ME recurrences were associated with a marked increase in the venular diameters in the unaffected quadrants.⁴⁶ These results corroborate our findings, and it would suggest that eyes with nonrecurrence of ME after an IVR injection cause more vasoconstriction because of lower VEGF concentration. 

Our results showed that the mean BCVA improved and the mean CMT decreased significantly at 2 months after the IVR. The improvement of the BCVA and the decrease of CMT peaked at 1 month and the effect became slightly less at 2 months after the IVR. Experimentally, the half-life of IVR in the vitreous cavity of the monkey eyes is 2.6 days⁴⁷ and 2.88 days in the rabbit eyes.⁴⁸ A concentration of >0.1 μg/mL ranibizumab is maintained in the vitreous humor of humans for 29 days.⁴⁸ These results support the time course of improvement of vision and decrease of CMT after IVR in our study. Taken together, IVR has a transient effect on the retinal microcirculation causing a decrease of the CMT and improvement of the BCVA. 

The concentration of the VEGF before IVR was significantly and negatively correlated with the preoperative venous RFV in the occluded area. There are very few studies that determined the correlation between the retinal blood flow and the VEGF concentration in eyes with a RVO. In eyes with a central RVO (CRVO), Yamada et al.⁴⁹ found that the MBR determined by LSFG was negatively correlated with the aqueous VEGF concentration before treatment, which supports our results. The aqueous and vitreous levels of VEGF are correlated with the size of the nonperfused area of eyes with BRVO,^50–53 and the VEGF concentrations in the aqueous humor and vitreous were correlated with the severity of ME.^50,52–54 Taken together, our results suggest that larger nonperfused areas in eyes with BRVO can lead to higher concentrations of VEGF and lower venous RFV values in the occluded quadrant. This would then result in a negative correlation between the VEGF concentration and the venous RFV. However, these mechanisms are not fully understood. Thus, further experiments are needed to determine the correlation between the retinal blood flow and the VEGF concentrations in the aqueous. 

In the fellow eyes, our results showed there were no significant alterations of any parameter determined by LSFG after the IVR injection. On the other hand, one recent clinical trial reported that an injection of IVB might affect the retrobulbar ocular blood flow in the injected eye and the untreated fellow eyes of patients with neovascular AMD.⁵⁵ Avery et al.⁵⁶ reported on the systemic pharmacokinetics following IVR and IVB; ranibizumab cleared very quickly, whereas bevacizumab and aflibercept had a longer systemic half-life and induced a marked reduction in plasma-free VEGF. The characteristics of the systemic pharmacokinetics following IVR should be related to the results of the fellow eyes. 

This study has several limitations. First, the sample size was not large and the follow-up period was short. Because some eyes can have a recurrence of the ME, they required several IVR injections during a long follow-up period. The use of data from these eyes would be confounding. Accordingly, we evaluated the retinal microcirculation for only 2 months after a single IVR injection. 

Second, the mean time after the onset of BRVO was 44 days, which suggests that many patients had relatively fresh BRVO (less than 2 months) in the present study. It is well known that some patients with BRVO have a spontaneous resolution of the ME without any treatment, and it was reported that the retinal blood flow and the velocity are different between fresh BRVO eyes with spontaneous ME resolution and persistent ME.¹² Accordingly, a possibility is not denied that ME was resolved spontaneously in some eyes of the nonrecurrent group, and it could have affected the results, especially in the few months after the onset of the BRVO.⁵⁷ 

Third, 14 patients with controlled systemic hypertension were included, and the effect of systemic medications on the retinal microcirculation in patients with BRVO needs to be studied in the future. Fourth, no control group in which the natural course of retinal vessel diameters after BRVO was observed. Finally, there are several studies reporting that the choroidal blood flow is reduced after anti-VEGF agent injection. We did not measure the choroidal blood flow, but it would be better to measure retinal and choroidal blood flow simultaneously after the injection of the anti-VEGF agent. 

In conclusion, this is the first study that evaluated the retinal blood flow and velocity after a single IVR injection in patients with ME secondary to BRVO. The injection led to a temporal vasoconstriction of the retinal arteries and veins and a reduction of the retinal blood flow and the velocity in both the occluded or nonoccluded quadrants. The changes in the retinal microcirculation may be related to the improvements in the ME and vision. 

Acknowledgments

Supported by a Grant-in-Aid for Scientific Research (C) (TI) and a Grant-in-Aid for Scientific Research (B) (HT). 

Disclosure: M. Fukami, None; T. Iwase, None; K. Yamamoto, None; H. Kaneko, None; S. Yasuda, None; H. Terasaki, None