Effect of Intravitreal Injection of High-Dose Bevacizumab in Monkey Eyes

Abstract

purpose. To evaluate the ocular complications of intravitreal injection of high dose bevacizumab in monkey eyes.

methods. Four healthy monkeys (Macaca fuscata) received an intravitreal injection of either 6.25 or 12.5 mg bevacizumab in the right eye; the left control eye received an intravitreal injection of the same volume of saline. The eyes were examined using slit lamp and funduscopy, optical coherence tomography (OCT), electroretinography (ERG), fundus photography (FP), fluorescein angiography (FA), and indocyanine green angiography (ICGA). The eyes were enucleated 28 days after the intravitreal injection and subjected to light microscopy.

results. No pathologic changes were observed by FP, FA, ICGA, OCT, and light microscopy in the eyes injected with either of the two bevacizumab doses. ERG showed no toxic change in the eyes that received the 6.25-mg dose. In two eyes that received the 12.5-mg dose, ERG showed no significant difference between the right and left eyes 4 weeks after injection, although there were transient changes in scotopic responses.

conclusions. No irreversible toxic effects were observed in monkey eyes receiving an intravitreal injection of high dose bevacizumab.

Vascular endothelial growth factor (VEGF) plays a major role in normal and pathologic angiogenesis, promoting the formation of abnormal vessels and increasing vascular permeability as seen, for example, in age-related macular degeneration (AMD), proliferative diabetic retinopathy (PDR), retinal vein occlusion, and other ocular diseases. ^{1 2 3 4} In diabetic retinopathy and retinal vein occlusion, the intraocular levels of VEGF are associated with proliferative retinal changes. ^{3 5} In animal models, the blockage of VEGF results in substantial inhibition of vascular permeability and reduced retinal neovascularization. ^{6 7}  

Bevacizumab (Avastin; Genentech, Inc., San Francisco, CA) is a full-length humanized murine monoclonal IgG antibody that binds all isoforms of VEGF. It was approved by the Food and Drug Administration as an antiangiogenesis drug that can be administered intravenously in combination with 5-fluorouracil to treat metastatic colorectal cancer and currently is being studied as a treatment for various solid tumors. ^{8 9 10 11} Bevacizumab is also in phase III trials for advanced breast cancer and advanced renal cancer. ^{10 11} Systemic bevacizumab therapy is effective for neovascular AMD and has resulted in a significant improvement in visual acuity and decreased central retinal thickness. ¹² However, the drug is associated with hypertension, thromboembolic events, gastrointestinal perforation, proteinuria, and other complications in patients with cancer. ^{8 13}  

Bevacizumab is administered intravitreally to reduce the adverse systemic effects and intravitreal route of administration results in a high concentration of the drug in the vitreous. Bevacizumab is far less expensive than other VEGF inhibitors, such as pegaptanib (Macugen; Eye Tech/Pfizer, New York, NY) and ranibizumab (rhuFab V2; Lucentis, Novartis, Basel, Switzerland), which have been been effective in clinical trials. ¹⁴ Recently, numerous studies of intravitreal bevacizumab for VEGF-associated diseases, such as PDR, ^{15 16 17 18} retinal vein occlusion, ^{19 20} AMD, ^{21 22 23} retinopathy of prematurity, ²⁴ and other diseases, ^{25 26 27 28 29} have been reported. 

Intravitreal injection of bevacizumab is performed under local anesthesia, and repeated intravitreal injections of bevacizumab (1.25 mg) are often required at 4- to 6-week intervals. To the best of our knowledge, no study has been reported on the use of intravitreal injections of high-dose bevacizumab over 5.0 mg in primates. In the present study, we proposed to inject high-dose bevacizumab intravitreally into monkey eyes and evaluate the retinal and choroidal vasculature and permeability by examining cross-sectional retinal images, retinal function, and histology. From these data, we hope to promote the clinical use of high-dose pharmacologic intervention with bevacizumab to inhibit progression of severe VEGF mediated diseases such as severe ROP. 

Methods

Animals

Four healthy juvenile (Macaca fuscata) monkeys were used in the study, which was conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The weights and ages of the monkeys were as follows: monkey 1: 3.3 kg, 42-month-old male; monkey 2: 5.9 kg, 9-year-old female; monkey 3: 3.2 kg, 44-month-old male; and monkey 4: 9.7 kg, 9-year-old male. Before all examinations, the monkeys were anesthetized by intramuscular injection of a mixture of ketamine hydrochloride (10 mg/kg) and xylazine hydrochloride (2 mg/kg). Topical ocular surface anesthesia (0.4% oxybuprocaine hydrochloride) was instilled to reduce discomfort. The pupils were dilated with 0.5% tropicamide and 0.5% phenylephrine. We measured the axial length of the eyes by ultrasonography (UD-6000; Tomey, Nagoya, Japan), with the animals under general anesthesia. The axial length of the right eye injected with bevacizumab (0.25 mL or 6.25 mg) and the left eye injected with saline (0.25 mL) were 18.9 and 18.3 mm, respectively, for monkey 1 and 20.3 and 20.1 mm, respectively, for monkey 2. The axial length of the right eye injected with bevacizumab (0.50 mL or 12.5 mg) and the left eye injected with saline (0.50 mL) were 18.6 and 18.5 mm, respectively, in monkey 3 and 21.6 and 21.5 mm, respectively, in monkey 4. 

The monkeys underwent clinical examination by slit lamp and indirect ophthalmoscopy, electroretinogram (ERG), fundus photography (FP), fluorescein angiography (FA), indocyanine green angiography (ICGA), and optical coherence tomography (OCT). Slit lamp and indirect ophthalmoscopy, ERG, and OCT were performed at baseline, 3 days, 1 week, and 4 weeks after injection. FP, FA, and ICGA were performed at baseline and 4 weeks after injection. After the animals were killed by intravenous injection of an overdose of pentobarbital under deep anesthesia 4 weeks after the injection, the eyes were enucleated and submitted for light microscopy. 

Intravitreal Injection of Bevacizumab

Bevacizumab solution (25 mg/mL, original vial concentration) was drawn into a tuberculin syringe with a 30-gauge needle. The eyes were sterilized with 5% povidone iodine. To prevent an acute increase in intraocular pressure, we injected 0.25 mL of bevacizumab (6.25 mg) 1.5 mm posterior to the limbus after withdrawing 0.15 mL of aqueous fluid with a 27-gauge needle. The same procedure was performed on the first day and the following day with regard to the monkey that received the higher dose of bevacizumab (0.50 mL, 12.5 mg). We referred to the day of the first injection as the first day. The left eye of each monkey served as a control. The left eye was treated in the same manner as the right eye, with the exception that it was injected intravitreally with saline. Levofloxacin ophthalmic solution 0.5% and ofloxacin ophthalmic ointment 0.3% were applied to the eyes immediately after the injections. 

IOP Measurement

Measurements of intraocular pressure were performed by Goldmann applanation tonometry for the experimental eyes of monkeys 2 and 4 before and just after the intravitreal injection, which was performed every 15 minutes. 

Slit Lamp and Funduscopic Examinations

At each time point, we examined the anterior segment and anterior vitreous of the eye by slit lamp to identify inflammation, infection, or cataract formation. The fundus was examined by indirect ophthalmoscopy. 

Fundus Photography and Angiography

Fundus photography (FP), fluorescein angiography (FA), and indocyanine green angiography (ICGA) of the posterior pole were taken with a digital fundus camera (TRC-50IX; Topcon, Tokyo, Japan) connected to an image acquisition system (Image net 2000 High Resolution System; Topcon). FA and ICGA were separately performed by intravenously injecting 1.0 mL of 10% fluorescein sodium and 10 mg/mL of ICG. The FA and ICGA images were captured at approximately every 1 and 5 minutes after an injection, respectively. 

OCT Imaging

At baseline, 3 days, 1 week, and 4 weeks after injection of bevacizumab, OCT macular images (monkeys 1 and 3, Stratus OCT Model 3000; monkeys 2 and 4, Cirrus HD-OCT, Carl Zeiss Meditec, Dublin, CA) were obtained to determine the presence of morphologic changes. We performed five measurements of foveal thickness at each time point and performed statistical analysis by t-test (baseline versus each point). P < 0.05 was considered significant. 

Electroretinography

At baseline, 3 days, 1 week, and 4 weeks after bevacizumab injection, full-field ERG responses were recorded simultaneously from the experimental and control eyes using a synchronized trigger and summing amplifier (Primus; Mayo, Nagoya, Japan) with a stimulation device (LS-W; Mayo). The ERG responses were recorded by using white light-emitting diode built-in corneal contact-typed bipolar electrodes, which were placed on each cornea with 1.5% hydroxymethylcellulose solution. Needle electrodes were placed in the SC skin of the forehead and earlobe as the negative and ground electrodes, respectively. The monkeys were dark-adapted for at least 60 minutes after pupillary dilation, and scotopic responses were examined. Measurement of the ERG responses was subject to the guidelines of the International Society for Clinical Electrophysiology of Vision. The colors of the scotopic-adapted rod stimulus, the photopic cone stimulus, and the flash stimulus were all white and the color temperature was 4000 to 9000 Kelvin. The scotopic-adapted bright white-flash stimulus was set at 8000 cd/m² for 0.5 ms and the scotopic-adapted rod stimulus was set at 316.2 cd/m² for 0.03 ms. All the monkeys were light adapted (prestimulus light background was 25 cd/m²) for 10 minutes before photopic ERG recordings. The cone stimulus was set at 1000 cd/m² for 3 ms. 

Electroretinographic analysis was based on amplitudes and implicit times of dark-adapted scotopic rod responses and dark-adapted bright-flash ERG responses (rod and cone) and of light-adapted photopic cone responses as independent cone ERG signals. The ERG was obtained at least three times. For each monkey, the experimental and control eye were compared at each time point, to minimize the effect of individual conditions, and statistical analysis about respective data sets was performed by t-test. P < 0.05 was considered significant. 

Histologic Analysis

Four weeks after injection, the eyes were enucleated and frozen in optimal cutting temperature embedding medium (Miles Diagnostics, Elkhart, IN). Sections of 10-μm thickness were stained with hematoxylin and eosin and examined by light microscopy. Measurements of total retinal thickness of the eyes (monkeys 1 and 3) at the point 1.5 mm from the optic disc were performed on five histologic sections. For each monkey, the total retinal thickness of the experimental and control eye were compared and statistical analysis about respective data sets was performed by t-test. P < 0.05 was considered significant. 

Results

Slit Lamp and Funduscopy Examinations

There were no obvious changes in slit lamp or fundus examinations at each time point in eyes intravitreally injected with bevacizumab or saline (Figs. 1 2 3 4 A 4B 4C 4Din each). The cornea, conjunctiva, lens, and vitreous appeared clear and normal. 

Intraocular Pressure Measurement

The levels of IOP in the experimental eyes of monkeys 2 and 4 were 8 and 11 mm Hg, respectively, just before injection; 49 and 46 mm Hg, respectively, just after injection; 14 and 12 mm Hg, respectively, 15 minutes after injection; and 10 and 10 mm Hg, respectively, 30 minutes after injection. 

FP, FA, and ICGA

There were no changes on FPs (Figs. 1 2 3 4, A 4B 4C 4Din each) in the eyes before injection and at 4 weeks after injection. FA and ICGA (Figs. 5 6 7 8)showed no obvious changes such as leakage of fluorescein and ICG dye at 4 weeks compared with baseline. 

OCT Imaging

There were no changes such as retinal swelling or serous retinal detachment at any time point on the OCT images of eyes intravitreally injected with 6.25 or 12.5 mg bevacizumab compared to control eyes (Figs. 1 2 3 4, E 4F 4G 4Hin each). The photoreceptor inner and outer segment junction (IS/OS) was preserved during the experiment. There were no significant differences by t-test in the foveal retinal thicknesses between baseline and each time point. Tables 1 and 2show average retinal thickness (±SE) and probabilities. 

Electroretinogram

Statistical studies of the group that received an intravitreal injection of 6.25 mg show average amplitudes, implicit times (±SE), and probabilities (Tables 3 4) . There were no significant differences by t-test in the responses between the experimental and control eyes at baseline, 3 days, 1 week, and 4 weeks after intravitreal injection of 6.25 mg bevacizumab. 

Statistical studies of the group that received an intravitreal injection of 12.5 mg show average amplitudes, implicit times (±SE) and probabilities (Tables 5 6) . Figures 9 10 and 11show the representative traces (monkey 3) of dark-adapted scotopic rod responses, dark-adapted bright flash ERG responses (rod and cone), and light-adapted photopic cone responses at baseline, 3 days, 1 week, and 4 weeks after intravitreal injection of 12.5 mg bevacizumab. There was a significant decrease in amplitudes of both b-wave of dark-adapted scotopic rod responses and a-wave of dark-adapted bright flash ERG responses (rod and cone) at 3 days and 1 week. Furthermore, there was a significant prolongation in implicit time for both dark-adapted scotopic rod responses and a-wave of dark-adapted bright flash ERG responses (rod and cone) at 1 week after injection. 

In monkey 4 (Table 6) , there was a significant decrease in both amplitude and prolongation in the implicit times in b-wave of dark-adapted scotopic rod ERG responses at 1 week after injection. There were no significant differences by t-test between the experimental and control eyes the a-wave of dark-adapted bright-flash ERG responses (rod and cone). However, no significant differences between the experimental and control eye were found at 4 weeks. No significant differences in light-adapted photopic cone responses were found after intravitreal injection of either 6.25 or 12.5 mg bevacizumab. 

Histologic Analysis

Light microscopic examinations of the experimental (bevacizumab 6.25 or 12.5 mg) and control eyes did not reveal any pathologic changes such as infiltration of inflammatory cells, or thinning, swelling, or vacuolization of the retina or choroid (Fig. 12) . The ganglion cells, the inner and outer nuclear layers, and the photoreceptor structures appeared normal. There were no significant differences in the total retinal thicknesses between the experimental and control eyes in monkeys 1 and 3 (Figs. 12G 12H) . 

Discussion

The current results suggest that an intravitreal injection of bevacizumab (6.25 mg) does not cause functional or morphologic retinal damage in monkey eyes, although the 12.5-mg dose of bevacizumab transiently reduced scotopic responses and prolonged their implicit times during 4 weeks of follow-up. The mean axial length of the monkey eyes in monkeys 1 and 3 was 18.6 mm; thus, the vitreous volume was calculated as half that of human eyes. As a result, the intravitreal concentrations of bevacizumab in the monkey eyes injected with the 6.25- and 12.5-mg doses were equivalent to 10 and 20 times the current clinical dose of 1.25 mg bevacizumab, respectively. Our results indicate that the toxicity of intravitreal bevacizumab appears to be limited even in the case of a dose 10 times higher than the 1.25-mg dose used clinically, and this is the first study to investigative toxicity for primates using OCT, FA, and ICGA after intravitreal bevacizumab injection that exceeds 5 mg. 

IOP increased immediately after intravitreal injection and then returned to normal within 15 minutes. Falkenstein et al. ³⁰ reported that in 122 intravitreal injections for treatment of neovascular AMD, the mean IOPs 3 minutes after intravitreal injection of bevacizumab (0.05 mL, 1.25 mg) increased over 35 mm Hg, but they decreased under 25 mm Hg 10 minutes after the injections, and all eyes were below 30 mm Hg at 15 minutes after injection. Hollands et al. ³¹ also reported that in 104 human eyes injected with bevacizumab intravitreally (0.05 mL, 1.25 mg), the mean IOP before injection and 2, 5, and 30 minutes after injection was 14.0, 36.1, 25.7, and 15.5 mm Hg, respectively. In our study, the behavior of IOP after injection of bevacizumab was very similar to that in these previous reports. To prevent an acute increase in intraocular pressure, we performed paracentesis before intravitreal injection of high volume bevacizumab solution (the concentration of bevacizumab solution we can now procure is 25 mg/mL). If we can use the higher concentration of bevacizumab in the future, the volume of bevacizumab solution will be decreased and it will be unnecessary to perform paracentesis. 

Shahar et al. ³² reported that intravitreal bevacizumab (2.5 mg) had no toxic effect on the ERG and visual evoked potential in rabbit eyes. Manzano et al. ³³ also reported that intravitreal bevacizumab doses of 0.5, 1.0, 2.5, and 5.0 mg had no apparent toxic effects on the ERG and light microscopy in rabbit eyes. Bakri et al. ³⁴ reported that intravitreal bevacizumab at doses of 1.25 and 2.5 mg showed no signs of retinal or optic nerve toxicity by light microscopy in rabbit eyes. Inan et al. ³⁵ reported that intravitreal bevacizumab at doses of 1.25 and 3.0 mg in rabbit eyes showed no signs of retinal toxicity by either ERG or light microscopy, but there was mitochondrial damage in the inner segments of the photoreceptors by electron microscopy and more intensive apoptotic protein expression in the outer retina than in the control eyes. We observed morphologic retinal changes at baseline, 3 days, 1 week, and 4 weeks using OCT, which showed an intact IS/OS line and retinal pigment epithelium (RPE) throughout the follow-up period. Similar to previous reports, no histologic changes were observed by light microscopy at 4 weeks. 

The physiological functions of VEGF, with regard to development of the eyeball and maintenance of the choriocapillaris and retinal vasculature, are important in a normal eye. ^{36 37} Lack of RPE-derived VEGF induces severe microphthalmia, absence of choroidal vasculature, and fenestrated vessels in mice. ³⁶ Interruption of VEGF-A causes decreased blood vessel branching and density in the retina. ³⁷ The intravitreal half-life of 1.25 mg of intravitreal bevacizumab is 4.32 days in rabbit eyes, and minute amounts of bevacizumab have been detected in the serum and in the fellow eye that did not receive an injection. ³⁸  

Heiduschka et al. ³⁹ showed that the bevacizumab molecule can penetrate the retina and is also transported into the RPE, the choroid, and, in particular, into the photoreceptor outer segments, after the drug is injected intravitreally. These studies raised the possibility that intravitreal bevacizumab may alter the retinal and choroidal circulation. We performed FA and ICGA at baseline and 4 weeks after injection, and FA showed no leakage of retinal vessels or obstruction of the capillary network, and the ICGA findings remained normal after injection. 

In previous reports, the retinal function has been assessed by flash ERG in rabbits after intravitreal injections of bevacizumab at various concentrations. We evaluated the retinal function using full-field ERG in primates. Both rod and cone responses were normal at 4 weeks, although significant transient reduction and prolongation of both scotopic b- and a-wave after flash stimulus were detected in the eyes that received the higher dose of bevacizumab (12.5 mg or 20 times higher in monkey 3 than that used clinically). 

Previous reports indicate the use of intravitreal injection of bevacizumab for severe ROP (stage 4). Sonmez et al. ⁴⁰ showed large increases in VEGF and SDF-1α concentrations in the vitreous samples obtained by vitrectomy for stage 4 ROP. This finding suggests that anti-VEGF treatment may be of benefit in some eyes that develop ROP. Lee et al. ⁴¹ reported an inhibitory effect of bevacizumab injected intraperitoneally on the angiogenesis and growth of retinoblastoma in mice model. This report suggests that bevacizumab can be used in the treatment of retinoblastomas, similar to its use in metastatic colorectal cancer. Recurrence of retinoblastoma can lead to visual loss, enucleation, and death. In addition, repeated intravitreal injections could increase the risk of metastasis, dissemination of carcinoma and complications from the general anesthesia required, especially in infants. If intravitreal bevacizumab can be shown effective in the treatment of ROP and retinoblastoma, a single high dose rather than multiple small doses is desirable. Since our study supports the use of high-dose bevacizumab, it provides hope to patients with either ROP or retinoblastoma. 

In conclusion, no irreversible toxic effects were observed in monkey eyes after intravitreal injection of high-dose bevacizumab. Our findings raise the possibility that high concentrations of bevacizumab may be useful to treat severe VEGF-induced retinopathy. A larger scale study using the preclinical primate and a longer follow-up period are needed to support our data. In addition, human clinical trials are necessary because of the small number of eyes used for each dose in our study. 

 Supported by a grant from the Health and Welfare Ministry, Japan.

 Submitted for publication December 3, 2007; revised May 19 and October 9, 2008, and February 18 and April 7, 2009; accepted August 18, 2009.

 Disclosure: K. Sakurai, None; H. Akiyama, None; Y. Shimoda, None; I. Yoshida, None; M. Kurabayashi, None; S. Kishi, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

FPs of eyes (monkey 1) at baseline (A, B) and 4 weeks (C, D) after injection. The right eye was injected intravitreally with 6.25 mg (0.25 mL) bevacizumab (C) and the left eye with the same volume of saline (D). No obvious changes such as retinal or vitreous hemorrhages, vitreous opacity, or periphlebitis developed. There were no morphologic changes in the OCT images (Stratus OCT Model 3000; Carl Zeiss Meditec) of the eye injected with 6.25 mg bevacizumab at baseline (E), 3 days (F), 1 week (G), and 4 weeks (H).

Figure 2.

 

FPs of monkey eyes (monkey 2) at baseline (A, B) and 4 weeks (C, D) after injection. The right eye was injected intravitreally with 6.25 mg (0.25 mL) bevacizumab (C) and the left eye with the same volume of saline (D). No obvious changes such as retinal or vitreous hemorrhages, vitreous opacity, or periphlebitis developed. There were no morphologic changes in the OCT (Cirrus HD-OCT; Carl Zeiss Meditec) images of the eye injected with 6.25 mg bevacizumab at baseline (E), 3 days (F), 1 week (G), and 4 weeks (H).

Figure 3.

 

FPs of monkey eyes (monkey 3) at baseline (A, B) and 4 weeks (C, D) after injection. The right eye was intravitreally injected with 12.5 mg (0.50 mL) bevacizumab (C) and the left eye with the same volume of saline (D). No obvious changes such as retinal or vitreous hemorrhage, vitreous opacity, or periphlebitis developed. There were no morphologic changes in the OCT (Stratus OCT Model 3000; Carl Zeiss Meditec) images of the eye injected with 12.5 mg bevacizumab at baseline (E), 3 days (F), 1 week (G), and 4 weeks (H) after injection.

Figure 4.

 

FPs of monkey eyes (monkey 4) at baseline (A, B) and 4 weeks (C, D) after injection. The right eye was intravitreally injected with 12.5 mg (0.50 mL) bevacizumab (C) and the left eye with the same volume of saline (D). No obvious changes such as retinal or vitreous hemorrhage, vitreous opacity, or periphlebitis developed. There were no morphologic changes in the OCT (Cirrus HD-OCT; Carl Zeiss Meditec) images of the eye injected with 12.5 mg bevacizumab at baseline (E), 3 days (F), 1 week (G), and 4 weeks (H) after injection.

Figure 5.

 

In monkey 1, FA showed no pathologic leakage or hyper- or hypofluorescein before injection (A, B) and 4 weeks after injection (C, D). The retinal vasculature was normal in the right eye injected with 6.25 mg (0.25 mL) bevacizumab (C) and in the left eye injected with saline (D). ICGA showed no pathologic leakage or hyper- or hypofluorescein before injection (E, F) or 4 weeks after injection (G, H). There were no marked changes in the choroidal vasculature in the right eye injected with 6.25 mg bevacizumab (G) or the control eye (H). The timing of the angiogram frames was as follows. (A) 3 minutes 15 seconds, (B) 3 minutes 52 seconds, (C) 3 minutes 33 seconds, (D) 3 minutes 59 seconds, (E) 4 minutes 28 seconds, (F) 5 minutes 6 seconds, (G) 4 minutes 48 seconds, and (H) 5 minutes 39 seconds.

Figure 6.

 

In monkey 2, FA showed no pathologic leakage or hyper- or hypofluorescein before injection (A, B) and 4 weeks after injection (C, D). The retinal vasculature was normal in the right eye injected with 6.25 mg bevacizumab (C) and in the left eye injected with saline (D). ICGA showed no pathologic leakage or hyper- or hypofluorescein before injection (E, F) or 4 weeks after injection (G, H). There were no marked changes in the choroidal vasculature in the right eye injected with 6.25 mg bevacizumab (G) or the control eye (H). The timing of the angiogram frames was as follows: (A) 3 minutes 18 seconds, (B) 3 minutes 43 seconds, (C) 2 minutes 54 seconds, (D) 3 minutes 43 seconds, (E) 5 minutes 40 seconds, (F) 4 minutes 29 seconds, (G) 4 minutes 33 seconds, and (H) 5 minutes 11 seconds.

Figure 7.

 

In monkey 3, FA showed no pathologic leakage or hyper- or hypofluorescein before injection (A, B) and 4 weeks after injection (C, D). The retinal vasculature was normal in the right eye injected with 12.5 mg bevacizumab (C) and in the left eye injected with saline (D). ICGA showed no pathologic leakage or hyper- or hypofluorescein before injection (E, F) and 4 weeks after injection (G, H). There were no marked changes in the choroidal vasculature in the right eye injected with 12.5 mg bevacizumab (G) and the control eye (H). The timing of angiogram frames was as follows: (A) 2 minutes 3 seconds, (B) 3 minutes 35 seconds, (C) 3 minutes 13 seconds, (D) 3 minutes 58 seconds, (E) 4 minutes 10 seconds, (F) 3 minutes 13 seconds, (G) 5 minutes 17 seconds, (H) 4 minutes 42 seconds.

Figure 8.

 

In monkey 4, FA showed no pathologic leakage or hyper- or hypofluorescein before injection (A, B) and 4 weeks after injection (C, D). The retinal vasculature was normal in the right eye injected with 12.5 mg bevacizumab (C) and in the left eye injected with saline (D). ICGA shows no pathologic leakage or hyper- or hypofluorescein before injection (E, F) and 4 weeks after injection (G, H). There were no marked changes in the choroidal vasculature in the right eye injected with 12.5 mg bevacizumab (G) and the control eye (H). The timing of the angiogram frames was as follows: (A) 2 minutes 47 seconds, (B) 3 minutes 29 seconds, (C) 2 minutes 46 seconds, (D) 3 minutes 9 seconds, (E) 4 minutes 17 seconds, (F) 5 minutes 56 seconds, (G) 5 minutes 3 seconds, and (H) 5 minutes 14 seconds.

Table 1.

 

Foveal Retinal Thickness in Monkeys 1 and 2, Treated with 6.25 mg Bevacizumab or 0.25 mL Saline

	Bevacizumab (Right Eye)	P (vs. baseline)	Control (Left Eye)	P (vs. baseline)
Monkey 1
Retinal thickness (μm)
Baseline	141.2 ± 7.7		139.2 ± 4.2
3 days	138.2 ± 6.2	0.771	139.0 ± 5.1	0.977
1 week	137.8 ± 6.4	0.743	138.0 ± 6.2	0.877
4 weeks	140.0 ± 6.2	0.907	140.0 ± 3.6	0.889
Monkey 2
Retinal thickness (μm)
Baseline	180.8 ± 2.4		183.6 ± 2.5
3 days	181.4 ± 2.2	0.848	178.4 ± 3.7	0.287
1 week	179.0 ± 2.4	0.619	178.9 ± 2.8	0.256
4 weeks	181.4 ± 2.8	0.863	179.4 ± 2.9	0.311

 Thickness (micrometers ± SE) was measured by OCT and probability determined by t-test (baseline versus each point).

Table 2.

 

Foveal Retinal Thickness in Monkeys 3 and 4, Treated with 12.5 mg Bevacizumab or 0.5 mL Saline

	Bevacizumab (Right Eye)	P (vs. baseline)	Control (Left Eye)	P (vs. baseline)
Monkey 3
Retinal thickness (μm)
Baseline	144.6 ± 6.6		141.2 ± 6.6
3 days	145.0 ± 5.8	0.911	143.8 ± 6.6	0.788
1 week	138.8 ± 4.6	0.528	146.0 ± 6.2	0.611
4 weeks	140.2 ± 5.3	0.661	141.6 ± 2.4	0.956
Monkey 4
Retinal thickness (μm)
Baseline	177.8 ± 2.2		180.2 ± 2.0
3 days	179.9 ± 2.0	0.507	181.0 ± 2.2	0.788
1 week	178.0 ± 1.6	0.942	179.4 ± 2.4	0.808
4 weeks	176.7 ± 2.1	0.723	177.6 ± 3.1	0.495

 Thickness (micrometers ± SE) was measured by OCT and probability determined by t-test (baseline versus each point).

Table 3.

 

Amplitudes and Implicit Times in Monkey 1, with 6.25 mg Bevacizumab or 0.25 mL Saline

	Bevacizumab (Right Eye)	Control (Left Eye)	P
Rod Response b-Wave
Amplitude (μV)
Baseline	256.4 ± 6.5	270.5 ± 7.7	0.232
3 days	231.7 ± 4.5	231.2 ± 3.6	0.937
1 week	329.2 ± 6.0	347.7 ± 12.7	0.260
4 weeks	242.0 ± 5.8	237.8 ± 4.7	0.599
Implicit time (msec)
Baseline	85.5 ± 2.2	83.3 ± 1.7	0.476
3 days	78.8 ± 1.6	84.2 ± 3.1	0.198
1 week	78.9 ± 2.2	89.3 ± 3.9	0.080
4 weeks	76.4 ± 0.4	79.8 ± 2.4	0.242
Flash ERG b-Wave
Amplitude (μV)
Baseline	436.0 ± 22.4	421.1 ± 19.4	0.640
3 days	364.3 ± 7.6	344.2 ± 11.3	0.218
1 week	431.6 ± 11.1	428.2 ± 9.4	0.829
4 weeks	432.9 ± 6.9	414.0 ± 1.4	0.055
Implicit time (msec)
Baseline	71.0 ± 1.0	68.5 ± 0.3	0.070
3 days	71 ± 7.8	69.5 ± 11.3	0.164
1 week	52.5 ± 0.2	63.6 ± 0.7	0.172
4 weeks	53.4 ± 0.3	53.8 ± 1.0	0.708
Flash ERG a-Wave
Amplitude (μV)
Baseline	−180.1 ± 11.1	−182.2 ± 6.1	0.873
3 days	−169.4 ± 4.2	−175.8 ± 1.0	0.211
1 week	−227.3 ± 4.1	−223.5 ± 17.2	0.841
4 weeks	−191.5 ± 17.3	−179.6 ± 4.8	0.546
Implicit time (msec)
Baseline	12.3 ± 0.1	12.0 ± 0.2	0.251
3 days	10.9 ± 0.0	11.7 ± 0.0	0.999
1 week	11.2 ± 0.1	11.4 ± 0.3	0.422
4 weeks	10.6 ± 0.1	10.9 ± 0.0	0.116
Photopic Cone Response
Amplitude (μV)
Baseline	114.8 ± 3.9	119.9 ± 2.9	0.350
3 days	100.6 ± 3.2	90 ± 2.4	0.056
1 week	111.0 ± 3.1	113.6 ± 4.1	0.640
4 weeks	112.7 ± 4.0	105.1 ± 2.4	0.178
Implicit time (msec)
Baseline	29.0 ± 0.1	29.2 ± 0.1	0.519
3 days	29.4 ± 0.5	29.8 ± 0.5	0.577
1 week	30.9 ± 0.4	30.7 ± 0.4	0.841
4 weeks	28.9 ± 0.0	29.0 ± 0.1	0.374

 Probabilities (6.25 mg of bevacizumab versus 0.25 mL saline at each point) dark-adapted scotopic rod, dark-adapted bright-flash (rod and cone), and light-adapted photopic cone ERG responses were determined by t-test. Data are the mean ± SE. No differences were significant.

Table 4.

 

Amplitudes and Implicit Times in Monkey 2, with 6.25 mg Bevacizumab or 0.25 mL Saline

	Bevacizumab (Right Eye)	Control (Left Eye)	P
Rod Response b-Wave
Amplitude (μV)
Baseline	228.8 ± 8.6	207.5 ± 0.2	0.070
3 days	257.3 ± 3.7	240.7 ± 16.2	0.375
1 week	251.5 ± 5.0	243.0 ± 7.7	0.409
4 weeks	200.2 ± 3.3	195.3 ± 6.1	0.519
Implicit time (msec)
Baseline	93.0 ± 2.0	88.5 ± 7.2	0.575
3 days	90.2 ± 2.6	91.6 ± 1.4	0.664
1 week	97.4 ± 1.1	94.1 ± 3.1	0.384
4 weeks	98.4 ± 1.2	99.5 ± 0.1	0.429
Flash ERG b-Wave
Amplitude (μV)
Baseline	285.2 ± 15.5	306.0 ± 1.5	0.252
3 days	300.6 ± 5.6	318.9 ± 9.7	0.178
1 week	327.1 ± 9.1	329.0 ± 2.3	0.846
4 weeks	297.9 ± 3.5	299.9 ± 15.3	0.903
Implicit time (msec)
Baseline	69.5 ± 0.3	70.8 ± 1.9	0.525
3 days	70.6 ± 0.3	72.9 ± 1.0	0.091
1 week	69.3 ± 0.3	67.4 ± 2.2	0.434
4 weeks	77.6 ± 5.6	73.9 ± 2.4	0.579
Flash ERG a-Wave
Amplitude (μV)
Baseline	−204.5 ± 2.9	−189.3 ± 5.0	0.572
3 days	−194.0 ± 8.2	−195.9 ± 8.3	0.883
1 week	−218.7 ± 6.2	−214.7 ± 1.2	0.564
4 weeks	−203.4 ± 6.8	−189.9 ± 13.1	0.413
Implicit time (msec)
Baseline	12.5 ± 0.2	12.0 ± 0.1	0.060
3 days	12.4 ± 0.2	12.0 ± 0.2	0.236
1 week	12.2 ± 0.1	12.4 ± 0.2	0.457
4 weeks	13.9 ± 0.3	13.5 ± 0.3	0.364
Photopic Cone Response
Amplitude (μV)
Baseline	58.4 ± 3.3	63.6 ± 6.6	0.519
3 days	79.4 ± 4.5	84.2 ± 1.8	0.388
1 week	58.9 ± 1.4	61.9 ± 3.3	0.455
4 weeks	46.0 ± 1.4	47.1 ± 2.5	0.713
Implicit time (msec)
Baseline	34.7 ± 1.2	31.5 ± 0.2	0.058
3 days	28.7 ± 0.2	28.4 ± 0.1	0.577
1 week	35.3 ± 1.1	33.2 ± 0.2	0.137
4 weeks	38.8 ± 0.5	40.6 ± 1.5	0.313

 Probabilities (6.25 mg of bevacizumab versus 0.25 mL saline at each point) of dark-adapted scotopic rod, dark-adapted bright-flash (rod and cone), and light-adapted photopic cone ERG responses were determined by t-test. Data are the mean ± SE. No differences were significant.

Table 5.

 

Amplitudes and Implicit Times in Monkey 3, with 12.5 mg Bevacizumab or 0.5 mL Saline

	Bevacizumab (Right Eye)	Control (Left Eye)	P
Rod Response b-Wave
Amplitude (μV)
Baseline	319.6 ± 3.5	327.0 ± 1.0	0.114
3 days	312.0 ± 4.0	339.5 ± 2.0	0.004*
1 week	229.1 ± 3.1	280.7 ± 6.1	0.002*
4 weeks	368.9 ± 5.6	365.1 ± 7.2	0.698
Implicit time (msec)
Baseline	72.4 ± 0.7	75.4 ± 0.8	0.052
3 days	78.9 ± 0.5	78.5 ± 0.6	0.629
1 week	90.5 ± 2.7	73.8 ± 0.5	0.004*
4 weeks	78.9 ± 3.4	85.3 ± 2.8	0.220
Flash ERG b-Wave
Amplitude (μV)
Baseline	436.8 ± 13.2	460.5 ± 8.5	0.204
3 days	456.6 ± 25.7	503.9 ± 34.3	0.332
1 week	362.7 ± 14.9	394.5 ± 12.4	0.178
4 weeks	486.7 ± 15.4	483.6 ± 18.3	0.901
Implicit time (msec)
Baseline	69.9 ± 0.4	68.8 ± 0.6	0.182
3 days	71.5 ± 0.4	71.3 ± 0.7	0.816
1 week	59.4 ± 5.9	56.3 ± 5.5	0.719
4 weeks	64.4 ± 5.7	64.3 ± 5.8	0.988
Flash ERG a-Wave
Amplitude (μV)
Baseline	−237.2 ± 8.4	−259.0 ± 1.8	0.063
3 days	−210.2 ± 3.2	−258.8 ± 8.5	0.006*
1 week	−171.7 ± 6.1	−195.1 ± 1.2	0.020*
4 weeks	−244.9 ± 8.6	−251.9 ± 10.2	0.624
Implicit time (msec)
Baseline	10.8 ± 0.1	10.8 ± 0.1	0.999<
3 days	11.3 ± 0.2	11.6 ± 0.1	0.284
1 week	11.8 ± 0.1	10.4 ± 0.1	0.001*
4 weeks	10.8 ± 0.1	10.6 ± 0.1	0.519
Photopic Cone Response
Amplitude (μV)
Baseline	99.2 ± 3.6	105.5 ± 2.5	0.226
3 days	94.9 ± 1.1	98.6 ± 2.2	0.203
1 week	92.0 ± 2.7	87.4 ± 2.7	0.301
4 weeks	88.0 ± 2.9	95.1 ± 5.1	0.288
Implicit time (msec)
Baseline	32.7 ± 0.5	32.0 ± 0.0	0.238
3 days	29.6 ± 0.1	29.2 ± 0.1	0.148
1 week	27.8 ± 0.1	28.7 ± 0.6	0.231
4 weeks	30.6 ± 0.1	28.9 ± 0.1	0.927

 Probabilities (12.5 mg of bevacizumab versus 0.50 mL saline at each point) of dark-adapted scotopic rod, dark-adapted bright-flash (rod and cone), and light-adapted photopic cone ERG responses were determined by t-test. Data are the mean ± SE.

*   P < 0.05.

Table 6.

 

Amplitudes and Implicit Times in Monkey 4, with 12.5 mg Bevacizumab or 0.5 mL Saline

	Bevacizumab (Right Eye)	Control (Left Eye)	P
Rod Response b-Wave
Amplitude (μV)
Baseline	247.5 ± 11.3	241.1 ± 24.8	0.824
3 days	198.7 ± 2.5	195.1 ± 2.1	0.324
1 week	117.0 ± 1.9	186.4 ± 3.4	<0.001*
4 weeks	208.3 ± 3.1	198.4 ± 4.9	0.163
Implicit time (msec)
Baseline	99.5 ± 0.1	97.7 ± 0.7	0.057
3 days	96.5 ± 0.8	96.6 ± 1.9	0.940
1 week	124.8 ± 2.1	99.3 ± 0.3	0.001*
4 weeks	93.4 ± 4.6	95.4 ± 0.1	0.676
Flash ERG b-Wave
Amplitude (μV)
Baseline	247.4 ± 3.6	243.3 ± 1.6	0.357
3 days	252.0 ± 9.7	249.8 ± 20.9	0.932
1 week	251.1 ± 11.2	270.6 ± 5.7	0.198
4 weeks	304.8 ± 2.2	289.9 ± 5.4	0.065
Implicit time (msec)
Baseline	69.2 ± 1.3	62.0 ± 2.6	0.066
3 days	64.5 ± 6.6	64.9 ± 6.1	0.723
1 week	71.7 ± 0.9	69.2 ± 0.4	0.080
4 weeks	66.7 ± 2.8	65.3 ± 2.9	0.752
Flash ERG a-Wave
Amplitude (μV)
Baseline	−160.2 ± 4.0	−149.1 ± 2.3	0.075
3 days	−159.7 ± 2.0	−146.3 ± 5.1	0.146
1 week	−143.7 ± 3.5	−151.5 ± 4.6	0.253
4 weeks	−216.8 ± 8.1	−201.6 ± 2.0	0.143
Implicit time (msec)
Baseline	12.6 ± 0.6	13.9 ± 0.2	0.091
3 days	14.2 ± 0.1	14.1 ± 0.4	0.691
1 week	13.7 ± 0.0	13.3 ± 0.6	0.548
4 weeks	12.4 ± 0.1	12.9 ± 0.2	0.116
Photopic Cone Response
Amplitude (μV)
Baseline	43.1 ± 0.5	40.5 ± 1.8	0.256
3 days	37.4 ± 0.9	36.1 ± 0.3	0.299
1 week	35.3 ± 1.1	38.1 ± 1.1	0.157
4 weeks	44.9 ± 1.1	43.5 ± 1.1	0.430
Implicit time (msec)
Baseline	29.3 ± 0.4	29.0 ± 0.1	0.561
3 days	31.2 ± 0.2	30.3 ± 0.1	0.062
1 week	32.6 ± 1.3	30.2 ± 0.1	0.156
4 weeks	29.7 ± 0.2	31.1 ± 0.5	0.061

 Probabilities (12.5 mg of bevacizumab versus 0.50 mL saline at each point) of dark-adapted scotopic rod, dark-adapted bright (rod and cone), and light-adapted photopic cone ERG responses were determined by t-test. Data are the mean ± SE.

*   P < 0.05.

Figure 9.

 

A representative ERG (monkey 3) showing b-wave of the dark-adapted scotopic rod responses at baseline (A), 3 days (B), 1 week (C), and 4 weeks (D) in the right eye injected with 12.5 mg bevacizumab and the left eye injected with 0.50 mL saline. Top trac e: experimental eye; bottom trac e: control eye.

Figure 10.

 

A representative ERG (monkey 3) showing dark-adapted bright flash responses between the experimental and control eyes at baseline (A), 3 days (B), 1 week (C), and 4 weeks (D). Top trace: experimental eye; bottom trace: control eye.

Figure 11.

 

A representative light-adapted photopic cone ERG (monkey 3) shows no detectable difference in the responses between the experimental and control eyes at baseline (A), 3 days (B), 1 week (C), and 4 weeks (D). Top trace: experimental eye; bottom trace: control eye.

Figure 12.

 

Histologic samples (A–F) from the posterior fundus of the monkey retinas 4 weeks after an intravitreal injection of 0.50 mL saline versus 6.25 or 12.5 mg bevacizumab. In the eyes injected with 6.25 mg bevacizumab (monkey 1) (C, D) or 12.5 mg bevacizumab (monkey 3) (E, F), all retinal layers were normal compared with the control eye of monkey 3 (A, B) injected with 0.50 mL of saline. Hematoxylin and eosin; magnification: (A, C, E) ×100; (B, D, F) ×200. Data indicate the retinal thickness (micrometers) ± SE measured by using histologic sections at a point 1.5 mm from the optic disc in the eye injected with 6.25 mg bevacizumab (left), 0.25 mL saline (right) (G), 12.5 mg bevacizumab (left) and 0.50 mL saline (right) (H).