ORIGINAL ARTICLE

JOURNAL OF OCULAR PHARMACOLOGY AND THERAPEUTICS Volume 00, Number 00, 2014 ª Mary Ann Liebert, Inc. DOI: 10.1089/jop.2014.0037

Intravitreal Bevacizumab in Acute Central/Hemicentral Retinal Vein Occlusions: Three-Year Results of a Prospective Clinical Study Dan Ca˘luga˘ru1 and Mihai Ca˘luga˘ru 2

Abstract

Purpose: To prospectively evaluate the effects of intravitreal bevacizumab (IVB, Avastin; Genentech, Inc., San Francisco, CA) injections in patients with acute central/hemicentral retinal vein occlusions (C/HCRVOs) ( £ 1 month after the occlusion was diagnosed) over the course of 3 years. Methods: The study included 57 patients with unilateral acute C/HCRVOs. Initially, the treatment for acute C/ HCRVO patients consisted of 4 consecutive IVB injections administered off-label at a dose of 2.5 mg per injection, with each injection spaced *45 days apart. Thereafter, IVB therapy was flexible, and subsequent injections were administered during scheduled visits whenever a best corrected visual acuity (BCVA) loss of ‡ 5 Early Treatment Diabetic Retinopathy Study (ETDRS) letters occurred and/or iris/angle neovascularization appeared (regardless of the intraocular pressure level). Changes in the BCVA and foveal thickness (FT), number of IVB injections administered, and incidence of neovascular glaucoma (NVG) were estimated. Results: The increase in the BCVA score at month 36 was 17.15 (ETDRS letters) (P < 0.0001) in cases of nonischemic and 26.81 (ETDRS letters) (P < 0.01) in cases of ischemic occlusions. At the end of the follow-up, the proportion of BCVA score improvements greater than 15 ETDRS letters was similar in patients with both forms of occlusions (measured in 45.5% of nonischemic and 45.8% of ischemic patients) (P = 0.977). There were significant reductions in FT from baseline values to 230 – 40.50 mm (P = 0.0001) in patients with nonischemic occlusions and 270 – 40.50 mm (P = 0.0001) in patients with ischemic forms. There was a significant difference (P < 0.03) in the number of IVB injections administered in patients with nonischemic C/HCRVOs (8.7 – 1.58) compared to patients with ischemic occlusions (9.7 – 1.78). NVG occurred in 2 cases of ischemic occlusions. Conclusions: The 3-year IVB therapy provided sustained vision and FT gains in most phakic patients with acute C/HCRVOs, making this treatment option a rational and viable therapeutic strategy. Bevacizumab was more effective in patients with ischemic occlusions who required a significantly higher number of injections.

during the acute phase of the disease ( < 3-month duration of symptoms of a venous occlusive event).1–3 However, most of the currently available studies include only patients with intermediate (3–12 months) or late ( > 1 year) stages of venous occlusions associated with macular edema.4–9 The aim of this study was to prospectively evaluate the effects of bevacizumab therapy in patients with acute central/hemicentral retinal vein occlusions (C/HCRVOs) ( £ 1 month after the occlusion was diagnosed) over the course of 3 years. The ischemic and nonischemic forms of the disease were evaluated separately.

Introduction

I

ntravitreal antivascular endothelial growth factor (VEGF) therapy with bevacizumab, ranibizumab, and aflibercept quickly become incorporated into the clinical management of central retinal vein occlusion (CRVO). We believe that eyes receiving therapy immediately after diagnosis (within 1 month) may benefit more than those receiving delayed treatment. Visual outcomes are better the sooner treatment is performed after the occlusion forms. A few studies have investigated the use of anti-VEGF therapy in CRVOs

1 2

Department of Ophthalmology, University ‘‘Grigore T. Popa’’ Iasxi, Romania. Department of Ophthalmology, University of Medicine ‘‘Iuliu Hatxieganu,’’ Cluj-Napoca, Romania.

1

˘ LUGA ˘ RU AND CA ˘ LUGA ˘ RU CA

2

Methods Patient population The study included 57 patients with unilateral acute C/HCRVOs, was approved by the University ‘‘Grigore T. Popa’’ Iasxi Ethics Committee, and conformed to the tenets of the Declaration of Helsinki. C/HCRVO was defined by the following ocular fundus findings: dotted and flameshaped intraretinal hemorrhages [which are demonstrated in all 4 retinal quadrants in cases of CRVO and usually found in only 2 quadrants in cases of hemicentral retinal vein occlusion (HCRVO), although intraretinal hemorrhages may involve 1/3–2/3 of the retina in HCRVO10], engorgement and tortuosity of the venous system, papilloretinal edema, telangiectatic capillary bed, cotton wool spots, and angiographic evidence of prolongation retinal circulation times.1,10–13 The study included patients diagnosed with both CRVO and HCRVO, as the 2 occlusion types are pathogenetically similar. In cases of CRVO, the only existing central retinal vein trunk within the optic nerve is involved, whereas patients with HCRVO have 2 central retinal vein trunks as a congenital anomaly and develop an occlusion in only one of them.14,15 The following eligibility criteria were used: unilateral CRVO and HCRVO patients with the duration of symptoms suggesting a venous occlusive event of £ l month, phakic eyes with clear ocular media, and the ability and willingness to comply with both the proposed therapy and the prospective follow-up process. All participants provided written informed consent. Patients were excluded from the study if any of the following criteria were met: concomitant disease that may compromise evaluations of the study eye or induce complications such as active ocular inflammation or infection, prior treatment (filtering surgery, corneal transplantation, pars plana vitrectomy, or photocoagulation), aphakia or pseudophakia, opacification of the media, existence of peripheral anterior synechiae in the study eye, inability of the patient to discontinue anticoagulation therapy during the study, any vascular retinal disorders in the study eye, or agerelated macular degeneration (namely, drusen, geographic atrophy, and neovascular form) in either of the eyes.

Patient examination All patients underwent a comprehensive ophthalmological examination of both eyes, for example, the affected eye and the contralateral uninvolved eye. Examination of the unaffected eye helped us to determine a condition associated, such as ocular hypertension, glaucoma, or age-related degeneration. The best corrected visual acuity (BCVA) score was determined using the Early Treatment Diabetic Retinopathy Study (ETDRS) protocol (Snellen equivalent plus ETDRS letters).16 The visual field was carefully tested using the Goldmann perimeter with I2e, I4e, and V4e isopters. In cases of suspected concomitant primary glaucoma, Humphrey static achromatic automatic perimetry was used. A biomicroscopic examination was performed with both dilated and nondilated pupils, with both the pupillary margin and the iris anterior surface assessed. Routine gonioscopy was performed with a nondilated pupil to highlight the angle neovascularization (NV), which in rare cases may initially and precociously develop within the chamber angle

and not at the iris level. The ocular fundus was thoroughly evaluated by direct and indirect ophthalmoscopy and, if needed, using a contact lens. Optical coherence tomography (Stratus OCT; Carl Zeiss Meditec, Dublin, CA) was used to assess the morphology and thickness of the macula, optic disc, and retinal nerve fiber layer. Fluorescein angiography assessments were focused on the disc and macula during early examination times, on the middle retinal periphery of each quadrant during the intermediate phases, and again on the disc and macula during late stages ( > 5 min after fluorescein injection); iris images were obtained during the recirculation phase. Retinal capillary nonperfusion was angiographically measured on standard photographic fields, with a retinal area equal to the optic disc diameter used as a template.12 Eyes with at least 10 disc areas of retinal capillary nonperfusion and/or intraocular NV were classified as having retinal ischemia. Intraocular pressure (IOP) was determined using a Perkins applanation tonometer and adjusted according to the corneal thickness. Most of the tests performed (determination of the BCVA score, perimetry, biomicroscopy, gonioscopy, ocular fundus examination, and measurement of the IOP) were carried out at each follow-up visit. Optical coherence tomography and fluorescein angiography were done every 6 months during the entire followup period. Regarding systemic evaluations, apart from a routine medical evaluation (namely, for systemic arterial hypertension, diabetes, and dyslipidemia), an extensive checkup for systemic disorders is unnecessary in most patients with retinal vein occlusion.17 Given that hematological risk factors for spontaneous systemic venous thrombosis are only sporadically present in patients with retinal vein occlusions,18 not all patients with retinal vein occlusions need to be subjected to exhaustive hematological investigations (namely, determination of plasma homocysteine, Leiden mutation of V factor, C and S protein deficiencies, activated protein C resistance, and antithrombin and antiphospholipid antibodies). Such tests are necessary only when clearly indicated.

C/HCRVO classification C/HCRVOs were divided into 2 groups: nonischemic and ischemic forms. The eligibility criteria for acute nonischemic C/HCRVOs were as follows: a BCVA score > 20/ 400 Snellen equivalent in the affected eye; normal peripheral visual field with or without relative central scotoma; mild to moderate intraretinal hemorrhages and venous tortuosity involving 4 (CRVO) or 2 (HCRVO) retinal quadrants; rare ( £ 4), if any, cotton wool spots; perfused retinal capillaries or small and very limited focal retinal capillary dropouts ( < 10 disc areas of retinal capillary nonperfusion); and optic disc edema and varied grades of macular edema.11 The inclusion criteria for the ischemic type of acute C/ HCRVOs were determined based on the angiography result. In cases with angiographyically clear evidence of retinal capillary nonperfusion zones, the criteria included 10 or more disc areas of nonperfusion.4,12 If intraretinal hemorrhages prevented a clear angiographic evaluation of retinal capillary nonperfusion, the following parameters were considered: a BCVA score £ 20/400 Snellen equivalent; ability to see £ V/4e isopter based on the Goldmann perimeter; the presence of relative afferent pupillary defects in

BEVACIZUMAB FOR RETINAL VEIN OCCLUSIONS

patients with a normal fellow eye; extensive ocular fundus changes [striking amount of hemorrhages, venous tortuosity, cotton wool spots ( > 5), disc and macular edema]; and an IOP reduction in the occluded eye of ‡ 4 mmHg compared with the congener eye.4,11,12,15 An eye was classified as having ischemic C/HCRVO by the presence of at least 4 of these 5 parameters. Neovascular glaucoma (NVG) was defined19 as NV of the iris and/or anterior chamber angle associated with increased IOP.

Treatment and follow-up Initially, the treatment for acute C/HCRVO patients consisted of 4 consecutive intravitreal bevacizumab (IVB, Avastin; Genentech, Inc., San Francisco, CA) injections administered off-label at a dose of 2.5 mg per injection, with each injection spaced *45 days apart. Thereafter, IVB therapy was flexible, and subsequent injections were administered during scheduled visits whenever a visual acuity loss of ‡ 5 ETDRS letters occurred and/or iris/angle NV appeared (regardless of the IOP level). Panretinal photocoagulation was performed as soon as intraocular NV was diagnosed, unless it subsided after 2 consecutive IVB injections, administered 30 days apart, and topical steroids and cycloplegics were prescribed. In cases of elevated IOP, a topical fixed combination of timolol and dorzolamide (FCTD; Cosopt, Merck & CO., Inc., Whitehouse Station, NJ) was added. Unless IOP normalized in response to these treatments, surgery was advised after an additional IVB injection. Patient follow-up was carried out every 2 months during the first year and every 6 months thereafter for the next 2 years.

Main outcome measures The primary outcome measure was the mean change from the baseline BCVA score in the study eye compared to month 36. The secondary outcome measures included the following: the proportion of patients with a BCVA letter score gain of > 15 at the end of the study, the mean decrease from baseline foveal thickness (FT) compared to month 36, the cumulative prevalence of conversion from nonischemic C/HCRVOs to ischemic forms, the number of IVB injections administered during the follow-up period, and the incidence of NVG.

Statistical analysis Distribution normality was tested using the Kolmogorov– Smirnov test. The continuous variables were tested using Student’s t-tests (normally distributed data), and both the Mann–Whitney U and Wilcoxon tests were used for nonparametric distributions. Categorical data were analyzed using the Pearson chi-square test and Fisher’s exact test. Cumulative prevalence calculations were performed with the Kaplan–Meier estimation method. A P value of < 0.05 was considered statistically significant. Statistical analysis was performed using SPSS statistical software (version 16) and Epi Info 2000.

Results The average time to start of treatment was 17.94 – 11.02 days after the onset of occlusion symptoms (16.16 – 11.51

3

days for nonischemic and 21 – 9.64 days for ischemic occlusions; P = 0.11). Baseline characteristics of the 57 patients with unilateral acute C/HCRVOs are shown in Table 1. Four patients had HCRVOs and 53 patients experienced CRVOs. At baseline, 36 and 21 patients exhibited nonischemic and ischemic forms of C/HCRVOs, respectively. The average age of patients with nonischemic occlusions was 57.28 – 14.08 years, and that of patients with ischemic C/HCRVOs was 64.05 – 13.12 years (P = 0.078). Twenty patients were 55 years old or younger and 37 patients were older than 55 years. Differences in the proportions of the nonischemic and ischemic forms of occlusions between the 2 age groups were not statistically significant (P = 0.43). There were no significant differences in the proportions of men (P = 0.79) and women (P = 0.79), exhibiting nonischemic versus ischemic forms of venous occlusions. The initial BCVA score impairment, which was determined just immediately after the patient developed C/ HCRVO, was significantly more evident in the ischemic patients (7.6 – 22.28 ETDRS letters) than the nonischemic patients (48.6 – 24.25 ETDRS letters) (P = 0.0001). There were no obvious differences between the nonischemic and ischemic forms, with respect to the rate of primary openangle glaucoma (P = 0.89), ocular hypertension (P = 0.849), and associated systemic comorbidities (namely, systemic arterial hypertension: P = 0.816; diabetes: P = 0.947; dyslipidemia: P = 0.823; cardiovascular diseases: P = 0.499; and cerebrovascular diseases: P = 0.63). A significant difference in baseline FT values was found between patients with nonischemic (530.64 – 150.522 mm) and ischemic (621.33 – 34.60 mm) (P = 0.0018) C/HCRVOs. Three cases of nonischemic occlusions (2 CRVOs and 1 HCRVO) evolved into retinal ischemia after 12, 10, and 6 months of therapy (Table 2). The cumulative prevalence of conversion to retinal ischemia was 10% [95% confidence interval (CI), 2.11–26.52], as shown in Figure 1. At baseline, retinal ischemia was found in 21 C/HCRVO patients (20 with CRVOs and 1 with a HCRVO) (36.8%; 95% CI, 25–50). The cumulative prevalence of retinal ischemia increased to 50% (95% CI, 35.22–64.77) (Fig. 2) in month 36, after 3 cases of nonischemic venous occlusions converted to retinal ischemia. Intention-to-treat analysis (Table 3) revealed a significant improvement in the BCVA score at month 36 in patients with both nonischemic (65.75 – 19.96 ETDRS letters; P < 0.0001) and ischemic (34.41 – 17.3 ETDRS letters; P < 0.01) occlusions; the difference in the BCVA score between the 2 groups was significant (P < 0.0001). The increase in the BCVA score at month 36 was 17.15 ETDRS letters in cases of nonischemic and 26.81 ETDRS letters in cases of ischemic occlusions. Taking into account only the BCVA score improvements at the end of the follow-up period greater than 15 ETDRS letters in patients with both forms of occlusions, the proportions were similar (45.5% in nonischemic and 45.8% in ischemic patients; P = 0.977). There were significant reductions in FT from baseline values to 230.15 – 17.69 mm (P = 0.0001) in patients with nonischemic occlusions and 270 – 40.506 mm (P = 0.0001) in patients with ischemic forms. There was a significant difference (P < 0.03) in the number of IVB injections administered in patients with nonischemic C/HCRVOs (8.7 – 1.58) compared to patients

˘ LUGA ˘ RU AND CA ˘ LUGA ˘ RU CA

4

Table 1.

Baseline Characteristic in 57 Central/Hemicentral Retinal Vein Occlusion Patients Treated with Intravitreal Bevacizumab Central/hemicentral retinal vein occlusion form

Variables Age (years) Mean – SD Median (range) Stratification, n (%) £ 55 > 55 Gender, n (%) Male Female Occlusion type, n (%) CRVO HCRVO BCVA score Mean – SD (Snellen equivalent) Median (range) (Snellen equivalent)

Nonischemic (n = 36)

Ischemic (n = 21)

P-value

Total (n = 57)

57.28 – 14.08 61.5 (21.5–75)

64.05 – 13.12 64 (41–85)

0.078 0.56

59.77 – 14.01 62 (21–85)

14 (38.8) 22 (61.1)

6 (28.5) 15 (71.4)

0.43 0.43

20 (35.8) 37 (64.91)

19 (52.7) 17 (47.2)

16 (76.1) 5 (23.8)

0.079 0.079

35 (61.40) 22 (38.5)

33 (91.6) 3 (8.3)

20 (95.2) 1 (4.7)

0.97 0.97

53 (92.98) 4 (7.01)

20/61 – 20/59 20/154 (20/317–20/20) Mean – SD (ETDRS letters) 48.6 – 24.25 Median (range) (ETDRS letters) 40 (25–85) Stratification, n (%) (Snellen equivalent [ETDRS letters]) £ 20/400 ( £ 20 letters) — > 20/400–20/200 (21–35 letters) 18 (50) > 20/200–20/100 (36–50 letters) 4 (11.1) > 20/100–20/50 (51–65 letters) 2 (5.5) > 20/50 ( ‡ 66 letters) 12 (33.3) Foveal thickness (mm) Mean – SD (mm) 530.64 – 150.522 Median (range) (mm) 505 (405–631.25) Primary open angle glaucoma, n (%) 7 (19.4) Ocular hypertension, n (%) 8 (22.2) Systemic comorbidities, n (%) Arterial systemic hypertension 20 (55.5) Diabetes 5 (13.8) Dyslipidemia 6 (16.6) Cardiovascular diseases 9 (25) Cerebrovascular diseases 1 (2.7)

20/435 – 20/476 20/500 (20/20000–20/125) 7.6 – 22.28 5 ( - 27)–( + 45) 15 (71.4) 5 (23.8) 1 (4.76) — —

< 0.0001 0.014 < 0.0001 0.014 < 0.0001 0.0518 0.739 0.52 0.002

20/89 – 20/66 20/317 (20/20000–20/20) 33.5 – 30.6 25 ( - 27)–( + 85) 15 (26.3) 23 (40.3) 5 (8.77) 2 (3.50) 12 (21.05)

621.33 – 34.60 610 (600–645) 3 (14.2) 5 (23.3)

0.0018 0.0001 0.89 0.849

564.05 – 128.59 586 (476–640) 10 (17.5) 13 (22.8)

11 (52.3) 2 (9.52) 2 (9.52) 7 (33.3) —

0.816 0.947 0.823 0.499 0.63

31 (54.3) 7 (12.2) 8 (14.03) 16 (28) 1 (1.75)

BCVA, best corrected visual acuity; CRVO, central retinal vein occlusion; ETDRS, Early Treatment Diabetic Retinopathy Study; HCRVO, hemicentral retinal vein occlusion.

with ischemic occlusions (9.7 – 1.78). NVG occurred in 2 cases of ischemic CRVOs, 1 at month 12 and the other at month 18; the incidence of NVG in patients with ischemic C/HCRVOs was 8.3. Three patients with nonischemic CRVOs were lost to follow-up: 1 in month 8, 1 in month 12, and 1 in month 24. One patient with ischemic CRVO withdrew from the study in month 10 of the follow-up period. Two patients with ischemic CRVOs died during the study period, at 12 and 30 months, after occlusion onset. No adverse effects or ocular toxicity, including clinically evident sterile or infectious endophthalmitis, traumatic cataract, intraocular inflammation, IOP increase, retinal ruptures, retinal detachment and systemic thromboembolic events, was encountered during the study.

Discussion In the present study, there was a statistically significant difference in the initial BCVA scores between the non-

ischemic and ischemic groups (P < 0.0001); this significant difference was also found at month 36 (P < 0.0001). On the other hand, the BCVA score improvements at month 36 were significant in patients with both nonischemic (P < 0.0001) and ischemic (P < 0.01) occlusions (Table 3). However, the magnitudes of response to treatment were different, namely, an increase in the BCVA score of 17.15 ETDRS letters in cases of nonischemic and 26.81 ETDRS letters in patients of ischemic occlusions. The proportions of BCVA score increases (from baseline values) were 35.2% and 352.7%, respectively. Such a comparison helped us to know that IVB therapy was more effective in patients with ischemic occlusions. This assertion was already mentioned by Ferrara et al.1 The authors noted that patients with the worst initial visual acuity showed the greatest benefit from the treatment. In contrast, the Rave trial20 reported that despite significant clinical benefits with intravitreal ranibizumab (Lucentis; Genentech, Inc., South San Francisco, CA) in eyes with ischemic CRVO, the risk of neovascular complications was not ameliorated by VEGF blockade, but was merely delayed.

6 > 10 disc areas of RCN

12

10 > 10 disc areas of RCN

10

20/125 (45 ETDRS letters) 20/160 (41 ETDRS letters) 20/125 (45 ETDRS letters) 11 12 Iris neovascularization

HCRVO Female 3

66

CRVO Female 2

73

Male 1

IVB, intravitreal bevacizumab; RCN, retinal capillary nonperfusion.

Systemic hypertension; Ischemic cardiopathy Diabetes; Anticoagulant treatment Systemic hypertension; Ischemic cardiopathy 20/200 (35 ETDRS letters) 20/320 (25 ETDRS letters) 20/320 (25 ETDRS letters) CRVO

Risk factors Gender

62

No. of IVB injections Time up to conversion (months) Stigmata of retinal ischemia Initial BCVA score Venous occlusion type Age (years) Case No.

Table 2.

Nonischemic Central/Hemicentral Retinal Vein Occlusions Converted to Ischemic Forms

Final BCVA score

BEVACIZUMAB FOR RETINAL VEIN OCCLUSIONS

5

The improvements of the BCVA EDTRS letter score of greater than 15 were similar in patients with both types of occlusions (measured in 45.5% of nonischemic and 45.8% of ischemic patients) (P = 0.977). However, it should be noted that to achieve this result, patients with ischemic C/ HRVOs required a significantly greater number of IVB injections (P < 0.03) than those with nonischemic forms. There was a significant difference in FT between patients with the 2 occlusion forms both initially (P = 0.0018) and at month 36 (P = 0.0001). The magnitudes of response to treatment were different, for example, a decrease of 300.49 mm in cases of nonischemic and 351.33 mm in patients of ischemic occlusions. However, the proportions of FT reductions (from baseline values) were similar in the nonischemic and ischemic groups (56.62% and 56.54%, respectively). So, the anatomical outcomes did not mirror the visual results mentioned above. There are a small number of studies reporting the results of anti-VEGF therapy in patients with CRVO in the acute phase. In a retrospective study of 5 CRVO cases treated with IVB, within 3 months of occurrence, Ferrara et al.,1 found dramatic improvements in the BCVA score at a 12-month follow-up especially in patients with ischemic occlusions. Algvere et al.,3 reported a significant improvement of 18 ETDRS letters in 13 patients (12 with nonischemic and 1 with ischemic occlusions) with a mean CRVO duration of 2.5 months; those patients received an average of 7.4 IVB injections during a follow-up period of 18 months. The Copernicus Study2 analyzed the benefit of intravitreal aflibercept (IVA, VEGF Trap-Eye; Regeneron Pharmaceuticals, Inc., Tarrytown, NY) injections in patients with macular edema secondary to CRVO. Patients received 6 monthly intravitreal injections of 2 mg aflibercept, and the therapy continued for 52 weeks on an as-needed basis (with 2.7 injections required on average). At the end of the followup period, 55.3% of patients gained at least 15 ETDRS letters, and the BCVA score improved by an average of 16.2 ETDRS letters. These results at week 52 are consistent with our results obtained after bevacizumab therapy, which indicate that 45.6% of patients gained more than 15 ETDRS letters, and that the mean BCVA score improved by 19.06 ETDRS letters. However, there are major differences between the 2 studies, with regard to study design, study population, study duration, and the mean time of initiation of the treatment. The Copernicus Study2 had a lower percentage of patients with posterior nonperfusion (15.5%), while such patients accounted for 50% of the cases in the present study. The durations of the Copernicus Study2 and our study were completely different (52 weeks vs. 36 months), but the primary endpoints (a gain of at least 15 compared to more than 15 ETDRS letters) were achieved with approximately the same number of intravitreal injections (8.7 aflibercept injections versus 9.1 bevacizumab injections). The mean time that elapsed between initiation of the treatment and the onset of occlusion symptoms was different in both studies, for example, 2.73 months in the Copernicus Study2 and 17.94 days in our series. The patients in the Copernicus Study were also analyzed after a follow-up period of 100 weeks. Between week 522 and week 100,21 the therapy was given on an as-needed basis (with 3.3 IVA injections required on average). At the end of the follow-up, a decline in the visual and anatomic improvements from those gained at week 52 occurred (mean

6

˘ LUGA ˘ RU AND CA ˘ LUGA ˘ RU CA

FIG. 1. The Kaplan–Meier estimate of cumulative probability of conversion to retinal ischemia along the entire course of the study in 36 patients with unilateral acute nonischemic central/hemicentral retinal vein occlusions, all of them being considered as patients at risk. Patients who did not develop conversion and were lost before the end of the study were censored from the last completed visit. The dotted lines represent the upper and lower bounds of the 95% confidence interval.

loss of 6.2% in patients who gained at least 15 ETDRS letters; mean loss of 3.2 ETDRS letters in BCVA score; mean increase of 23 mm in FT). The 100-week results of the Copernicus Study21 are different from our data obtained after bevacizumab therapy. The inconsistency existing be-

FIG. 2. The Kaplan–Meier estimate of the cumulative prevalence of global retinal ischemia along the entire course of the study in 57 patients with unilateral acute central/hemicentral retinal vein occlusions. Patients at risk are those who did not present with retinal ischemia at the beginning of the study period. Patients who were lost before the end of the study were censored from the last completed visit, and those who died were censored at their date of death. The dotted lines represent the upper and lower bounds of the 95% confidence interval.

tween our data and those of the Copernicus Study21 is likely owing to differences in the mean time from CRVO diagnosis to initiation of the therapy (CRVO was diagnosed within 9 months before initiation of the treatment in the Copernicus Study21 and within 1 month in our series). Any

BEVACIZUMAB FOR RETINAL VEIN OCCLUSIONS

Table 3.

7

Intention-to-Treat Analysis in 57 Central/Hemicentral Retinal Vein Occlusion Patients Treated with Intravitreal Bevacizumab Central/hemicentral retinal vein occlusion form

Variables

Nonischemic (n = 33)

Ischemic (n = 24)

BCVA score at 36 months Mean – SD (Snellen equivalent) 20/32 – 20/36 20/144 – 20/104 Median (range) (Snellen equivalent) 20/50 (20/125–20/12) 20/160 (20/4000–20/20) Mean – SD (ETDRS letters) 65.75 – 19.96a 34.41 – 17.3b Median (range) (ETDRS letters) 65 (45–96) 40 (5–85) Stratification, n (%) (Snellen equivalent [ETDRS letters]) £ 20/400 ( £ 20 letters) — 7 (29.1) > 20/400–20/200 (21–35 letters) — 4 (16.6) > 20/200–20/100 (36–50 letters) 12 (36.3) 11 (45.8) > 20/100–20/50 (51–65 letters) 8 (24.2) 1 (4.16) > 20/50 ( ‡ 66 letters) 13 (39.3) 1 (4.16) Visual improvement > 15 ETDRS letters% (95% CI) 45.5 (28.1–63.6) 45.8 (22.5–67.2) Foveal thickness (mm) at month 36 Mean – SD (mm) 230.15 – 17.69d 270 – 40.50e Median (range) (mm) 223 (214–248.50) 267.50 (240–300) Follow-up/months Mean – SD 34.7 – 6.857 34.6 – 7.55 Median (range) 36 (8–39) 37 (10–39) No injections of bevacizumab per patient Mean – SD 8.7 – 1.58 9.70 – 1.78 Median (range) 9 (4–12) 9.5 (5–13) NVG incidence, n (%) 0 (0) 2 (8.3)

P-value

Total (n = 57)

< 0.0001 20/48 – 20/40 0.001 20/118 (20/4000–20/12) < 0.0001 52.56 – 24.3c 0.001 45 (5–95) < 0.001 < 0.02 0.739 0.52 < 0.002

7 (12.2) 4 (7.01) 23 (40.3) 9 (15.7) 14 (24.5)

0.977

45.6 (32.4–59.3)

0.0001 0.006

246.93 – 35.30f 247 (220–262.50)

0.49 0.703

34.68 – 7.09 36 (8–39)

< 0.03 0.139 0.17

9.14 – 1.72 9 (4–13) 2 (3.5)

Mean change in BCVA score from baseline values (Table 1): aP < 0.0001; bP < 0.01; cP < 0.0001; mean change in foveal thickness from baseline values (Table 1): dP = 0.0001; eP = 0.0001; fP < 0.0001. CI, confidence interval; NVG, neovascular glaucoma.

delay in initiating anti-VEGF treatment may result in a visual penalty, which hardly can be recuperated even with subsequent treatment, highlighting the importance of early treatment in patients with acute C/HRVOs. However, anti-VEGF therapy has been extensively used to treat patients with venous occlusions in the intermediate (3– 12-month duration of symptoms) and late ( > 1 year) stages, which are associated with macular edema. A small series of uncontrolled and nonrandomized cases were treated with IVB,4,6 with follow-up periods between 3 and 12 months. Short-term results have shown that there is a marked reduction in macular edema and an improvement in the BCVA score within 4–6 weeks of treatment. In the long-term studies, patients treated with bevacizumab were followed for 12 months,22,23 18 months,3 and 24 months.24 Long-term studies have revealed that there is a significant gain in the BCVA score following IVB injections. To maintain visual improvement, repeated injections are necessary as long as the disease is active. The same result has been shown to be true for ranibizumab therapy in patients with CRVOs associated with macular edema. Sham-controlled studies8,9 and uncontrolled studies5,7 have shown clinically significant anatomical and visual improvements, but they stressed the need for repeated, monthly intravitreal injections to maintain the initial positive outcomes. In the long-term study by Chang et al.,25 there was significant improvement in visual acuity and central retinal thickness, which persisted up to 2 years (with minimal side effects) in patients with CRVOs, who received an average of 10.2 injections of 0.5 mg ranibizumab during the first year and 6.6 injections during the second year.

The CRUISE Study,26 a 12-month trial, analyzed the benefit of intravitreal ranibizumab injections in patients with macular edema secondary to CRVO, who were treated at a mean time of 3.3 months since CRVO diagnosis. Patients were divided into 3 groups (sham, 0.3 mg, and 0.5 mg of ranibizumab) and received 6 monthly injections of 0.3 or 0.5 mg ranibizumab or sham injections. After 6 months, all patients received ranibizumab 0.5 mg on an as-needed basis (with a mean number of 3.7, 3.8, and 3.3 injections, respectively, in the sham/0.5, 0.3, and 0.5 mg groups). At the end of the follow-up period, 33.1%, 47%, and 50.8% of patients gained at least 15 ETDRS letters, respectively, the BCVA score improved by an average of 7.3, 13.9, and 13.9 ETDRS letters, respectively, and FT decreased with a mean of 427.2, 452.8, and 462.1 mm, respectively. The 12-month results of the CRUISE Study26 are consistent with our data achieved after bevacizumab therapy. However, the Horizon trial,27 which is a 12-month, open-label, single-arm extension of the CRUISE Study26 for the treatment of macular edema following CRVO, revealed different outcomes. Patients, who were divided into 3 groups (sham, 0.3 mg of ranibizumab, and 0.5 mg of ranibizumab), received a mean number of 2.9, 3.8, and 3.5 injections, respectively, of 0.5 mg intravitreal ranibizumab during the follow-up period on an as-needed basis. At month 12, there was a worsening of outcome measures [decrease in the BCVA score ( - 4.2, - 5.2, and - 4.1, ETDRS letters, respectively), and increase in the FT (63, 89, and 72 mm, respectively)]. Most patients included in this study (98.5%) experienced nonischemic occlusions; cases with a positive, brisk, relative afferent

˘ LUGA ˘ RU AND CA ˘ LUGA ˘ RU CA

8

pupillary defect, who required treatment, were excluded. Exclusion of these patients from the study may have eliminated patients with extensive retinal capillary nonperfusion (eg, those who are mainly at high risk of developing NVG). In fact, patients with nonischemic occlusions have a much better prognosis than those with ischemic forms, even without treatment. The inconsistency existing between our data and those of the Horizon Study27 may be explained by the different durations of the period of time from CRVO diagnosis to initiation of the treatment. So, in the Horizon Study,27 the diagnosis of CRVO was made within 12 months before initiation of screening and within 1 month in our series. Brynskov et al.28 reported, in clinical practice, a mean improvement in the BCVA score of 1.8 ETDRS letters (P = 0.50) with intravitreal ranibizumab for CRVOs after a 1-year follow-up period. The results of this study for CRVO failed to reproduce the outcomes of the CRUISE trial.26 Sophie and al.29 performed a 6-year follow-up of CRVO patients treated monthly with intravitreal ranibizumab for 3 months, followed by the administration of injections, as needed, for recurrent/persistent macular edema. The authors reported that 25% of the patients were cured only with injections after an average of 14.0 months. However, many patients eventually lost some of the early benefits, partly due to the progression of retinal nonperfusion, which was prevented by monthly rather than intermittent injections of ranibizumab. We believe that these patients with intermediate and late stages of venous occlusions associated with macular edema most likely had permanent retinal capillaropathy, which was temporarily relieved by reduction of the edematous component with bevacizumab or ranibizumab. However, the pathology was incurable due to the irreversible ischemic degenerative cystoid macular changes, which sometimes resulted in the appearance of a lamellar macular hole. The findings of the Copernicus Study2 support the aforementioned assertion. Thus, this study demonstrated that higher proportions of eyes gained at least 15 ETDRS letters at week 52 in the subgroup of patients who received IVA £ 2 months after the occurrence of CRVO symptoms (64.1%) compared with those who were treated > 2 months after the disease was diagnosed (42.9%). The authors of this study suggest that a 6-month delay in providing IVA therapy may be too long because damages from chronic macular edema may limit the recovery of visual acuity. The rationale for administering early IVB treatment to patients with acute occlusions included the following: initial abrogation of the increased VEGF levels in the acute phase, which are responsible for the main symptoms and complications, most of which occur in the natural clinical course during the first 7–8 months of the disease (macular edema, retinal capillary nonperfusion, NV, and NVG30); binding of the bevacizumab to all VEGF-A isoforms, preventing their attachment to receptors situated on the endothelial cell surface31; rapid, effective, and direct blocking of the neovascular process and its complications32; reversal of increased vascular permeability mediated by VEGF,27,31 ensuring the stability and integrity of the inner blood–retinal barrier; maintenance of a relatively normal or almost normal foveal anatomy during the acute phase of occlusion, when the VEGF levels are increased, until improvement of the draining circulation; prevention of acute, functional curable retinal capillaropathy, which is present immediately after the onset of occlusion to

develop into a permanent capillaropathy with limited reversal; and normalization of the long-term physiological VEGF expression, which is essential for vascular endothelial homeostasis,8 blood pressure homeostasis,33 and neuroprotection of retinal ganglion cells.34 To our knowledge, the assessment of the visual results of IVB treatment in patients with acute nonischemic and ischemic retinal vein occlusions, who are followed for at least 3 years, has not been previously reported. Our case series has several limitations. Although the study was prospective and included the long-term follow-up of patients with acute C/HCRVOs, it lacked both randomization and a control group. We were unable to create a control group because, all patients with acute C/HCRVOs since 2008 have been treated with anti-VEGF agents and carefully monitored.

Conclusions In conclusion, the 3-year IVB therapy provided sustained vision and anatomic gains in most phakic patients with acute C/HCRVOs, making this treatment option a rational and viable therapeutic strategy. Bevacizumab was more effective in patients with ischemic occlusions who required a significantly higher number of injections. Larger prospective studies are warranted to confirm our findings.

Acknowledgments D.C. and M.C. were involved in the design and conduct of the study; collection, management, analysis and interpretation of the data; preparation, review, or approval of the manuscript. Written informed consent according to the tenets of the Declaration of Helsinki was obtained from patients enrolled. Approval for the study was received from the Institutional Review Board/Ethics Committee of the University ‘‘Grigore T. Popa’’ Iasxi/Romania. This study has been performed in accordance with the ethical standards set forth in the Declaration of Helsinki.

Author Disclosure Statement There are no commercial associations that might create a conflict of interest in connection with the submitted manuscript. The authors have no proprietary or commercial interest in any of the materials discussed in this article. No competing financial interests exist.

References 1. Ferrara, D.C., Koizumi, H., and Spaide, R.F. Early bevacizumab treatment of central retinal vein occlusion. Am. J. Ophthalmol. 144:864–871, 2007. 2. Brown, D.M., Heier, J.S., Clark, W.L., et al. Intravitreal aflibercept injection for macular edema secondary to central retinal vein occlusion: 1-year results from the phase 3 COPERNICUS Study. Am. J. Ophthalmol. 155:429–437, 2013. 3. Algvere, P.V., Epstein, D., von Wendt, G., Seregad, S., and Kvanta, A. Intravitreal bevacizumab in central retinal vein occlusion: 18 month results of a prospective clinical trial. Eur. J. Ophthalmol. 21:789–795, 2011. 4. Iturralde, D., Spaide, R.F., Meyerle, C.B., et al. Intravitreal bevacizumab (Avastin) treatment of macular edema in central retinal vein occlusion. Retina. 26:279–284, 2006.

BEVACIZUMAB FOR RETINAL VEIN OCCLUSIONS

5. Campochiaro, P.A., Hafiz, G., Shah, S.M., et al. Ranibizumab for macular edema due to retinal vein occlusions: implication of VEGF as a critical stimulator. Mol. Ther. 16:791–799, 2008. 6. Kriechbaum, K., Prager. F., Geitzenauer, W., et al. Association of retinal sensitivity and morphology during antiangiogenic treatment of retinal vein occlusion over one year. Ophthalmology. 116:2415–2421, 2009. 7. Spaide, R.F., Chang, L.K., Klancnik, J.M., et al. Prospective study of intravitreal ranibizumab as a treatment for decreased visual acuity secondary to central retinal vein occlusion. Am. J. Ophthalmol. 147:298–306, 2009. 8. Kinge. B., Stordahl, P.B., Forsaa, V., et al. Efficacy of ranibizumab in patients with macular edema secondary to central retinal vein occlusion: results from the sham-controlled ROCC Study. Am. J. Ophthalmol. 150:310–314, 2010. 9. Brown, D.M., Campochiaro, P.A., Singh, R.P., et al. Ranibizumab for macular edema following central retinal vein occlusion. Ophthalmology. 117:1124–1133, 2010. 10. Hayreh, S.S., and Hayreh, M.S. Hemi-central retinal vein occlusion. Pathogenesis, clinical features, and natural history. Arch. Ophthalmol. 98:1600–1609, 1980. 11. The Central Vein Occlusion Study. Baseline and early natural history report. Arch. Ophthalmol. 111:1087–1095, 1993. 12. The Eye Disease Case-Control Study Group. Risk factors for central retinal vein occlusion. Arch. Ophthalmol. 114:545–554, 1996. 13. Recchia, F.M., Carvalho-Recchia, C.A., and Hassan, T.S. Clinical course of younger patients with central retinal vein occlusion. Arch. Ophthalmol. 122:317–321, 2004. 14. Coscas, G., and Dhermy, P. Occlusions Veineuses Retiniennes. Paris, New York, Barcelone, Milan: Masson; 1978; p. 277–345. 15. Hayreh, S.S., March, W., and Phelps, C.D. Ocular hypotony following retinal vein occlusion. Arch. Ophthalmol. 96:827–833, 1978. 16. Beck, R.W., Moke, P.S., Turpin, A.H. et al. A Computerized method of visual acuity testing: adaptation of the Early Treatment of Diabetic Retinopathy Study testing protocol. Am. J. Ophthalmol. 135:194–205, 2003. 17. Hayreh, S.S., Zimmerman, B., McCarthy, M.J., and Podhajsky, P. Systemic diseases associated with various types of retinal vein occlusion. Am. J. Ophthalmol. 131:61–77, 2001. 18. Hayreh, S.S., Zimmerman, M.B., and Podhajsky, P. Hematologic abnormalities associated with various types of retinal vein occlusion. Graefes Arch. Clin. Exp. Ophthalmol. 240:180–196, 2002. 19. Netland, P.A. The Ahmed glaucoma valve in neovascular glaucoma. Trans. Am. Ophthalmol. Soc. 107:325–342, 2009. 20. Brown, D.M., Wykoff, C.C., Wong, T.P., et al. Ranibizumab in preproliferative (ischemic) central retinal vein occlusion: the rubeosis anti-VEGF (Rave) trial. Retina. 34: 1728–1735, 2014. 21. Heier, J.S., Clark, W.L., Boyer, D.S., et al. Intravitreal aflibercept injection for macular edema due to central retinal vein occlusion. Two-year results from the Copernicus Study. Ophthalmology. 121:1414–1420, 2014. 22. Epstein, D.L., Algvere, P.V., von Wendt, G., Seregard, S., and Kvanta, A. Benefic from bevacizumab for macular edema in central retinal vein occlusion: twelve-month results of a prospective, randomized study. Ophthalmology. 119:2587–2591, 2012.

9

23. Daien, V., Navarre, S., Fesler, P., et al. Visual acuity outcome and predictive factors after bevacizumab for central retinal vein occlusion. Eur. J. Ophthalmol. 22:1013–1018, 2002. 24. Wu, L., Arevalo, J.F., Beroccal, M.H., et al. Comparison of two doses of intravitreal bevacizumab as primary treatment for macular edema secondary to central retinal vein occlusion; results of the pan American collaborative retina study group at 24 months. Retina. 30:1002–1011, 2010. 25. Chang, L.K., Spaide, R.F., Klancnik, J.M., et al. Longerterm outcomes of a prospective study of intravitreal ranibizumab as a treatment for decreased visual acuity secondary to central retinal vein occlusion. Retina. 31:821– 828, 2011. 26. Campochiaro, P.A., Brown, D.M., Awh, C.C., et al. Sustained benefits from ranibizumab for macular edema following central retinal vein occlusion: twelve-month outcomes in a phase III study. Ophthalmology. 118:2041– 2049, 2011. 27. Heier, J.S., Campochiaro, P.A., Yau, L., et al. Ranibizumab for macular edema due to retinal vein occlusions: long-term follow-up in the Horizon trial. Ophthalmology. 119:802– 809, 2012. 28. Brynskov, T.Y., Kemp, H., and Sorensen, T. Intravitreal ranibizumab for retinal vein occlusion through 1 year in clinical practice. Retina. 34:1637–1643, 2014. 29. Sophie, R., Hafiz, G., Scott, A.W., et al. Long-term outcomes in ranibizumab-treated patients with retinal vein occlusion: the role of progression of retinal nonperfusion. Am. J. Ophthalmol.156:693–705, 2013. 30. Hayreh, S.S., Klugman, M.R., Beri, M., Kimura, A., and Podhajsky, P. Differentiation of ischemic from non-ischemic central retinal vein occlusion during the early acute phase. Graefes Arch. Clin. Exp. Ophthalmol. 228:201–217, 1990. 31. Kaur, C., Foulds, W.S., and Ling, E.A. Blood-retinal barrier in hypoxic ischaemic conditions; basic concepts, clinical features and management. Prog. Retin. Eye Res. 27:622– 647, 2008. 32. Ouhadj, O., Chergui, I., Mendil, L., and Nouri, M.T. Intravitreal bevacizumab in the treatment of neovascular glaucoma (in French). J. Fr. Ophtalmol. 32:112–116, 2009. 33. Csaky, K., and Do, D.V. Safety implications of vascular endothelial growth factor blockade to subjects receiving intravitreal anti-vascular endothelial growth factor therapies. Am. J. Ophthalmol. 148:647–656, 2009. 34. Nishijima, K., Ng, Y.S., Zhong, L., et al. Vascular endothelial growth factor-A is a survival factor for retinal neurons and a critical neuroprotectant during the adaptive response to ischemic injury. Am. J. Pathol. 171:53–67, 2007.

Received: May 7, 2014 Accepted: September 10, 2014 Address correspondence to: Dr. Mihai Ca˘luga˘ru Department of Ophthalmology University of Medicine ‘‘Iuliu Hatxieganu’’ Strada Brancoveanu 11 Cluj-Napoca 3400 Romania E-mail: [email protected]