RANIBIZUMAB IN PREPROLIFERATIVE (ISCHEMIC) CENTRAL RETINAL VEIN OCCLUSION The Rubeosis Anti-VEGF (RAVE) Trial DAVID M. BROWN, MD, CHARLES C. WYKOFF, MD, PHD, TIEN P. WONG, MD, ANGELINE F. MARIANI, BA, DANIEL E. CROFT, BA, KARRI L. SCHUETZLE; FOR THE RAVE STUDY GROUP Purpose: To analyze the efficacy and safety of ranibizumab in eyes with preproliferative (ischemic) central retinal vein occlusion. Methods: In this prospective, phase I/II, open-label clinical trial, eyes at high risk of neovascular complications were identified; all eyes met $3 of 4 high-risk criteria: 1) the best-corrected visual acuity being #20/200, 2) loss of the 1-2e isopter on Goldmann visual field, 3) relative afferent pupillary defect being $0.9 log units, and 4) electroretinogram B-wave reduction to #60% of the corresponding A-wave. Monthly intravitreal ranibizumab treatment for 9 months, monthly monitoring for 3 months, and then monthly examination with pro re nata retreatment on evidence of disease activity for 24 months were performed. Therefore, the total study duration was 36 months. Results: The main outcome measures were mean change in the best-corrected visual acuity and central macular thickness by optical coherence tomography, proportion of patients with neovascular complications, and the incidence and severity of ocular and nonocular adverse events. Twenty patients were enrolled in the Rubeosis Anti-VEgf trial, and the mean number of intravitreal treatments administered through Months 24 and 36 were 14.1 and 17.2, respectively. The mean best-corrected visual acuity letters gained were +21.1 and +21.4 at 9 and 36 months, respectively. The mean central macular thickness improved −294 mm from baseline after 9 monthly treatments. Subsequently, after 3 months of observation, the mean central macular thickness increased +203 mm. On initiation of pro re nata ranibizumab retreatment, the mean central macular thickness then improved −191 mm at Month 36 compared with Month 12. Nine patients developed neovascular complications, being diagnosed after a mean of 24-month follow-up (range, 3–44 months), with 2 patients developing neovascularization after completion of the 36-month trial endpoint (at Months 42 and 44 after study enrollment). Conclusion: Intravitreal ranibizumab therapy can improve retinal anatomy and vision in eyes with severe central retinal vein occlusion. Despite significant clinical benefit with antivascular endothelial growth factor therapy, the risk of neovascular complications was not ameliorated by vascular endothelial growth factor blockade, but was merely delayed. RETINA 34:1728–1735, 2014
C
hemorrhage or anterior rubeosis with secondary neovascular glaucoma. Previous natural history and interventional studies have attempted to differentiate CRVO eyes that have a higher risk of neovascular complications. Hayreh et al1 identified criteria that predicted and differentiated CRVO eyes with a propensity for devastating neovascular complications (“hemorraghic retinopathy” or
entral retinal vein occlusion (CRVO) is a common vascular event that presents acutely with intraretinal hemorrhages throughout the retinal fundus and marked macular edema associated with decreased visual acuity. Although some eyes with CRVO spontaneously resolve, many have persistent macular edema or develop proliferative complications, including posterior neovascularization with vitreous
1728
1729
THE RAVE TRIAL BROWN ET AL
“ischemic CRVO” eyes) from eyes with a more benign CRVO course classified as “venous stasis retinopathy” or “non-ischemic CRVO.” The Central Retinal Vein Occlusion Study (CVOS) used an angiographic criteria of 10 disk areas or more of capillary nonperfusion on fundus fluorescein angiography to identify “ischemic CRVO,”2 whereas Hayreh identified functional tests (visual acuity, visual fields, relative afferent pupillary defect [RAPD], and electroretinography [ERG]) that demonstrated a high sensitivity and specificity for predicting “ischemic” CRVO. Intravitreal anti-vascular endothelial growth factor (VEGF) treatments have provided an extremely effective therapy for macular edema secondary to CRVO with reduction and/or elimination of macular edema and improved visual acuity with monthly injections.3–5 The phase III ranibizumab for CRVO trial (CRUISE) specifically excluded patients with a RAPD and those with visual acuity worse than 20/320. Only 2 of 334 eyes in CRUISE with adequately gradable fundus fluorescein angiography had .10 disk area of nonperfusion demonstrating how effective these 2 functional criteria were at excluding patients who would have been previously identified as “ischemic” by the CVOS. Although almost all of the eyes in the CRUISE trial would have historically been classified as “non-ischemic” by the CVOS or Hayreh criteria, the universal anatomical improvements seen with VEGF blockade imply that the thrombus in the central retinal vein must lead to variable amounts of retinal ischemia in all CRVO patients leading to the production of VEGF with subsequent macular edema. Because these trials implicate ischemia and VEGF in all patients with CRVO with macular edema, we use the term “pre-proliferative CRVO” to describe CRVO eyes with a high risk of neovascular complications. Hayreh et al6 demonstrated that in high-risk eyes (what in this report is referred to as preproliferative CRVO eyes), neovascular complications occur in approximately 2 of 3 eyes within 3 months without From the Retina Consultants of Houston, Houston Methodist Hospital, Houston, Texas. The authors received a research grant for this trial from Genentech. The funding organization had no role in the design or conduct of this research. Clinical Trial Registration—http://www.clinicaltrials.gov. Identifier: NCT00406471, NCT01225146. DMB: Consultant for Alcon, Allergan, Bayer, Genentech, Regeneron. Lecturer for Bayer and Roche. Research Support from Alcon, Allergan, Genentech, Regeneron. CCW: Consultant for Alcon, Allergan, Genentech. Lecturer for Genentech and Regeneron. Research Support from Alcon, Allergan, Genentech, Regeneron. TPW: Research Support from Alcon, Allergan, Genentech, Regeneron. AFM: None. DEC: None. KLS: None. Reprint requests: David M. Brown, MD, The Methodist Hospital, 6560 Fannin Street, Suite 750, Houston, TX 77030; e-mail: [email protected]
treatment. However, approximately 1 of 3 of these eyes avoids neovascular glaucoma and stabilizes within approximately 270 days presumably secondary to collateral formation or recanalization of the thrombus. The Rubeosis Anti-VEgf (RAVE) trial tested the efficacy and safety of VEGF blockade in eyes with severe CRVO. As natural history data imply that many preproliferative CRVO eyes stabilize after approximately 270 days, the RAVE protocol was designed (after discussions with Hayreh) to test whether VEGF blockade for 9 months could alter the natural history of neovascular complications in these high-risk eyes. To determine the durability of this response, patients were monitored monthly for 3 months after 9 monthly intravitreal ranibizumab injections before 24 months of pro re nata retreatment for a study duration of 36 months. In the design of the RAVE trial, the greatest expectations were that VEGF blockade would probably not improve mean visual acuity, but might prevent the devastating neovascular complications seen in eyes with the most severe CRVO. Materials and Methods This study was a prospective, phase I/II, open-label, clinical trial (Food and Drug Administration Investigational New Drug #12246, http://clinicaltrials.gov/ show/NCT00406471). After obtaining approval of the study protocol and consent by the Patient Advocacy Council Institutional Review Board, 20 patients identified at the Retina Consultants of Houston in the Texas Medical Center were provided with informed consent documents and were enrolled. In the study by Avery et al,7 bevacizumab was associated with a rapid regression of retinal and iris neovascularization in patients with diabetes, with complete resolution of neovascularization of the iris in 82% eyes. We expected a comparably large effect size (d = 0.82). Thus, the required sample size to detect changes in the development of neovascularization with a 2-tailed analysis, a = 0.05, and a power 1-b = 0.80 would be N = 21. Inclusion and Exclusion Criteria Within 3 months of CRVO onset, eyes at high-risk of neovascular complications were identified prospectively. Patients were at least 18 years old. Only one eye was eligible for the study enrollment. To maximize sensitivity for such enrollment, inclusion criteria required that eyes meet at least 3 of the 4 high-risk criteria defined by Hayreh et al: best-corrected visual acuity (BCVA) being #20/200, loss of the 1-2e isopter on Goldmann visual
1730 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
field, RAPD being $0.9 log units determined with neutral density filters,8,9 and electroretinogram (ERG) B-wave reduction to #60% of the corresponding A-wave.1,6 The ocular exclusion criteria included previous laser photocoagulation, any previous intravitreal injection, previous vitrectomy, the presence of .3 clock hours of anterior chamber angle neovascularization by gonioscopy, inability to assess iris neovascularization, intraocular pressure .30 mmHg, previous intracapsular cataract extraction, significant diabetic retinopathy in the fellow eye (diabetic macular edema, proliferative diabetic retinopathy, or severe nonproliferative diabetic retinopathy), history of retinal detachment, intraocular surgery within 90 days before the onset of the CRVO, significant cardiovascular disease or cancer that would limit follow-up visits or completion of the 36 months study, or previous radiation treatment to the head or neck. Protocol Visits and Testing Procedures On enrollment, patients received intravitreal injections of ranibizumab (Genentech, Inc, South San Francisco, CA) monthly (±7 days) for 9 treatments from Day 0 to Month 8. Although it was expected that these eyes would probably need the highest possible dose of anti-VEGF agent available, a dose escalation was designed at the request of the Food and Drug Administration regulatory authorities. At baseline, patients were sequentially enrolled based on study entry to receive 1 of 3 ranibizumab doses: 0.3 mg, 0.5 mg, or 1.0 mg. Entering Patients 1–5 and 11–15 were assigned to the 0.5 mg/0.05 mL group, Patients 6–10 were assigned to the 0.3 mg/0.05 mL group, and Patients 16–20 were assigned to the 1.0 mg/0.1 mL group. At all visits, subjects underwent Early Treatment Diabetic Retinopathy Study (ETDRS) refracted BCVA testing at 4 m, RAPD testing by swinging flashlight test with neutral density filters to balance, Goldman applanation tonometry, undilated slit-lamp biomicroscopy including gonioscopy specifically assessing the presence or absence of anterior segment neovascularization, dilated binocular indirect ophthalmoscopy, and spectral domain optical coherence tomography (OCT) (Spectralis; Heidelberg Engineering, Heidelberg, Germany). Goldmann visual field, color fundus photographs, fluorescein angiograms, and ERG were performed every 6 months. Wide-field fundus fluorescein angiography was performed with the Heidelberg Spectralis scanning laser ophthalmoscope (Heidelberg Engineering) using a Staurenghi contact lens (Ocular Staurenghi 230 SLO Retina Lens; Ocular Instruments Inc, Bellevue, WA). At Months 9 and 10, patients were examined clinically but not treated with intravitreal injections.
2014 VOLUME 34 NUMBER 9
Beginning at Month 12 (after a 3-month observation period), patients were examined monthly until the 36-month trial endpoint. Pro re nata retreatments were performed between Months 12 and 36 on evidence of cystoid macular edema (identified as any intraretinal cysts or subretinal fluid seen on OCT), an increase in central macular thickness (CMT) $100 mm (by OCT compared with the lowest previous CMT), or for any posterior or anterior segment neovascularization. After completion of the 36-month RAVE trial, patients were followed clinically to capture as many neovascular conversions as possible; these events were recorded. Eyes with iris neovascularization .3 clock hours, angle neovascularization, or posterior segment neovascularization were treated with panretinal photocoagulation in addition to intravitreal ranibizumab. Procedure Details of Intravitreal Injections Sterile surgical technique was followed for every intravitreal injection. Patients self-administered topical antimicrobials 4 times daily for 3 days before the treatment. After topical anesthesia, the periocular skin, eyelid, and eyelashes were disinfected with 10% povidone–iodine swabs and 5% povidone–iodine ophthalmic solution was applied to the ocular surface. Before intravitreal injection, anterior chamber paracentesis of 0.05 mL was performed to prophylax against intraocular pressure rise. After intravitreal injection, counting fingers was tested to confirm central retinal artery perfusion. When indicated, panretinal photocoagulation was applied 360° to the mid-peripheral retina according to the standard practice. Endpoints and Statistical Analysis The major clinical endpoints included change in the mean BCVA ETDRS letters, the percentage of patients who experienced a gain or loss of $15 letters ETDRS BCVA, mean change CMT, the proportion of patients with neovascular complications requiring panretinal photocoagulation, and the incidence and severity of ocular and nonocular adverse events. Statistical comparisons were performed with Student’s t-tests, paired Student’s t-tests, chi-square, and McNemar’s test using SAS 9.1.3 (SAS, Inc, Cary, NC) where appropriate. Results Twenty patients were enrolled in the RAVE trial. Baseline demographics and clinical findings were similar among patients. The mean age was 65 years (range, 44–83 years), 13 were male (65%), 6 (30%) had the diagnosis of diabetes mellitus at screening, and
THE RAVE TRIAL BROWN ET AL Table 1. Patient Disposition
Randomized Completed Exited Death Significant Adverse Event Noncompliant
Total
0.3 mg
0.5 mg
1.0 mg
20 13 7 2 4
5 3 2 0 2
10 7 3 2 0
5 3 2 0 2
1
0
1
0
16 (80%) had a diagnosis of hypertension. All patients met $3 of Hayreh’s 4 high-risk CRVO criteria and 11 (55%) met all 4. Baseline mean BCVA was 15 ETDRS letters (20/502 Snellen equivalent, range, 0–37 ETDRS letters). Nineteen patients (95%) had loss of the entire 1-2e isopter on Goldmann visual field testing. All binocular patients had a RAPD of $0.9 log units. Thirteen patients (65%) had a reduction of their ERG B-wave to ,60% of their A-wave. The mean CMT by spectral domain OCT was 485 mm (range, 168–893 mm). Eighteen (90%), 17 (85%), 15 (75%), and 13 (65%) patients completed the 9, 12, 24, and 36month trial time points, respectively (Table 1). The mean and median follow-up intervals during the RAVE trial were 30 and 35 months, respectively (range, 10–36 months). Reasons for not completing the trial included: death (2 patients at Week 1 and Month 11), medical indications prohibiting study follow-up visits (4 patients withdrew at Months 1, 20, 24, and 30), and noncompliance with loss of follow-up (1 patient after Month 20); corresponding data for these patients are included until the date of
1731
study withdrawal. To capture as many patients who developed neovascularization, patients were followed clinically after the completion of RAVE. The mean and median total follow-up lengths, including the 36 months RAVE trial and clinically beyond, were 57 and 59 months, respectively (range, 8–89 months). During the first 9 months of the trial, monthly treatment was mandatory and the mean number of treatments administered was 9 (range, 8–9). The mean number of intravitreal treatments administered through Months 24 and 36 was 14.1 (range, 9–20) and 17.2 (range, 9–32), respectively. There were no significant differences between the number of treatments administered to patients within the different dosage groups, although clinically there did seem to be more CMT fluctuation in the 0.3 mg ranibizumab–dosed patients compared with the 0.5 mg– dosed patients. Because of the few patients in each dosage arm and the observation that there were no significant differences in visual outcomes between the different dosage groups, results from the different dosage groups are pooled.
Visual Outcomes Mean ETDRS BCVA letters gained were +21.1 (range, −4 to +69) at 9 months, +14.1 (range, −21 to +63) at 12 months, +13.4 (range, −23 to +40) at 24 months, and +21.4 (range, −23 to +74) at 36 months (Figure 1). Patients gaining $15 ETDRS letters were 10 (63%), 7 (41%), 5 (50%), and 5 (56%) at 9, 12, 24, and 36 months, respectively. Patients losing $15 ETDRS letters were 0, 1 (6%), 1 (10%), and 1 (11%)
Fig. 1. Mean change in the bestcorrected visual acuity (ETDRS letters) over 36 months. Three phases are depicted: patients received 9 monthly doses of ranibizumab followed by a 3-month observation phase, then followed by a 24-month pro re nata phase. During the pro re nata phase, all patients underwent monthly visits and pro re nata retreatment with intravitreal ranibizumab.
1732 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
2014 VOLUME 34 NUMBER 9
Fig. 2. Change in the central retinal subfield thickness (spectral domain optical coherence tomography) over 36 months. Three phases are depicted: patients received 9 monthly doses of ranibizumab, followed by a 3-month observation phase, then followed by a 24-month pro re nata phase. During the pro re nata phase, all patients underwent monthly visits and pro re nata retreatment with intravitreal ranibizumab.
at 9, 12, 24, and 36 months, respectively. Seven eyes (39%) had ultimate visual acuity worse than 20/400. Anatomical Outcomes Initial anatomical improvement by OCT with dramatic reduction of intraretinal and subretinal fluid after ranibizumab treatment was similar among all patients (Figure 2). Mean CMT improved −294 mm (range, −47 to −652 mm) from baseline by Month 9. Subsequently, after 3 months of observation, the mean CMT increased +203 mm (range, −27 to +716 mm) at Month 12. Eight patients (44%) had recurrent macular
edema of .50 mm (range, 62–716 mm) during the observation period. On initiation of pro re nata retreatment, the mean CMT then improved −163 mm (range, −636 to −602 mm) and −191 mm (range, −623 to −58 mm) at Months 24 and 36 compared with Month 12. All patients demonstrated extensive retinal nonperfusion on fundus fluorescein angiography, consistent with the associated clinical findings. Development of Neovascular Complications During the course of follow-up, 9 patients developed neovascular complications (Figures 3 and 4). With
Fig. 3. Illustrative clinical case: 63-year-old man with a history of hypertension and severe CRVO with cystoid macular edema in his right eye. (A) Optical coherence tomography (above) and fluorescein angiography (below) demonstrate severe cystoid macular edema with blockage secondary to intraretinal hemorrhages. After monthly ranibizumab intravitreal injections for 9 months, (B) optical coherence tomography (above) and fluorescein angiogram (below) demonstrate the resolution of cystoid macular edema and improved intraretinal hemorrhages with areas of capillary nonperfusion. Clinical observation led to no cystoid macular edema recurrence, but development of anterior segment neovascularization at Month 12 (C), 3 months after the last intravitreal ranibizumab injection.
1733
THE RAVE TRIAL BROWN ET AL
Fig. 4. Illustrative clinical case: 49-year-old man with severe CRVO in his right eye. After 9 monthly intravitreal ranibizumab injections, 3 months of observation, and 1 pro re nata retreatment for recurrent cystoid macular edema, clinical examination and fluorescein angiography at Month 13 (A) demonstrates development of posterior segment neovascularization with neovascularization of the optic nerve head and the retinal periphery. After application of pan-retinal photocoagulation, neovascularization has resolved (B).
exclusion of the 2 patients who completed ,1 month of the trial, this represents 50% (9/18) of enrolled eyes; of these, 6 (33%) developed posterior segment neovascularization, 5 (28%) developed anterior segment neovascularization. Two eyes (11%) developed both posterior and anterior segment neovascularization. These were first diagnosed after a mean of 24 months follow-up (range, 3–44 months), with 2 patients developing neovascularization after completion of the 36-month trial endpoint (at Months 42 and 44 after study enrollment) (Figure 5). Patients who developed neovascularization subsequently underwent panretinal photocoagulation. Adverse Events Ocular and systemic adverse events are reported in Table 2. There were no cases of endophthalmitis, traumatic cataract, or intraocular inflammation. Two deaths during the trial were related to a myocardial infarction at Week 1 and a cerebral hemorrhage after a fall at Month 11. Medical reasons for study with-
drawal included antiphospholipid antibody syndrome, end stage renal disease, pneumonia with congestive heart failure, and congestive heart failure.
Discussion The RAVE trial was designed based on the natural history studies of Hayreh. The initial goals were to determine whether nine continuous injections of a potent anti-VEGF agent would eliminate retinal edema and protect the photoreceptors while allowing the clot in the central retinal vein to canalize and/or venous collaterals to form. It was not expected that significant visual acuity gains would occur because of concomitant retinal vasculature compromise. The visual acuity gains in the RAVE trial were consistent and directly correlated to the ability of ranibizumab to deturgesce the retina. When ranibizumab injections were withheld (Months 9–11), approximately half of patients had recurrent edema with subsequent loss of the initial visual acuity gains. Once
Fig. 5. Cumulative development of ocular neovascularization during clinical follow-up. A. Development of any anterior or posterior segment neovascularization in the RAVE trial. B. Reported natural history of the development of neovascularization after the onset of ischemic CRVO. Reproduced from Hayreh SS and Zimmerman BM. Ocular neovascularization associated with central and hemi-CRVO. Retina 2012;32:1553–1565.
1734 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES Table 2. Ocular and Systematic Adverse Events Ocular Adverse Events Total ocular adverse events Patients with ocular adverse events Any ocular neovascularization Posterior neovascularization Anterior neovascularization Neovascular glaucoma Subretinal hemorrhage Vitreous hemorrhage Increase in nuclear sclerosis Increase in posterior subcapsular cataract Posterior vitreous detachment Ocular serious adverse events Systemic Serious Adverse Events Total serious adverse events Patients with serious adverse events Death Myocardial infarction Congestive heart failure Arrhythmia Pneumonia
2014 VOLUME 34 NUMBER 9
Identification of Preproliferative CRVO Eyes 30 14 9 6 5 0 5 5 4 2 5 0 7 5 2 1 2 1 1
ranibizumab therapy was reinstated (on a pro re nata basis), anatomical and visual acuity improvements were restored in most eyes, and overall, patients did very well with close monitoring and pro re nata retreatments. An inherent weakness of this study is the lack of a control arm. However, the identification of “pre-proliferative” or early proliferative CRVO (those with presumably enough vascular drop-out to lead to neovascular complications with the stringent criteria) was used to identify eyes at highest risk for neovascular complications. Hayreh’s natural history studies of eyes meeting his functional criteria (and used as the entry criteria for this prospective interventional trial) demonstrated that 93% of eyes ended up with visual acuity worse than 20/400, approximately 2 of 3 of these eyes develop ocular neovascularization and 50% of untreated preproliferative eyes develop neovascular glaucoma usually in the first 8 months. In this study, only 39% of eyes had ultimate visual acuity worse than 20/400. Despite demonstrating this significant visual acuity benefit compared with the historical natural course of “ischemic CRVO”, RAVE essentially shows that this risk of neovascular complications is not ameliorated by VEGF blockade alone, but is only delayed in many patients. The fact that anti-VEGF therapy extends “90-day glaucoma” to “two-year-or-more glaucoma” in high-risk eyes makes patient education very important to avoid severe complications if the patient becomes noncompliant to monitoring or therapy. Although not the focus of this trial, we believe prophylactic pan-retinal photocoagulation should be considered in cases of questionable patient compliance.
The RAVE trial demonstrates that despite impressive visual acuity gains (71% 3-line gains at 9 months with monthly therapy), the preproliferative eyes in the RAVE trial still demonstrated at least a 50% incidence of neovascular complications with continued long-term follow-up. These paradoxical findings highlight the importance of identifying high-risk eyes before initiating treatment of CRVO with anti-VEGF agents. These data also reinforces the teachings of Hayreh that highrisk eyes can often be identified by clinical testing. The CVOS used a definition of $10 disk areas of nonperfusion on 90–120° fundus montages to identify eyes that were thought to be at risk for developing ocular neovascularization.10 Although 44% of the untreated eyes using this CVOS criteria (classified as “ischemic”) developed ocular neovascularization over 3 years follow-up,2 it is of note that of the entire CVOS study 48% (sensitivity of 52%) of eyes that developed iris or angle neovascularization were originally classified as “non-ischemic” or “perfused”.11 Hayreh et al’s1,6 multifactorial functional criteria seem to be more specific (as with follow-up 70% of eyes meeting these criteria develop neovascularization of the iris and 30% develop posterior neovascularization). Importantly, Hayreh’s classification was particularly sensitive with only 2.8% of “non-ischemic” patients developing neovascularization over follow-up. The most sensitive single test, RAPD, is actually the easiest and one of the most cost-effective tests in ophthalmology.8 While we quantified the magnitude of the RAPD by neutralizing the RAPD with the fellow eye using neutral density filters, we believe the “Hayreh” sign (a large RAPD in a patient with CRVO) should be elicited in any patient being treated for CRVO, particularly in those presenting with poor visual acuity. In patients with only one eye, the use of the ERG criteria, the loss of 1-2e isopter on Goldmann visual fields, or wide-field angiography looking for extensive capillary nonperfusion should be used. If these high-risk eyes are not identified, the improvements in visual acuity and clinical appearance with anti-VEGF injections can lull the physician and the patient into a false sense of security and decrease vigilance in the detection of neovascular complications. Central retinal vein occlusion occurs when a thrombus forms in the central retinal vein of the optic nerve, which drains the retinal circulation. This occlusion of the normal venous outflow of the eye increases venous pressure to a variable degree depending on where the actual occlusion occurs. It was classically assumed that increased venous pressure or “venous stasis” caused macular edema and
1735
THE RAVE TRIAL BROWN ET AL
resultant decreased visual acuity. However, as intravitreal anti-VEGF rapidly eliminates retinal edema in all CRVO, the mechanism for macular edema is probably not a direct result of the ocular intravenous hydrostatic pressure but is secondary to vasoproliferative cytokines released from ischemic retina because of decreased retinal perfusion. Thus, the classification of “ischemic” versus “non-ischemic” CRVO is a misnomer because it is likely that all CRVO eyes with macular edema are ischemic to some extent with resultant VEGF production. The mere delay of the neovascular complications in high-risk eyes illustrates that anti-VEGF therapy only treats the secondary cytokine release and does not treat the underlying blockage of flow in the central retinal vein. As there is not a current way to image the actual blockage in the CRVO, it is possible that anti-VEGF therapy might even delay collateral channel formation or clot resolution. In conclusion, the results seen with intravitreal ranibizumab in preproliferative eyes in the RAVE trial demonstrate that it is possible to improve anatomy and vision even in eyes that would have been excluded from previous CRVO trials. The most important finding, however, was that the neovascular complications in these high-risk “pre-proliferative” eyes are only delayed by VEGF blockade. Key words: RAVE, central retinal vein occlusion, ranibizumab, rubeosis, anti-VEGF, neovascularization, ischemic, macular edema, CRVO, ranibizumab. Acknowledgments The authors thank Sohan S. Hayreh, MD, FRCS, for his lifetime of work on CRVO and his advice on the development of this interventional protocol.
References 1. Hayreh SS, Klugman MR, Beri M, et al. Differentiation of ischemic from non-ischemic central retinal vein occlusion during the early acute phase. Graefes Arch Clin Exp Ophthalmol 1990;228:201–217. 2. The Central Vein Occlusion Study Group N Report. A randomized clinical trial of early panretinal photocoagulation for ischemic central vein occlusion. Ophthalmology 1995;102: 1434–1444. 3. Brown DM, Campochiaro PA, Singh RP, et al; CRUISE Investigators. Ranibizumab for macular edema following central retinal vein occlusion: six-month primary end point results of a phase III study. Ophthalmology 2010;117:1124–1133. 4. Campochiaro PA, Brown DM, Awh CC, et al. Sustained benefits from ranibizumab for macular edema following central retinal vein occlusion: twelve-month outcomes of a phase III study. Ophthalmology 2011;118:2041–2049. 5. Boyer D, Heier J, Brown DM, et al. Vascular endothelial growth factor trap-eye for macular edema secondary to central retinal vein occlusion: six-month results of the phase 3 COPERNICUS study. Ophthalmology 2012;119: 1024–1032. 6. Hayreh SS, Rojas P, Podhajsky P, et al. Ocular neovascularization with retinal vascular occlusion-III. Incidence of ocular neovascularization with retinal vein occlusion. Ophthalmology 1983;90:488–506. 7. Avery RL, Pearlman J, Pieramici DJ, et al. Intravitreal bevacizumab (Avastin) in the treatment of proliferative diabetic retinopathy. Ophthalmology 2006;113:1695. 8. Servais GE, Thompson HS, Hayreh SS. Relative afferent pupillary defect in central retinal vein occlusion. Ophthalmology 1986;93:301–303. 9. Thompson HS, Corbett JJ, Cox TA. How to measure the relative afferent pupillary defect. Surv Ophthalmol 1981;26:39–42. 10. Central Vein Occlusion Study Group. Central vein occlusion study of photocoagulation. Manual of operations. Online J Curr Clin Trials 1993;Doc No 92. http://www.ncbi.nlm.nih.gov/ pubmed/7508321. 11. The Central Vein Occlusion Study Group. Natural history and clinical management of central retinal vein occlusion. Arch Ophthalmol 1997;115:486–491.
C
hemorrhage or anterior rubeosis with secondary neovascular glaucoma. Previous natural history and interventional studies have attempted to differentiate CRVO eyes that have a higher risk of neovascular complications. Hayreh et al1 identified criteria that predicted and differentiated CRVO eyes with a propensity for devastating neovascular complications (“hemorraghic retinopathy” or
entral retinal vein occlusion (CRVO) is a common vascular event that presents acutely with intraretinal hemorrhages throughout the retinal fundus and marked macular edema associated with decreased visual acuity. Although some eyes with CRVO spontaneously resolve, many have persistent macular edema or develop proliferative complications, including posterior neovascularization with vitreous
1728
1729
THE RAVE TRIAL BROWN ET AL
“ischemic CRVO” eyes) from eyes with a more benign CRVO course classified as “venous stasis retinopathy” or “non-ischemic CRVO.” The Central Retinal Vein Occlusion Study (CVOS) used an angiographic criteria of 10 disk areas or more of capillary nonperfusion on fundus fluorescein angiography to identify “ischemic CRVO,”2 whereas Hayreh identified functional tests (visual acuity, visual fields, relative afferent pupillary defect [RAPD], and electroretinography [ERG]) that demonstrated a high sensitivity and specificity for predicting “ischemic” CRVO. Intravitreal anti-vascular endothelial growth factor (VEGF) treatments have provided an extremely effective therapy for macular edema secondary to CRVO with reduction and/or elimination of macular edema and improved visual acuity with monthly injections.3–5 The phase III ranibizumab for CRVO trial (CRUISE) specifically excluded patients with a RAPD and those with visual acuity worse than 20/320. Only 2 of 334 eyes in CRUISE with adequately gradable fundus fluorescein angiography had .10 disk area of nonperfusion demonstrating how effective these 2 functional criteria were at excluding patients who would have been previously identified as “ischemic” by the CVOS. Although almost all of the eyes in the CRUISE trial would have historically been classified as “non-ischemic” by the CVOS or Hayreh criteria, the universal anatomical improvements seen with VEGF blockade imply that the thrombus in the central retinal vein must lead to variable amounts of retinal ischemia in all CRVO patients leading to the production of VEGF with subsequent macular edema. Because these trials implicate ischemia and VEGF in all patients with CRVO with macular edema, we use the term “pre-proliferative CRVO” to describe CRVO eyes with a high risk of neovascular complications. Hayreh et al6 demonstrated that in high-risk eyes (what in this report is referred to as preproliferative CRVO eyes), neovascular complications occur in approximately 2 of 3 eyes within 3 months without From the Retina Consultants of Houston, Houston Methodist Hospital, Houston, Texas. The authors received a research grant for this trial from Genentech. The funding organization had no role in the design or conduct of this research. Clinical Trial Registration—http://www.clinicaltrials.gov. Identifier: NCT00406471, NCT01225146. DMB: Consultant for Alcon, Allergan, Bayer, Genentech, Regeneron. Lecturer for Bayer and Roche. Research Support from Alcon, Allergan, Genentech, Regeneron. CCW: Consultant for Alcon, Allergan, Genentech. Lecturer for Genentech and Regeneron. Research Support from Alcon, Allergan, Genentech, Regeneron. TPW: Research Support from Alcon, Allergan, Genentech, Regeneron. AFM: None. DEC: None. KLS: None. Reprint requests: David M. Brown, MD, The Methodist Hospital, 6560 Fannin Street, Suite 750, Houston, TX 77030; e-mail: [email protected]
treatment. However, approximately 1 of 3 of these eyes avoids neovascular glaucoma and stabilizes within approximately 270 days presumably secondary to collateral formation or recanalization of the thrombus. The Rubeosis Anti-VEgf (RAVE) trial tested the efficacy and safety of VEGF blockade in eyes with severe CRVO. As natural history data imply that many preproliferative CRVO eyes stabilize after approximately 270 days, the RAVE protocol was designed (after discussions with Hayreh) to test whether VEGF blockade for 9 months could alter the natural history of neovascular complications in these high-risk eyes. To determine the durability of this response, patients were monitored monthly for 3 months after 9 monthly intravitreal ranibizumab injections before 24 months of pro re nata retreatment for a study duration of 36 months. In the design of the RAVE trial, the greatest expectations were that VEGF blockade would probably not improve mean visual acuity, but might prevent the devastating neovascular complications seen in eyes with the most severe CRVO. Materials and Methods This study was a prospective, phase I/II, open-label, clinical trial (Food and Drug Administration Investigational New Drug #12246, http://clinicaltrials.gov/ show/NCT00406471). After obtaining approval of the study protocol and consent by the Patient Advocacy Council Institutional Review Board, 20 patients identified at the Retina Consultants of Houston in the Texas Medical Center were provided with informed consent documents and were enrolled. In the study by Avery et al,7 bevacizumab was associated with a rapid regression of retinal and iris neovascularization in patients with diabetes, with complete resolution of neovascularization of the iris in 82% eyes. We expected a comparably large effect size (d = 0.82). Thus, the required sample size to detect changes in the development of neovascularization with a 2-tailed analysis, a = 0.05, and a power 1-b = 0.80 would be N = 21. Inclusion and Exclusion Criteria Within 3 months of CRVO onset, eyes at high-risk of neovascular complications were identified prospectively. Patients were at least 18 years old. Only one eye was eligible for the study enrollment. To maximize sensitivity for such enrollment, inclusion criteria required that eyes meet at least 3 of the 4 high-risk criteria defined by Hayreh et al: best-corrected visual acuity (BCVA) being #20/200, loss of the 1-2e isopter on Goldmann visual
1730 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
field, RAPD being $0.9 log units determined with neutral density filters,8,9 and electroretinogram (ERG) B-wave reduction to #60% of the corresponding A-wave.1,6 The ocular exclusion criteria included previous laser photocoagulation, any previous intravitreal injection, previous vitrectomy, the presence of .3 clock hours of anterior chamber angle neovascularization by gonioscopy, inability to assess iris neovascularization, intraocular pressure .30 mmHg, previous intracapsular cataract extraction, significant diabetic retinopathy in the fellow eye (diabetic macular edema, proliferative diabetic retinopathy, or severe nonproliferative diabetic retinopathy), history of retinal detachment, intraocular surgery within 90 days before the onset of the CRVO, significant cardiovascular disease or cancer that would limit follow-up visits or completion of the 36 months study, or previous radiation treatment to the head or neck. Protocol Visits and Testing Procedures On enrollment, patients received intravitreal injections of ranibizumab (Genentech, Inc, South San Francisco, CA) monthly (±7 days) for 9 treatments from Day 0 to Month 8. Although it was expected that these eyes would probably need the highest possible dose of anti-VEGF agent available, a dose escalation was designed at the request of the Food and Drug Administration regulatory authorities. At baseline, patients were sequentially enrolled based on study entry to receive 1 of 3 ranibizumab doses: 0.3 mg, 0.5 mg, or 1.0 mg. Entering Patients 1–5 and 11–15 were assigned to the 0.5 mg/0.05 mL group, Patients 6–10 were assigned to the 0.3 mg/0.05 mL group, and Patients 16–20 were assigned to the 1.0 mg/0.1 mL group. At all visits, subjects underwent Early Treatment Diabetic Retinopathy Study (ETDRS) refracted BCVA testing at 4 m, RAPD testing by swinging flashlight test with neutral density filters to balance, Goldman applanation tonometry, undilated slit-lamp biomicroscopy including gonioscopy specifically assessing the presence or absence of anterior segment neovascularization, dilated binocular indirect ophthalmoscopy, and spectral domain optical coherence tomography (OCT) (Spectralis; Heidelberg Engineering, Heidelberg, Germany). Goldmann visual field, color fundus photographs, fluorescein angiograms, and ERG were performed every 6 months. Wide-field fundus fluorescein angiography was performed with the Heidelberg Spectralis scanning laser ophthalmoscope (Heidelberg Engineering) using a Staurenghi contact lens (Ocular Staurenghi 230 SLO Retina Lens; Ocular Instruments Inc, Bellevue, WA). At Months 9 and 10, patients were examined clinically but not treated with intravitreal injections.
2014 VOLUME 34 NUMBER 9
Beginning at Month 12 (after a 3-month observation period), patients were examined monthly until the 36-month trial endpoint. Pro re nata retreatments were performed between Months 12 and 36 on evidence of cystoid macular edema (identified as any intraretinal cysts or subretinal fluid seen on OCT), an increase in central macular thickness (CMT) $100 mm (by OCT compared with the lowest previous CMT), or for any posterior or anterior segment neovascularization. After completion of the 36-month RAVE trial, patients were followed clinically to capture as many neovascular conversions as possible; these events were recorded. Eyes with iris neovascularization .3 clock hours, angle neovascularization, or posterior segment neovascularization were treated with panretinal photocoagulation in addition to intravitreal ranibizumab. Procedure Details of Intravitreal Injections Sterile surgical technique was followed for every intravitreal injection. Patients self-administered topical antimicrobials 4 times daily for 3 days before the treatment. After topical anesthesia, the periocular skin, eyelid, and eyelashes were disinfected with 10% povidone–iodine swabs and 5% povidone–iodine ophthalmic solution was applied to the ocular surface. Before intravitreal injection, anterior chamber paracentesis of 0.05 mL was performed to prophylax against intraocular pressure rise. After intravitreal injection, counting fingers was tested to confirm central retinal artery perfusion. When indicated, panretinal photocoagulation was applied 360° to the mid-peripheral retina according to the standard practice. Endpoints and Statistical Analysis The major clinical endpoints included change in the mean BCVA ETDRS letters, the percentage of patients who experienced a gain or loss of $15 letters ETDRS BCVA, mean change CMT, the proportion of patients with neovascular complications requiring panretinal photocoagulation, and the incidence and severity of ocular and nonocular adverse events. Statistical comparisons were performed with Student’s t-tests, paired Student’s t-tests, chi-square, and McNemar’s test using SAS 9.1.3 (SAS, Inc, Cary, NC) where appropriate. Results Twenty patients were enrolled in the RAVE trial. Baseline demographics and clinical findings were similar among patients. The mean age was 65 years (range, 44–83 years), 13 were male (65%), 6 (30%) had the diagnosis of diabetes mellitus at screening, and
THE RAVE TRIAL BROWN ET AL Table 1. Patient Disposition
Randomized Completed Exited Death Significant Adverse Event Noncompliant
Total
0.3 mg
0.5 mg
1.0 mg
20 13 7 2 4
5 3 2 0 2
10 7 3 2 0
5 3 2 0 2
1
0
1
0
16 (80%) had a diagnosis of hypertension. All patients met $3 of Hayreh’s 4 high-risk CRVO criteria and 11 (55%) met all 4. Baseline mean BCVA was 15 ETDRS letters (20/502 Snellen equivalent, range, 0–37 ETDRS letters). Nineteen patients (95%) had loss of the entire 1-2e isopter on Goldmann visual field testing. All binocular patients had a RAPD of $0.9 log units. Thirteen patients (65%) had a reduction of their ERG B-wave to ,60% of their A-wave. The mean CMT by spectral domain OCT was 485 mm (range, 168–893 mm). Eighteen (90%), 17 (85%), 15 (75%), and 13 (65%) patients completed the 9, 12, 24, and 36month trial time points, respectively (Table 1). The mean and median follow-up intervals during the RAVE trial were 30 and 35 months, respectively (range, 10–36 months). Reasons for not completing the trial included: death (2 patients at Week 1 and Month 11), medical indications prohibiting study follow-up visits (4 patients withdrew at Months 1, 20, 24, and 30), and noncompliance with loss of follow-up (1 patient after Month 20); corresponding data for these patients are included until the date of
1731
study withdrawal. To capture as many patients who developed neovascularization, patients were followed clinically after the completion of RAVE. The mean and median total follow-up lengths, including the 36 months RAVE trial and clinically beyond, were 57 and 59 months, respectively (range, 8–89 months). During the first 9 months of the trial, monthly treatment was mandatory and the mean number of treatments administered was 9 (range, 8–9). The mean number of intravitreal treatments administered through Months 24 and 36 was 14.1 (range, 9–20) and 17.2 (range, 9–32), respectively. There were no significant differences between the number of treatments administered to patients within the different dosage groups, although clinically there did seem to be more CMT fluctuation in the 0.3 mg ranibizumab–dosed patients compared with the 0.5 mg– dosed patients. Because of the few patients in each dosage arm and the observation that there were no significant differences in visual outcomes between the different dosage groups, results from the different dosage groups are pooled.
Visual Outcomes Mean ETDRS BCVA letters gained were +21.1 (range, −4 to +69) at 9 months, +14.1 (range, −21 to +63) at 12 months, +13.4 (range, −23 to +40) at 24 months, and +21.4 (range, −23 to +74) at 36 months (Figure 1). Patients gaining $15 ETDRS letters were 10 (63%), 7 (41%), 5 (50%), and 5 (56%) at 9, 12, 24, and 36 months, respectively. Patients losing $15 ETDRS letters were 0, 1 (6%), 1 (10%), and 1 (11%)
Fig. 1. Mean change in the bestcorrected visual acuity (ETDRS letters) over 36 months. Three phases are depicted: patients received 9 monthly doses of ranibizumab followed by a 3-month observation phase, then followed by a 24-month pro re nata phase. During the pro re nata phase, all patients underwent monthly visits and pro re nata retreatment with intravitreal ranibizumab.
1732 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
2014 VOLUME 34 NUMBER 9
Fig. 2. Change in the central retinal subfield thickness (spectral domain optical coherence tomography) over 36 months. Three phases are depicted: patients received 9 monthly doses of ranibizumab, followed by a 3-month observation phase, then followed by a 24-month pro re nata phase. During the pro re nata phase, all patients underwent monthly visits and pro re nata retreatment with intravitreal ranibizumab.
at 9, 12, 24, and 36 months, respectively. Seven eyes (39%) had ultimate visual acuity worse than 20/400. Anatomical Outcomes Initial anatomical improvement by OCT with dramatic reduction of intraretinal and subretinal fluid after ranibizumab treatment was similar among all patients (Figure 2). Mean CMT improved −294 mm (range, −47 to −652 mm) from baseline by Month 9. Subsequently, after 3 months of observation, the mean CMT increased +203 mm (range, −27 to +716 mm) at Month 12. Eight patients (44%) had recurrent macular
edema of .50 mm (range, 62–716 mm) during the observation period. On initiation of pro re nata retreatment, the mean CMT then improved −163 mm (range, −636 to −602 mm) and −191 mm (range, −623 to −58 mm) at Months 24 and 36 compared with Month 12. All patients demonstrated extensive retinal nonperfusion on fundus fluorescein angiography, consistent with the associated clinical findings. Development of Neovascular Complications During the course of follow-up, 9 patients developed neovascular complications (Figures 3 and 4). With
Fig. 3. Illustrative clinical case: 63-year-old man with a history of hypertension and severe CRVO with cystoid macular edema in his right eye. (A) Optical coherence tomography (above) and fluorescein angiography (below) demonstrate severe cystoid macular edema with blockage secondary to intraretinal hemorrhages. After monthly ranibizumab intravitreal injections for 9 months, (B) optical coherence tomography (above) and fluorescein angiogram (below) demonstrate the resolution of cystoid macular edema and improved intraretinal hemorrhages with areas of capillary nonperfusion. Clinical observation led to no cystoid macular edema recurrence, but development of anterior segment neovascularization at Month 12 (C), 3 months after the last intravitreal ranibizumab injection.
1733
THE RAVE TRIAL BROWN ET AL
Fig. 4. Illustrative clinical case: 49-year-old man with severe CRVO in his right eye. After 9 monthly intravitreal ranibizumab injections, 3 months of observation, and 1 pro re nata retreatment for recurrent cystoid macular edema, clinical examination and fluorescein angiography at Month 13 (A) demonstrates development of posterior segment neovascularization with neovascularization of the optic nerve head and the retinal periphery. After application of pan-retinal photocoagulation, neovascularization has resolved (B).
exclusion of the 2 patients who completed ,1 month of the trial, this represents 50% (9/18) of enrolled eyes; of these, 6 (33%) developed posterior segment neovascularization, 5 (28%) developed anterior segment neovascularization. Two eyes (11%) developed both posterior and anterior segment neovascularization. These were first diagnosed after a mean of 24 months follow-up (range, 3–44 months), with 2 patients developing neovascularization after completion of the 36-month trial endpoint (at Months 42 and 44 after study enrollment) (Figure 5). Patients who developed neovascularization subsequently underwent panretinal photocoagulation. Adverse Events Ocular and systemic adverse events are reported in Table 2. There were no cases of endophthalmitis, traumatic cataract, or intraocular inflammation. Two deaths during the trial were related to a myocardial infarction at Week 1 and a cerebral hemorrhage after a fall at Month 11. Medical reasons for study with-
drawal included antiphospholipid antibody syndrome, end stage renal disease, pneumonia with congestive heart failure, and congestive heart failure.
Discussion The RAVE trial was designed based on the natural history studies of Hayreh. The initial goals were to determine whether nine continuous injections of a potent anti-VEGF agent would eliminate retinal edema and protect the photoreceptors while allowing the clot in the central retinal vein to canalize and/or venous collaterals to form. It was not expected that significant visual acuity gains would occur because of concomitant retinal vasculature compromise. The visual acuity gains in the RAVE trial were consistent and directly correlated to the ability of ranibizumab to deturgesce the retina. When ranibizumab injections were withheld (Months 9–11), approximately half of patients had recurrent edema with subsequent loss of the initial visual acuity gains. Once
Fig. 5. Cumulative development of ocular neovascularization during clinical follow-up. A. Development of any anterior or posterior segment neovascularization in the RAVE trial. B. Reported natural history of the development of neovascularization after the onset of ischemic CRVO. Reproduced from Hayreh SS and Zimmerman BM. Ocular neovascularization associated with central and hemi-CRVO. Retina 2012;32:1553–1565.
1734 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES Table 2. Ocular and Systematic Adverse Events Ocular Adverse Events Total ocular adverse events Patients with ocular adverse events Any ocular neovascularization Posterior neovascularization Anterior neovascularization Neovascular glaucoma Subretinal hemorrhage Vitreous hemorrhage Increase in nuclear sclerosis Increase in posterior subcapsular cataract Posterior vitreous detachment Ocular serious adverse events Systemic Serious Adverse Events Total serious adverse events Patients with serious adverse events Death Myocardial infarction Congestive heart failure Arrhythmia Pneumonia
2014 VOLUME 34 NUMBER 9
Identification of Preproliferative CRVO Eyes 30 14 9 6 5 0 5 5 4 2 5 0 7 5 2 1 2 1 1
ranibizumab therapy was reinstated (on a pro re nata basis), anatomical and visual acuity improvements were restored in most eyes, and overall, patients did very well with close monitoring and pro re nata retreatments. An inherent weakness of this study is the lack of a control arm. However, the identification of “pre-proliferative” or early proliferative CRVO (those with presumably enough vascular drop-out to lead to neovascular complications with the stringent criteria) was used to identify eyes at highest risk for neovascular complications. Hayreh’s natural history studies of eyes meeting his functional criteria (and used as the entry criteria for this prospective interventional trial) demonstrated that 93% of eyes ended up with visual acuity worse than 20/400, approximately 2 of 3 of these eyes develop ocular neovascularization and 50% of untreated preproliferative eyes develop neovascular glaucoma usually in the first 8 months. In this study, only 39% of eyes had ultimate visual acuity worse than 20/400. Despite demonstrating this significant visual acuity benefit compared with the historical natural course of “ischemic CRVO”, RAVE essentially shows that this risk of neovascular complications is not ameliorated by VEGF blockade alone, but is only delayed in many patients. The fact that anti-VEGF therapy extends “90-day glaucoma” to “two-year-or-more glaucoma” in high-risk eyes makes patient education very important to avoid severe complications if the patient becomes noncompliant to monitoring or therapy. Although not the focus of this trial, we believe prophylactic pan-retinal photocoagulation should be considered in cases of questionable patient compliance.
The RAVE trial demonstrates that despite impressive visual acuity gains (71% 3-line gains at 9 months with monthly therapy), the preproliferative eyes in the RAVE trial still demonstrated at least a 50% incidence of neovascular complications with continued long-term follow-up. These paradoxical findings highlight the importance of identifying high-risk eyes before initiating treatment of CRVO with anti-VEGF agents. These data also reinforces the teachings of Hayreh that highrisk eyes can often be identified by clinical testing. The CVOS used a definition of $10 disk areas of nonperfusion on 90–120° fundus montages to identify eyes that were thought to be at risk for developing ocular neovascularization.10 Although 44% of the untreated eyes using this CVOS criteria (classified as “ischemic”) developed ocular neovascularization over 3 years follow-up,2 it is of note that of the entire CVOS study 48% (sensitivity of 52%) of eyes that developed iris or angle neovascularization were originally classified as “non-ischemic” or “perfused”.11 Hayreh et al’s1,6 multifactorial functional criteria seem to be more specific (as with follow-up 70% of eyes meeting these criteria develop neovascularization of the iris and 30% develop posterior neovascularization). Importantly, Hayreh’s classification was particularly sensitive with only 2.8% of “non-ischemic” patients developing neovascularization over follow-up. The most sensitive single test, RAPD, is actually the easiest and one of the most cost-effective tests in ophthalmology.8 While we quantified the magnitude of the RAPD by neutralizing the RAPD with the fellow eye using neutral density filters, we believe the “Hayreh” sign (a large RAPD in a patient with CRVO) should be elicited in any patient being treated for CRVO, particularly in those presenting with poor visual acuity. In patients with only one eye, the use of the ERG criteria, the loss of 1-2e isopter on Goldmann visual fields, or wide-field angiography looking for extensive capillary nonperfusion should be used. If these high-risk eyes are not identified, the improvements in visual acuity and clinical appearance with anti-VEGF injections can lull the physician and the patient into a false sense of security and decrease vigilance in the detection of neovascular complications. Central retinal vein occlusion occurs when a thrombus forms in the central retinal vein of the optic nerve, which drains the retinal circulation. This occlusion of the normal venous outflow of the eye increases venous pressure to a variable degree depending on where the actual occlusion occurs. It was classically assumed that increased venous pressure or “venous stasis” caused macular edema and
1735
THE RAVE TRIAL BROWN ET AL
resultant decreased visual acuity. However, as intravitreal anti-VEGF rapidly eliminates retinal edema in all CRVO, the mechanism for macular edema is probably not a direct result of the ocular intravenous hydrostatic pressure but is secondary to vasoproliferative cytokines released from ischemic retina because of decreased retinal perfusion. Thus, the classification of “ischemic” versus “non-ischemic” CRVO is a misnomer because it is likely that all CRVO eyes with macular edema are ischemic to some extent with resultant VEGF production. The mere delay of the neovascular complications in high-risk eyes illustrates that anti-VEGF therapy only treats the secondary cytokine release and does not treat the underlying blockage of flow in the central retinal vein. As there is not a current way to image the actual blockage in the CRVO, it is possible that anti-VEGF therapy might even delay collateral channel formation or clot resolution. In conclusion, the results seen with intravitreal ranibizumab in preproliferative eyes in the RAVE trial demonstrate that it is possible to improve anatomy and vision even in eyes that would have been excluded from previous CRVO trials. The most important finding, however, was that the neovascular complications in these high-risk “pre-proliferative” eyes are only delayed by VEGF blockade. Key words: RAVE, central retinal vein occlusion, ranibizumab, rubeosis, anti-VEGF, neovascularization, ischemic, macular edema, CRVO, ranibizumab. Acknowledgments The authors thank Sohan S. Hayreh, MD, FRCS, for his lifetime of work on CRVO and his advice on the development of this interventional protocol.
References 1. Hayreh SS, Klugman MR, Beri M, et al. Differentiation of ischemic from non-ischemic central retinal vein occlusion during the early acute phase. Graefes Arch Clin Exp Ophthalmol 1990;228:201–217. 2. The Central Vein Occlusion Study Group N Report. A randomized clinical trial of early panretinal photocoagulation for ischemic central vein occlusion. Ophthalmology 1995;102: 1434–1444. 3. Brown DM, Campochiaro PA, Singh RP, et al; CRUISE Investigators. Ranibizumab for macular edema following central retinal vein occlusion: six-month primary end point results of a phase III study. Ophthalmology 2010;117:1124–1133. 4. Campochiaro PA, Brown DM, Awh CC, et al. Sustained benefits from ranibizumab for macular edema following central retinal vein occlusion: twelve-month outcomes of a phase III study. Ophthalmology 2011;118:2041–2049. 5. Boyer D, Heier J, Brown DM, et al. Vascular endothelial growth factor trap-eye for macular edema secondary to central retinal vein occlusion: six-month results of the phase 3 COPERNICUS study. Ophthalmology 2012;119: 1024–1032. 6. Hayreh SS, Rojas P, Podhajsky P, et al. Ocular neovascularization with retinal vascular occlusion-III. Incidence of ocular neovascularization with retinal vein occlusion. Ophthalmology 1983;90:488–506. 7. Avery RL, Pearlman J, Pieramici DJ, et al. Intravitreal bevacizumab (Avastin) in the treatment of proliferative diabetic retinopathy. Ophthalmology 2006;113:1695. 8. Servais GE, Thompson HS, Hayreh SS. Relative afferent pupillary defect in central retinal vein occlusion. Ophthalmology 1986;93:301–303. 9. Thompson HS, Corbett JJ, Cox TA. How to measure the relative afferent pupillary defect. Surv Ophthalmol 1981;26:39–42. 10. Central Vein Occlusion Study Group. Central vein occlusion study of photocoagulation. Manual of operations. Online J Curr Clin Trials 1993;Doc No 92. http://www.ncbi.nlm.nih.gov/ pubmed/7508321. 11. The Central Vein Occlusion Study Group. Natural history and clinical management of central retinal vein occlusion. Arch Ophthalmol 1997;115:486–491.
