The Diabetic Retinopathy Clinical Research Network Randomized Clinical

The Diabetic Retinopathy Clinical Research Network Randomized Clinical Trial Evaluating Intravitreal Ranibizumab or Intravitreal Saline for Vitreous Hemorrhage from Proliferative Diabetic Retinopathy Supported through a cooperative agreement from the National Eye Institute and the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services EY 14231, EY 14229, EY 018817 1

Background Vitreous Hemorrhage and Vitrectomy In at least 5% of cases, proliferative diabetic retinopathy (PDR) can lead to vitreous hemorrhage even after panretinal photocoagulation (PRP) is initiated which • can substantially affect vision and • may preclude performing complete PRP. Current treatment for non-clearing vitreous hemorrhage is vitrectomy • Surgical complications • Recovery time • Some patients may fear surgery 2

Background Use of anti-VEGF for vitreous hemorrhage Intravitreal anti-vascular endothelial growth factor (anti-VEGF) is a potential alternative to vitrectomy • VEGF is shown to be major causative factor in PDR • While hemorrhage absorbs, anti-VEGF drug may cause temporary or permanent regression of NVD/NVE, reducing the likelihood of additional hemorrhage until PRP is applied • Small case series suggest anti-VEGF causes short-term clearing of vitreous hemorrhage 3

Background Use of Saline for vitreous hemorrhage Unknown if effects seen with anti-VEGF are from fluid injection alone Prior study found that saline injected eyes were 5 times more likely to have visual improvement at 2 months (26% vs. 6%) and 3 times more likely at 3 months (36% vs. 11%) compared with eyes under observation alone 4

Study Objectives To determine if intravitreal injections of ranibizumab decrease the proportion of eyes in which vitrectomy is performed compared with saline injections in eyes presenting with vitreous hemorrhage from PDR. • Note: This trial was not a comparison of anti-VEGF with observation or sham injection; rather the trial was a comparison of intravitreal anti-VEGF with intravitreal saline injection To assess the efficacy and safety of anti-VEGF therapy as treatment for vitreous hemorrhage due to PDR. 5

Study Design, Enrollment, Follow-up Randomized Multicenter Double-Masked Trial At least one eye that met all of the following criteria: § Vitreous hemorrhage causing vision impairment, presumed to be from PDR, and precluding completion of PRP § Immediate vitrectomy is not required § Visual acuity is light perception or better § No prior anti-VEGF treatment for VH Primary Outcome: Proportion of eyes with vitrectomy by 16 weeks. 6

Study Enrollment 261 Eyes Randomized (61 Sites) Intravitreal Injection of 0. 5 ranibizumab N = 125 Intravitreal Injection 0. 9% sodium chloride N = 136 12 Week Visit Completion 95% Overall* (96% Ranibizumab Injection; 95% Saline Injection) * One death occurred prior to the 12 week visit. 7

Follow-up Schedule Phase 1 Phase 2 4 WK VISIT Treatment for VH at investigator discretion 8 WK VISIT 12 WK VISIT 26 WK Phone Call 16 WK Primary Outcome Time Point 52 WK VISIT 8

Follow-Up Treatment Follow-up injections performed at 4 and 8 weeks unless: • Vitreous hemorrhage had cleared enough to complete PRP or • Vitrectomy had been performed. All eyes were to be treated with complete PRP as soon as possible. 9

Guidelines for Vitrectomy Prior to 8 Week Study Visit Vitrectomy was not to be performed unless one of the following was present: • Traction retinal detachment involving or threatening the macula • Rhegmatogenous retinal detachment • Angle neovascularization • Progressive iris neovascularization • Neovascular glaucoma or ghost cell glaucoma resulting in increased IOP that could not be controlled medically 10

Guidelines for Vitrectomy 8 to 16 Weeks Vitrectomy could be performed (but was not required) if there was failure of substantial clearing of vitreous hemorrhage such that the study participant’s need for vitrectomy and its associated benefits outweighed its risks. Primary outcome time point was at 16 weeks to include cases where decision was made at 12 weeks to perform vitrectomy. 11

PRP Treatment PRP was to be initiated as soon as possible “Complete” PRP defined as • 500 micron size burns on the retina placed no further than 1 to 2 burn widths apart beginning ~3000 µm from macular center and extending to equator for 12 clock hours. If vitreous hemorrhage cleared and it was determined that “complete” PRP was already given, further treatment was not required. 12

10 -J u 10 n -J 10 ul -A 10 ug -S e 10 p -O 10 ct -N 10 ov -D 11 ec -J 11 an -F 11 eb -M 11 ar -A 11 pr -M 11 ay -J u 11 n -J 11 ul -A 11 ug -S e 11 p -O 11 ct -N 11 ov -D ec Recruitment Summary 300 250 200 Original goal Original Goal reached Recruitment Resumed 150 100 50 0 13

Baseline Subject Characteristics Ranibizumab N=125 52% 61 (50, 67) Saline N=136 51% 58 (49, 64) Race White African-American 54% 16% 51% 16% Hispanic or Latino Other Type 1 Diabetes Type 2 Diabetes Uncertain 26% 4% 17% 81% 2% 25% 8% 23% 73% 4% Women Age (yrs) (Median 25, 75 th percentile) 14

Baseline Subject Characteristics Duration of Diabetes - yrs Ranibizumab N=125 19 (12, 28) Saline N=136 21 (16, 27) (Median 25 th, 75 th percentile) Hemoglobin A 1 c - % 7. 7 (6. 7, 8. 7) 8. 0 (6. 9, 9. 3) Pre-existing cardiovascular condition 35% 44% Pre-existing hypertension 84% 86% (Median 25 th, 75 th percentile) 15

Baseline Study Eye Characteristics Ranibizumab N=125 Saline N=136 Prior PRP Prior Treatment for DME 50% 42% 57% 42% Prior treatment with anti. VEGF for DME 8% 13% E-ETDRS Visual Acuity Median (75 th, 25 th Percentile); Snellen Equivalent OCT Signal Strength = 0 > 0 34 (61, 0) 28 (59, 0) 20/200 20/320 (20/63, <20/800) 60% 40% 72% 28% 16

Baseline Study Eye Characteristics Ranibizumab N=125 Duration of Vitreous Hemorrhage <1 Month 53% 1 -3 Months 4 -6 Months >6 Months IOP - mm. Hg Median (25 th, 75 th percentile) 33% 6% 9% 15 (12, 17) Saline N=136 55% 29% 8% 8% 14 (12, 17) 17

Total Number of Intravitreal Injections In Eyes without vitrectomy or “complete PRP” prior to 8 weeks Number of Injections 0 1 2 3 4 Ranibizumab Injection Saline Injection N = 91 0 3 7 81 0 N = 103 0 7 14 81 1† † One subject was given masked intravitreal study drug at the 12 -week follow-up visit when a study injection was not scheduled per protocol. Since an injection was not scheduled, a drug number according to the randomized treatment was not obtained, and it is possible that the wrong randomized treatment was performed. The investigator did believe that an injection was medically 18 indicated for the study participant.

Total Number of Intravitreal Injections In Eyes with vitrectomy or “complete PRP” prior to 8 weeks Number of Injections 0 1 2 3 Ranibizumab Injection Saline Injection N = 33 0 19 13 1 N = 33 1¥ 20 9 3 ¥ After consenting and randomizing, 1 subject changed their mind and did not want to receive study drug – the participant remained in the study but was never injected. 19

Non-Protocol Study Eye Treatment for DME Prior to 16 weeks Ranibizumab Saline Injection N = 4 Number of Eyes with Non-Protocol Study Eye Treatment for DME 3 4 Intravitreal Bevacizumab (Avastin) 1* 1** Intravitreal Triamcinolone Acetonide 1* 0 Focal/Grid Laser Photocoagulation 2 3 Note: only focal/grid laser was permitted per protocol prior to 16 weeks *Same eye received bevacizumab and triamcinolone for DME after vitreous hemorrhage cleared and PRP was considered “complete” **DME treatment performed 1 week post-vitrectomy 20

Vitrectomy Primary Outcome 21

Cumulative Probability of Vitrectomy by 16 Weeks 60% Ranibizumab Saline Cumulative Probability 50% P = 0. 38 40% 30% 17% 20% 13% 0% -10% 0 4 8 12 16

Vitrectomy Status Over Time Vitrectomy performed up to 8 weeks Vitrectomy performed between 8 weeks and 16 weeks Vitrectomy performed by 16 weeks Ranibizumab Injection N = 125 Saline Injection N = 136 2% 2% 10% 14% 12% 16% 23

Cumulative Probability of Vitrectomy by 20 Weeks 60% Ranibizumab Saline Cumulative Probability 50% P = 0. 35 40% 28% 30% 23% 10% 0% -10% 0 4 8 12 16 20 24

“Complete” Panretinal Photocoagulation Secondary Outcome 25

Cumulative Probability of “Complete” PRP (in absence of vitrectomy) by 16 Weeks 60% Ranibizumab Saline Cumulative Probability 50% P = 0. 07 44% 40% 31% 30% 20% 10% 0% 0 4 8 12 16

Visual Acuity at Follow-up Visits Secondary Outcome 27

Mean Change in Visual Acuity from Mean Change in Letter Score Baseline 25 Ranibizumab 20 †P = 0. 04 Saline 15 10 5 0 Baseline 4 -weeks 8 -weeks 12 -weeks † Treatment comparison for the mean change in visual acuity at the 12 week visit was performed using a longitudinal mixed model adjusting for baseline visual acuity. 28

Visual Acuity Binary Outcomes at 12 Week Visit Ranibizumab Saline P Injection value* N = 119 N = 129 VA better than 69 letter score (>20/50) and no vitrectomy 45% 33% 0. 10 Severe VA deficiency ≤ 38 letter score (<20/200) 20% 27% 0. 30 Very severe VA deficiency ≤ 4 letter score (<20/800) 11% 16% 0. 25 *P-values were obtained using a logistic regression model adjusted by baseline visual acuity. 29

OCT Signal Strength Exploratory Outcome 30

OCT Signal Strength at Follow Up Baseline = 0 Ranibizumab Injection Saline Injection N =124 N = 134 60% 72% Among Participants with Baseline OCT = 0 N =70 N =89 4 Week Visit 36% 30% > 0* N = 69 N = 87 8 Week Visit 46% 38% > 0* N = 66 N = 88 12 Week Visit 52% 50% > 0* * Defined as a composite outcome of OCT >0 and no vitrectomy prior to the study visit. P value - 0. 50 0. 33 0. 87 31

Safety Outcomes 32

Ocular Adverse Events of Interest Prior to 16 weeks Ocular Events Endophthalmitis Angle or iris neovascularization Neovascular glaucoma Cataract Surgery Recurrent vitreous hemorrhage¥ Ranibizumab Injection N = 125 0 Saline Injection N = 136 1 (< 1%) 1 ( 1%) 4 (3%) 1 (1%) 0 8 (6%) 1 (<1%) 2 (2%) 23 (17%) ¥ Treatment comparison for recurrent vitreous hemorrhage was performed using Fisher Exact test (P-value = 0. 01) 33

Retinal Detachments Prior to 16 weeks Ranibizumab Injection N = 125 Traction retinal detachment 8 (6%) only Saline Injection N = 136 9 (7%) Rhegmatogenous retinal detachment only Combined retinal detachment 1(<1%) 2(2%) 1(<1%) 0 Any Retinal Detachment 10 (8%) 11(8%) 34

Elevated Intraocular Pressure Prior to 16 weeks Ranibizumab Injection N = 125 Saline Injection N = 136 Increase of IOP ≥ 10 mm Hg from baseline 6% 8% IOP ≥ 30 mm. Hg 3% 3% Currently on IOP-lowering medication, but not at baseline 10% 0 0 13% 14% Glaucoma surgery at anytime Elevated IOP/Glaucoma Percentages represent the number of eyes with at least 1 occurrence.

Systemic Adverse Events Cardiovascular Events Non-fatal myocardial infarction Non-fatal stroke –ischemic or hemorrhagic Vascular death Any APTC Event¥ Ranibizumab Injection N = 125 0 Saline Injection N = 136 4 1 2 2 2 1 6 ¥ According to Antiplatelet Trialists’ Collaboration 36

Systemic Adverse Events of Interest Ranibizumab Injection N = 125 7 Saline Injection N = 136 9 Hypertension-related event 4 9 Cardiovascular-related event 12 14 Kidney-related events Cell counts are the number of participants with at least one of the given event during the study Summary of Causes of Deaths Age Treatment Group Cause of Death Days in Study 53 Ranibizumab Unknown 7

Discussion 38

Discussion Rationale for Saline Injections Controls for potential effects of drug vehicle and any physical fluid dynamic effect of an intravitreal injection on VH Prior study found that saline injected eyes had higher rate of visual acuity improvement than eyes under observation alone Allowed complete masking of all study staff, which was essential given the subjective primary outcome measure, i. e. , undergoing vitrectomy 39

Rates of Vitrectomy Trials DRCR- Protocol N Outcome Vitrectomy by 16 weeks Persistence of Vitrase Study VH by 12 weeks Anti-VEGF Saline Observation Injection Group N = 125 13% N = 136 17% - - N = 193 68% - Bevacizumab† Vitrectomy by 12 -14 weeks N = 40 10% - N = 40 45% Pegaptanib¥ Vitrectomy by 16 weeks N = 15 40% - - † Intravitreal bevacizumab and Panretinal Photocoagulation for PDR associated with vitreous hemorrhage, Huang (2009) ¥ Use of pegaptanib for recurrent and non-clearing vitreous hemorrhage in proliferative diabetic retinopathy, Hornan (2010). 40

Discussion Rates of Vitrectomy Rate of vitrectomy following intravitreal ranibizumab or saline in DRCR. net trial appears lower than expected with observation 41

Conclusions Primary Outcome Rate of vitrectomy by 16 weeks was not reduced by intravitreal ranibizumab compared with intravitreal saline in eyes with vitreous hemorrhage due to PDR. 42

Conclusions Secondary Outcomes Secondary outcomes such as short-term change in visual acuity, completion of PRP, and frequency of recurrent vitreous hemorrhage suggest faster hemorrhage clearing with intravitreal ranibizumab. Intravitreal ranibizumab does not appear to increase the risk of retinal detachment 43

Other Considerations Outcomes of intravitreal ranibizumab or saline compared with observation or sham treatment cannot be determined from this study 44

Acknowledgements Top 10 Site Enrollers • • • Site 46; Maturi Site 100; Friedman Site 127; Gonzalez Site 19; Kim Site 39; Sharuk Site 152; Marcus Site 212; Jamal Site 14; Bhavsar Site 111; Elman Site 207; Stoltz Top 10 enrollers with 100% visit completion. (4, 8, and 12 week visits) • • Dr. Friedman Dr. Elman Dr. Marcus Dr. Bhavsar 45

Thank You on Behalf of Diabetic Retinopathy Clinical Research Network (DRCR. net) 61 clinical study sites Study participants who volunteered to participate in this trial DRCR. net Data and Safety Monitoring Committee Genentech provided clinical site funding and the ranibizumab for the study and collaborated in a manner consistent with the Network’s DRCR. net Industry Collaboration Guidelines, the DRCR. net had complete control over the design of the protocol, ownership of the data, and all editorial content of presentations and publications related to the protocol. DRCR. net investigators and staff 46