1. The Diabetic Retinopathy Clinical Research Network Randomized Trial Evaluating Ranibizumab Plus Prompt or Deferred Laser or Triamcinolone Plus Prompt Laser for Diabetic Macular Edema Michael Elman, MD 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 EY14231, EY14229, EY018817 1

2. Background: Protocol B – IVT vs Laser 2  Intravitreal triamcinolone was evaluated previously as a “monotherapy” treatment for DME in a randomized trial conducted by DRCR.net ():  Results suggested that IVT without laser was not superior to focal/grid photocoagulation  Focal/grid photocoagulation in eyes with center- involved DME produces gradual visual acuity improvement of ≥2 lines in about 30% of eyes after 2 years, although approximately 20% of laser treated eyes worsen by ≥2 lines

3. 3 Laser-Ranibizumab-Triamcinolone Randomized Clinical Trial for DME: Laser-Ranibizumab-Triamcinolone Randomized Clinical Trial for DME: Study Objective Evaluate efficacy and safety of: 0.5-mg intravitreal ranibizumab plus prompt laser (within 1 week); or 0.5-mg intravitreal ranibizumab plus deferred laser (≥24 weeks); or 4-mg intravitreal triamcinolone plus prompt (within 1 week) laser, in comparison with sham plus prompt laser for treatment of diabetic macular edema.

4. 4 Study Design Primary outcome: Change in visual acuity from baseline to 1 year (intent to treat analysis) Randomized, multi-center clinical trial At least one eye meeting all of the following criteria: Electronic-ETDRS © best corrected visual acuity letter score of 78 to 24 (~20/32 to 20/320) Definite retinal thickening due to diabetic macular edema involving the center of the macula on clinical examination Central subfield (Stratus OCT™) ≥250 µm

5. Visit/Treatment Schedule: Year 1 – the “4:2:7” Guide 5 Sham+Prompt Laser Ranibizumab +Prompt Laser Ranibizumab +Deferred Laser Triamcinolone +Prompt Laser 048 12 ‘4’ required injections 1620 ‘2’ required injections if not a success* 2428323640444852 ‘7’ additional follow-up visits every 4 weeks; required injection if improvement † but not success* since last injection; otherwise optional Primary Endpoint  Visits were every 4 weeks regardless of whether the eye status was successful, improved, or failed. *Success: Visual acuity letter score ≥84 (~20/20) or OCT CSF <250 µm; retreatment at investigator discretion. † Improvement: OCT central subfield thickness decreased by >10% or visual acuity letter score improved by >5. Sham+prompt laser Ranibizumab+promt laser Ranibizumab+deferred laser Triamcinolone+prompt laser

6. Follow-up Visits at and After 52 Week Visit if Ranibizumab Injection Given 6 Ranibizumab +Deferred or Prompt Laser Drug 5256  Four wks after any ranibizumab injection, the study eye is evaluated for possible additional ranibizumab injection using retreatment criteria as in year 1: If not a success*, but improvement † since last injection, retreatment required; otherwise, retreatment is up to investigator. 606468…………104 *Success: Visual acuity letter score ≥84 (~20/20) or OCT CSF <250 µm; retreatment at investigator discretion. † Improvement: OCT central subfield thickness decreased by >10% or visual acuity letter score improved by >5. Drug Visit Visit*

7. Follow-up Visits at and After 52 Week Visit if Ranibizumab Not Given (4 to 8 to 16 Weeks) 7 Ranibizumab +deferred or Prompt Laser No Drug 52 56  If a ranibizumab injection is not given at the current and previous 2 visits (e.g. week 60 above), the next follow-up visit is in 8 weeks.  If at the next 8 week interval visit the injection is deferred again, the next follow-up visit is in 16 weeks; visits continue every 16 weeks unless a ranibizumab injection is given, at which point the visit schedule goes back to 4 week intervals. No Drug 606468……………….……84 No Drug Visit Skip VisitVisit Skip 16 wks Visit

8. Study Enrollment and Completion 8 Ranibizumab +Prompt Laser N = 187 Ranibizumab +Prompt Laser N = 187 Ranibizumab +Deferred Laser N = 188 Ranibizumab +Deferred Laser N = 188 Sham +Prompt Laser N = 293 Sham +Prompt Laser N = 293 Triamcinolone +Prompt Laser N = 186 Triamcinolone +Prompt Laser N = 186 Eyes Randomized: N = 854 (691 Participants) Eyes Randomized: N = 854 (691 Participants) 1 Year Visit (Primary Outcome) Completion: 94%* 2 Year Visit Completion: 87%** * Includes deaths ** Includes deaths and excludes pending and dropped who are not yet in window

9. 9 Baseline Characteristics Sham +Prompt Laser Ranibizumab +Prompt Laser Ranibizumab +Deferred Laser Triamcinolone +Prompt Laser Median age 63626462 Diabetes type Type I 9%6%8% Type II 89%92%90%89% Uncertain 3%2% 3% Median E-ETDRS © visual acuity letter score (Snellen equivalent) 65 (20/50)66 (20/50) Median OCT CSF thickness (µm) 407371382374

10. 10 Injections/Sham Prior to 1 Year Sham +Prompt Laser N = 274 Ranibizumab +Prompt Laser N = 171 Ranibizumab +Deferred Laser N = 178 Triamcinolone +Prompt Laser N =176 Maximal possible # of sham/injections 13 sham*13 drug 9 sham/ 4 drug Median number of sham/study drug injections to 1 year 11*895 sham/ 3 drug AE Precluding Study Drug Injection † NA2% 15% Compliance with sham/drug injection when required by protocol 96%95%97% Masked participant with 1 study eye identified correct assignment at 1 year 10%88%90%44% *Excludes 56 eyes among 163 participants with 2 study eyes unmasked at baseline when assigned ranibizumab + deferred laser. † % of visits reported; 12% of eyes in the triamcinolone group compared with 3% and 4% in the ranibizumab groups

11. 11 Additional Information On Injections/Sham to 2 Yrs Sham +Prompt Laser N = 274 Ranibizumab +Prompt Laser N = 171 Ranibizumab +Deferred Laser N = 178 Triamcinolone +Prompt Laser N = 176 % of eyes meeting failure criteria* at 1- year study visit 4%2%1%2% Maximal # of drug injections prior to 2 years n/a25 8 Median drug injections prior to 1 year n/a893 Median drug injections prior to 2 years n/a11134 * Failure is defined as: VA 10 or more worse than baseline, OCT CSF ≥250 microns, DME present on clinical exam that is the cause of the visual loss, complete laser given AND ≥13w since last laser treatment with no improvement since the last laser treatment Out of 26 maximum Out of 26 maximum

12. 12 Laser Treatments Prior to 1 Year Sham +Prompt Laser Ranibizumab +Prompt Laser Ranibizumab +Deferred Laser * (permitted starting at 24- week visit) Triamcinolone +Prompt Laser Median number of laser treatments including baseline 3202 Proportion of eyes receiving laser at 48-week visit 26%16%8%21% No, only 1, only 2 or 3 more lasers after baseline 13%,27%,31%,32%,70%,20%,26%,30%, 40%,20%27%,11%10%,1%28%,15% * 3 eyes deviated from the protocol and received laser prior to 24 weeks (2 were given laser at the 1 week safety visit and 1 at the 20 week visit). Out of 4 maximum Out of 4 maximum Out of 2 maximum Out of 4 maximum

13. Visual Acuity 13

14. Mean Change in Visual Acuity (Letters) * at Follow-up Visits 14 * Values that were ±30 letters were assigned a value of 30 P-values for difference in mean change in visual acuity from sham+prompt laser at the 52-week visit: ranibizumab+prompt laser <0.001; ranibizumab+deferred laser <0.001; and triamcinolone+prompt laser=0.31.

15. Change in Visual Acuity (LOCF) at 1 Year* 15 Change in Visual Acuity (letters) Sham +Prompt Laser N = 293 Ranibizumab +Prompt Laser N = 187 Ranibizumab +Deferred Laser N = 188 Triamcinolone +Prompt Laser N = 186 Mean+3+9 +4 Difference in mean change from Sham +Prompt Laser [P Value]** +5.8 [P<0.001] +6.0 [P<0.001] +1.1 [P = 0.31] *Visits occurring between 308 and 420 days from randomization were included as 1-year visits. When more than 1 visit occurred in this window, data from the visit closest to the 1-year target date were used. For other eyes with out any 1-year data (19 eyes in the sham+prompt laser group, 16 eyes in the ranibizumab+prompt laser group, 10 eyes in the ranibizumab+deferred laser group, and 10 eyes in the triamcinolone+prompt laser group) the last observation carried forward (LOCF) method was used to impute data for the primary analysis. **Analysis of covariance adjusted for correlation between 2 study eyes and baseline visual acuity

16. Change in Visual Acuity at 2 Years* 16 Change in Visual Acuity (letters) Sham +Prompt Laser N = 163 Ranibizumab +Prompt Laser N = 106 Ranibizumab +Deferred Laser N = 112 Triamcinolone +Prompt Laser N = 103 Mean+2+7+100 Difference in mean change from Sham +Prompt Laser [P Value]** +5.0 [P = 0.01] +7.2 [P<0.001] -1.6 [P = 0.43] *Visits occurring between 616 and 840 days from randomization were included as 2-year visits **Analysis of covariance adjusted for correlation between 2 study eyes and baseline visual acuity

17. ≥10 Letter Improvement in Visual Acuity at Follow-up Visits 17 P values for the difference in proportion of 10 letter improvement in visual acuity from sham+prompt laser at the 52-week visit: ranibizumab+prompt laser <0.001; ranibizumab+deferred laser <0.001; triamcinolone+prompt laser = 0.16 Visit Week

18. ≥10 Letter Worsening in Visual Acuity at Follow-up Visits 18 Visit Week P values for the difference in proportion of 10 letter worsening in visual acuity from sham+prompt laser at the 52-week visit: ranibizumab+prompt laser <0.001; ranibizumab+deferred laser =0.001; triamcinolone+prompt laser = 0.75

19. Visual Acuity Subgroup Analyses 19

20. 20 Subgroup Analyses   No obvious clinically important difference in results at 1-year primary outcome visit for any of the following subgroups: Prior treatment for DME Baseline visual acuity Baseline OCT-measured central subfield thickening Baseline level of diabetic retinopathy on photos Description of edema by ophthalmologist as predominantly focal or predominantly diffuse   In the subset of pseudophakic eyes at baseline (n = 273), visual acuity improvement in the triamcinolone+prompt laser group appeared comparable to the ranibizumab groups

21. Change in Visual Acuity at 1 Year Stratified by Pseudophakic at Baseline 21

22. Retinal Thickening 22

23. Mean Change in Central Subfield Thickening at Follow-up Visits 23 Visit Week P values are for the difference in mean change in OCT CSF retinal thickness from sham+prompt laser at the 52-week visit: ranibizumab+prompt laser <0.001, ranibizumab+deferred laser <0.001, and triamcinolone+prompt laser <0.001.

24. Change in Retinal Thickening at 1 Year* 24 Change in OCT Central Subfield Thickening a Sham +Prompt Laser N = 271 Ranibizumab +Prompt Laser N = 171 Ranibizumab +Deferred Laser N = 175 Triamcinolone +Prompt Laser N = 173 Mean change from baseline (µm) -102-131-137-127 Difference in mean change from Sham Prompt+Laser [P Value]** -55 [P<0.001] -49 [P<0.001] -52 [P<0.001] Thickness <250 µm with at least a 25 µm decrease from baseline 27%53%42%47% *Visits occurring between 308 and 420 days from randomization were included as 1 year visits. When more than 1 visit occurred in this window, data from the visit closest to the 1 year target date were used. **Analysis of covariance adjusted for baseline OCT retinal thickness and visual acuity and correlation between 2 study eyes a Missing data for 22 eyes in the sham+prompt laser group, 16 eyes in the ranibizumab+prompt laser group, 13 in the ranibizumab+deferred Laser, and 13 eyes in the triamcinolone+prompt laser group (includes missing and ungradeable data [3 in sham+prompt laser, 2 in ranibizumab+deferred laser and 2 in triamcinolone+prompt laser]

25. Change in Retinal Thickening at 2 Years* 25 Change in OCT Central Subfield Thickening a Sham +Prompt Laser N = 152 Ranibizumab +Prompt Laser N = 99 Ranibizumab +Deferred Laser N = 100 Triamcinolone +Prompt Laser N = 93 Mean change from baseline (µm) -133-144-170-95 Difference in mean change from Sham + Laser [P Value]** -31 [P = 0.01] -36 [P = 0.004] -3 [P = 0.81] Thickness <250 µm with at least a 25 µm decrease from baseline 38%54%55%44% *Visits occurring between 616 and 840 days from randomization were included as 2-year visits. When more than 1 visit occurred in this window, data from the visit closest to the 2-year target date were used. ** Analysis of covariance adjusted for baseline OCT retinal thickness and visual acuity and correlation between 2 study eyes ª Excluding pending- Missing data for 2 eyes in the sham+prompt laser group, 2 eyes in the ranibizumab+prompt laser group, 2 in the ranibizumab +deferred laser, and 6 eyes in the triamcinolone+prompt laser group; Ungradeable data for 1 in the ranibizumab+prompt laser, 1 in ranibizumab+deferred laser and 2 in triamcinolone+prompt laser Both groups still have central DME in about 45% at 2 yr

26. Retinopathy (hidden pearl within the trial) 26

27. 27 Step Changes of Improvement/Worsening in Diabetic Retinopathy by Baseline Severity Change from baseline to 1-year visit* Sham +Prompt Laser Ranibizumab +Prompt Laser or Deferred Laser Triamcinolone +Prompt Laser Baseline Severity: Moderately Severe NPDR or Better N = 150N = 182N = 80 Improved by ≥2 levels4%25% Worsened by ≥2 levels7%3% P value for comparison with Sham P = 0.08P =0.17 Baseline Severity: Severe NPDR or worse N = 83N = 121N = 70 Improved by ≥2 levels19%28%13% Worsened by ≥2 levels8%1%3% P value for comparison with Sham P = 0.03P = 0.17 *Photos were missing or ungradeable for 61 eyes in the sham+prompt laser group, 72 eyes in the ranibizumab groups, and 33 eyes in the triamcinolone+prompt laser group

28. Retinopathy Progression During 1 Year of Follow-up 28 Sham N = 293 Ranibizumab N = 375 Triamcinolone N = 186 Reported vitreous hemorrhage OR received PRP 8%3% P Value for comparison with sham --0.0020.02

29. Safety 29

30. 30 Major Ocular Adverse Events During 2-Years of Follow-up 30 Sham +Prompt Laser N = 293 Ranibizumab +Prompt Laser N = 187 Ranibizumab +Deferred Laser N = 188 Triamcinolone +Prompt Laser N = 186 Number of injections 18332140685 Endophthalmitis* 1 (<1%)2 (1%) 0 Pseudoendophthalmitis † 1(<1%)001 (1%) Ocular vascular event ‡ 1 (<1%)1 (1%) 3 (2%) Retinal detachment § 001 (1%)0 Vitrectomy 15 (5%)4 (2%)7 (4%)2 (1%) Vitreous Hemorrhage 27 (9%)6 (3%)8 (4%)7 (4%) * One case unrelated to study drug injection (following cataract extraction) in the sham+prompt laser group; 1 case related to study drug injection and 1 case unrelated to injection (following cataract surgery) in the ranibizumab+prompt laser group; 2 cases related to study drug injection in the ranibizumab+deferred laser group. The 3 cases related to study drug injection in the ranibizumab groups are 0.08% of ranibizumab study drug injections given. † One case unrelated to the study drug injection (vitreous opacity with hypopyon) and one case related to study drug injection in the triamcinolone group. ‡ Includes 2 central retinal vein occlusions and 4 branch retinal vein occlusions. § Includes 1 traction retinal detachment with proliferative diabetic retinopathy and prior panretinal photocoagulation at baseline.

31. 31 Elevated Intraocular Pressure/Glaucoma During 2-Years of Follow-up 31 Elevated Intraocular Pressure/Glaucoma Sham +Prompt Laser N = 293 Ranibizumab +Prompt Laser N = 187 Ranibizumab +Deferred Laser N = 188 Triamcinolone +Prompt Laser N = 186 Increase ≥10 mmHg from baseline 8%9%6%42% IOP ≥30 mmHg 3%2%3%27% Initiation of IOP- lowering meds at any visit* 5% 3%28% Number of eyes meeting ≥1 of the above 11% 7%50% Glaucoma surgery** <1%1%0 *Excludes eyes with IOP lowering medications at baseline **Includes 2 filter and 2 cilliary body destruction

32. 32 Cataract Surgery During 2-Years of Follow-up 32 Sham +Prompt Laser Ranibizumab +Prompt Laser Ranibizumab +Deferred Laser Triamcinolone +Prompt Laser Phakic at baseline N = 192N = 131N = 134N = 124 Eyes that had cataract surgery 12% 13%55%

33. Number of Deaths 33 Sham N = 130 Ranibizumab N = 375 Triamcinolone N = 186 Deaths*7 (5%)15 (4%)6 (3%) *Study participants with 2 study eyes are counted in their injection group.

34. 34 Cardiovascular or Cerebrovascular Events According to Antiplatelet Trialists’ Collaboration through 2-Years 34 Sham ‡ N * = 130 Ranibizumab N * = 375 Triamcinolone N * = 186 Non-fatal myocardial infarction 3%1%3% Non-fatal cerebrovascular accident-ischemic or hemorrhagic (or unknown) 6%2% Vascular death (from any potential vascular or unknown cause † ) 5%2% Any APTC event 12%5%6% * N=Number of Study Participants. Study participants with 2 study eyes are assigned to the non-sham group. Multiple events within a study participant are only counted once per event. ‡One participant had a non-fatal myocardial infarction and a non-fatal stroke (only counted once in the any cardiovascular event row) †Four of the vascular deaths in the sham group, 1 of the vascular deaths in the ranibizumab group, and 1 of the vascular deaths in the triamcinolone group were from an unknown cause

35. Discussion 35

36. 36 Intravitreal Ranibizumab Summary   Intravitreal ranibizumab with prompt or deferred (≥24 weeks) focal/grid laser had superior VA and OCT outcomes compared with focal/grid laser treatment alone. ~50% of eyes had substantial improvement (≥10 letters) while ~30% gained ≥15 letters Substantial visual acuity loss (≥10 letters) was uncommon Results were similar whether focal/grid laser was given starting with the first injection or it was deferred >24 weeks So what is your Standard Care for Center Involved DME? 36

37. 37 Intravitreal Ranibizumab Summary   If ranibizumab is to be given as it was in this study, the data indicate a need to follow eyes continuously undergoing this treatment Additional ranibizumab and/or laser were needed in most eyes through ≥2 years, even if ‘success’ criteria were met early in the course of treatment. So how are you following patients and determining when to re-treat? 37

38. 38 Intravitreal Ranibizumab Treatment Protocol   Results are based on rigorous adherence to a detailed retreatment protocol on a web-based real- time data entry system that provided feedback to the treating physician regarding the treatment to be prescribed at each follow-up visit.   The underlying rationale of the treatment algorithm is to place 4 mandatory injections and then continue treatment, as needed, until stabilization or lack of further improvement is noted.

39. 39 Intravitreal Ranibizumab Treatment Protocol   Once a retreatment is withheld, the algorithm is designed to identify when there is a need to re- initiate treatment. The goal was to avoid substantial vision loss while also avoiding a regimen which requires monthly treatments regardless of the clinical course.   The impact of different retreatment approaches or use of other anti-VEGF drugs (such as bevacizumab) in clinical practice cannot be determined from this study. 39

40. 40 Intravitreal Triamcinolone Summary   Intravitreal triamcinolone combined with focal/grid laser did not result in superior VA outcomes compared with laser alone, despite greater reduction in retinal thickening at 1 year (but not 2 years) compared with laser alone.  but is associated with an increased risk of intraocular pressure elevation.  In an analysis limited to pseudophakic eyes, the triamcinolone group’s outcome for VA appeared to be of similar magnitude to that of the 2 ranibizumab groups, but is associated with an increased risk of intraocular pressure elevation. So what will your first line management recommendation be for pseudophakic patients? 40