pdf mata

htRyMichael S. Ip, MD,1 Amitha Domalpally, MD,1 Jennifer K. Sun, MD, MPH,2 Jason S. Ehrlich, MD, PhD3

Purpose: To assess the effects of intravitreal ranibizumab on diabetic retinopathy (DR) severity whenadministered for up to 3 years, evaluate the effect of delayed initiation of ranibizumab therapy on DR severity, andidentify baseline patient characteristics associated with the development of proliferative DR (PDR).

Design: Exploratory analyses of phase III, randomized, double-masked, sham-controlled multicenter clinicaltrials.

Participants: Adults with diabetic macular edema (DME) (N 759), baseline best-corrected visual acuity20/40 to 20/320 Snellen equivalent, and central foveal thickness 275 mm.

Methods: Patients were randomized to monthly 0.3 or 0.5 mg ranibizumab or sham injections. Sham par-ticipants could switch to 0.5 mg ranibizumab during the third year (sham/0.5 mg crossover). Baseline risk factorswere evaluated to explore potential associations with development of PDR. Time to rst development of PDR wasanalyzed by KaplaneMeier methods to calculate cumulative probabilities by group.

Main Outcome Measures: Study eye change on the Early Treatment Diabetic Retinopathy Study severityscale and a composite clinical outcome evaluating progression to PDR based on photographic changes plusclinically important events dening PDR.

Results: At month 36, a greater proportion of ranibizumab-treated eyes had 2- or 3-step DR improvementcompared with sham/0.5 mg crossover. A 3-step improvement was achieved at 36 months by 3.3%, 15.0%,and 13.2% of sham/0.5 mg, 0.3 mg, and 0.5 mg ranibizumab-treated eyes, respectively (P < 0.0001). Through 36months, 39.1% of eyes in the sham/0.5 mg group developed PDR, as measured by composite outcome,compared with 18.3% and 17.1% of eyes treated with 0.3 or 0.5 mg ranibizumab, respectively. The presence ofmacular capillary nonperfusion at baseline seems to be associated with progression to PDR in ranibizumab-treated eyes but did not meaningfully inuence visual acuity improvement in eyes with DME after ranibizumabtherapy.

Conclusions: Ranibizumab, as administered to patients with DME for 12 to 36 months in these studies, canboth improve DR severity and prevent worsening. Prolonged delays in initiation of ranibizumab therapy may limitthis therapeutic effect. Although uncommon, the development of PDR still occurs in a small percentage of eyesundergoing antievascular endothelial growth factor therapy and may be related to the presence of macularnonperfusion. Ophthalmology 2015;122:367-374 2015 by the American Academy of Ophthalmology. This is anopen access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Supplemental material is available at www.aaojournal.org.

Diabetic retinopathy (DR) is a leading cause of visual loss inthe United States, with a prevalence of more than 40% inpatients aged >40 years with diabetes.1 Two general DRsubtypes exist: nonproliferative DR (NPDR) and prolifera-tive DR (PDR). Diabetic macular edema (DME) may bepresent in eyes with either subtype and is frequently theprimary cause of vision loss due to DR. Another majorcause of vision loss in patients with DR is the developmentof retinal neovascularization (e.g., PDR) and its accompa-nying complications. The natural history of NPDR in manypatients is a slow, inexorable worsening, with characteristic

changes that occur in well-dened and discrete steps. Thelevel of DR severity, as observable on retinal color fundusphotographs, is described by the standardized Early Treat-ment Diabetic Retinopathy Study (ETDRS) DR severityscale.2 The discrete changes that occur with diseaseprogression leading up to the development of frankneovascularization provide an opportunity to evaluate theeffectiveness of new therapies that may arrest the progres-sion of the disease or even reverse it.

In a prior report from our group,3 we described the effectsof monthly intravitreal antievascular endothelial growth

367 2015 by the American Academy of OphthalmologyThis is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/3.0/). Published by Elsevier Inc.

http://dx.doi.org/10.1016/j.ophtha.2014.08.048ISSN 0161-6420/14Long-term Effects of TRanibizumab on DiabeSeverity and BaselineWorsening Retinopatherapy withic Retinopathyisk Factors for

factor (VEGF) therapy with ranibizumab for 24 months on Methods

Ophthalmology Volume 122, Number 2, February 2015DR severity using data from the RIDE and RISE phase IIIclinical trials that evaluated the efcacy and safety of rani-bizumab for DME. In that exploratory analysis, we showedthat monthly ranibizumab for 24 months had profound andbenecial effects on DR severity: treatment with ranibizu-mab prevented the worsening of DR (i.e., ETDRS severitylevel progression) and led to DR improvement (i.e., ETDRSseverity level reduction). A composite measure evaluatingPDR development also was used to demonstrate the benetof ranibizumab; this composite outcome included not onlythe described fundus photographic changes, but also clinicalmeasures, such as the need for panretinal laser photocoag-ulation or vitrectomy as treatment for complications ofPDR. We noted that sham-treated patients were 3-fold morelikely to develop PDR than patients treated with ranibizu-mab over 24 months (33.8% vs. 11.2%e11.5%, respec-tively).3 Retarding the progression of DR has been reportedin analyses from other studies of intravitreal agents (i.e.,anti-VEGF therapies, steroids), as well as with systemictherapies, such as candesartan and fenobrate.4e8

Compelling preclinical and clinical data suggest that theretinal pathophysiology of DR is mediated in substantialpart by VEGF.9e12 Therefore, based on the biologicalplausibility that VEGF plays an important role in the clinicalcourse of diabetic eye disease, further studies are needed toevaluate anti-VEGF therapies for the modication of DRprogression.

In the RIDE and RISE trials, the active-treatment armswere assigned to monthly ranibizumab therapy for 36 months.Patients randomized to sham for the rst 24 months wereeligible for crossover to 0.5 mg ranibizumab monthly(sham/0.5 mg) starting at month 25. Thus, in this report wecompare the effect of a 2-year delay in the initiation oftreatment with ranibizumab on retinopathy severity levelbetween the ranibizumab treatment arms and the sham/0.5 mg crossover arm at month 36. Although in RIDE andRISE ranibizumab therapy signicantly reduced the rate ofprogression to PDR at 24 months versus sham, a smallpercentage of eyes treated with monthly intravitreal ranibi-zumab nevertheless experienced a progression from non-proliferative to proliferative disease. Thus, in the currentanalysis we also sought to determine baseline risk factorsassociated with the development of PDR. Previously, clin-ical factors that have been associated with an increasedlong-term risk of developing PDR included elevated he-moglobin A1c (HbA1c), longer duration of diabetes, othermarkers of diabetes severity and microvascular damage (i.e.,proteinuria, neuropathy), and elevated blood pressure.13e19

Further exploration of potential risk factors for progressionto PDR despite treatment with anti-VEGF therapy isimportant because identication of eyes at risk may allowfor intensied therapy and/or closer monitoring of patientswhen needed to reduce the likelihood of developing thisvision-threatening complication. In addition, identicationof subgroups at higher risk of developing PDR even in thesetting of anti-VEGF therapy is important because thesepatients may have unique genetic or other characteristicsthat could help identify additional target pathways for futuretherapeutics in retinal vascular disease.

368Clinical Trial Design

RIDE and RISE were methodologically identical, randomized,phase III, double-masked, sham injectionecontrolled clinical trialsof ranibizumab in patients with DME; the design, baseline patientcharacteristics, and core efcacy and safety outcomes of the trialshave been described elsewhere.3,20 Study protocols were approvedby institutional review boards and ethics committees, and partici-pants provided written informed consent. RIDE and RISEare registered on ClinicalTrials.gov with registration identiersNCT00473382 and NCT00473330, respectively.

Patients and Treatment

Individuals aged 18 years and older with decreased vision dueto DME (study eye best-corrected visual acuity [BCVA] of20/40e20/320 approximate Snellen equivalent) and central fovealthickness 275 mm on time-domain optical coherence tomography(OCT) were eligible for enrollment. One eye per patient wasrandomized to monthly sham injections or intravitreal injectionsof 0.3 or 0.5 mg ranibizumab through month 24. From months24 to 36, patients originally randomized to ranibizumab continuedwith monthly therapy at their assigned dosage. Patients initiallyrandomized to sham were eligible to switch to 0.5 mg ranibizumabmonthly starting at month 25. In this report, this is referred to as thesham/0.5 mg or sham/0.5 mg crossover group.

Grading Protocol and Clinical Assessment ofDiabetic Retinopathy Progression

Stereoscopic 7-eld color fundus photographs were obtained ateach patients screening visit and at months 3, 6, 12, 18, 24, 30,and 36. Photographs were graded according to the ETDRS severityscale for retinopathy level and were evaluated and dened in thesame manner as previously described.3 To include the clinicallyimportant DR progression events occurring between the periodicphotographic assessments, we measured DR progression using thesame composite outcome as previously described.3,21

Statistical Analyses

Baseline distributions of retinopathy severity were assessed and weresimilar across the RIDE and RISE studies; thus, data were pooled forthese analyses. Unless otherwise specied, analyses of the outcomesare based on the assessment of the study eye only. The ETDRSretinopathy severity level was summarized over time. The number ofeyes worsening (i.e., ETDRS level progression) or improving (i.e.,ETDRS level reduction) by 2 or 3 steps from baseline weresummarized at month 36. CochraneManteleHaenszel chi-squaretests stratied for baseline study eye visual acuity (55 vs. >55ETDRS letters), baseline HbA1c level (8% vs. >8%), and priortreatment for DME in the study eye (yes vs. no) were used to comparethe rates of DR worsening and improvement among patients treatedwith ranibizumab versus sham/0.5 mg; Pearson chi-square tests wereused to compare results between the ranibizumab groups and thesham/0.5 mg crossover group. Missing data were imputed using thelast observation carried forward method.

The cumulative probability of developing PDR at month 36 wasanalyzed in each treatment group using KaplaneMeier methods.The log-rank test was used to compare the risk of developing PDRamong the treatment groups. Univariate and multivariate Cox pro-portional hazard models were used to evaluate baseline risk factorsfor progression to PDR for the sham- and ranibizumab-treated pa-tients. As with the assessment of DR improvement/worsening, this

athy

Ip et al Long-term Effects of Ranibizumab on DR SeverityFigure 1. Distribution of patients by change in severity of diabetic retinopevaluation of progression to PDR was also stratied by baselinestudy eye visual acuity letter score (55 vs. >55 ETDRS letters),baseline HbA1c level (8% vs.>8%), and prior treatment for DMEin the study eye (yes vs. no). Baseline factors emerging as signicantin the univariate model were then carried forward into the multi-variate analysis. The model used 3 approaches for assessing DR

scale from baseline to 36 months. Reported values are the percentage of patientsthe y-axis) at each visit; the size of the circle represents the percentage of patiencolumn total 100%. Black lines represent the median level of DR severity over tiproliferative diabetic retinopathy.

Figure 2. Baseline and month 36 fundus photographs of a patient treated with raseverity. CFT central foveal thickness.(DR) on Early Treatment Diabetic Retinopathy Study (ETDRS) severityseverity: continuous (per-step change), ETDRS severity level47 versus 53, and ETDRS severity level 53 versus 60.

Mean change in BCVA from baseline over time was summa-rized by baseline macular capillary nonperfusion status amongpatients treated with ranibizumab. A t test was used to compare themean change in BCVA at month 24 between the patients with and

(numbers above circles) with the stated level of severity (in parentheses onts in the DR severity category at each respective visit. Percentages in eachme. BL baseline; NPDR nonproliferative diabetic retinopathy; PDR

nibizumab. Images indicate substantial regression in the level of retinopathy

369

without macular nonperfusion. This analysis was limited to dataat 24 months to enable comparison of risk factors for progression

patients in the sham group had already developed PDR during therst 24 months and thus could no longer signicantly improve onthe DR severity scale.

Diabetic Retinopathy Severity in Patients Treatedwith Ranibizumab from Study Entry

Consistent with observations after 24 months of therapy,3 longer-term (36-month) treatment with ranibizumab was associated withsubstantial improvement in DR severity level (Fig 2). Eyes treatedwith ranibizumab were substantially more likely to have a 2- and3-step improvement (regression) in DR severity level (Fig 3).Approximately 39% of ranibizumab-treated eyes had 2-stepimprovement in DR severity at 36 months compared with only24% of those in the sham/0.5 mg crossover group (P 0.0003 foreach comparison of ranibizumab dose vs. sham). Diabetic reti-nopathy severity improved by 3 ETDRS levels in 15% and 13%of eyes treated with 0.3 and 0.5 mg ranibizumab, respectively. Thisis a signicantly greater reduction in severity compared with the3% of eyes achieving this level of improvement in the sham/0.5 mgcrossover group (P < 0.0001 for each comparison of ranibizumabdosage vs. sham), and corresponds to a 4- to 5-fold greater like-lihood of achieving 3-step improvement in DR severity withmonthly ranibizumab therapy for 3 years. Of note, this is comparedwith eyes that received only 1 year of 0.5 mg ranibizumab after a

Figure 3. Percentage of patients with improvement and worsening of diabeticretinopathy severity level measured by change from baseline in Early Treat-ment Diabetic Retinopathy Study (ETDRS) letters. Data at 36 months areshown for patients treated with sham/0.5 mg (n 239), 0.3 mg ranibizumab(n 234), and 0.5 mg ranibizumab (n 234). *P < 0.001 versus controlgroup. **P < 0.05 versus control group. Vertical bars are unadjusted 95%condence intervals. Study eye P values versus sham/0.5 mg crossover wereadjusted for baseline study eye visual acuity (55 vs. >55 letters), baselinehemoglobin A1c (8% vs.>8%), and study eye prior to treatment for diabeticmacular edema (yes vs. no; CochraneManteleHaenszel chi-square test).

etin

Ophthalmology Volume 122, Number 2, February 2015to PDR between the ranibizumab groups and sham (beforecrossover).

Results

Diabetic Retinopathy Severity in Patients withDelayed Treatment (Sham/0.5 mg Crossover)

As shown in Figure 1, the median baseline DR severity level in all3 treatment groups was moderately severe NPDR (ETDRS level47). The median DR category improved by 2 levels of ETDRSseverity in participants treated with 0.3 or 0.5 mg ranibizumabmonthly (from moderately severe to mild NPDR), whereas in thesham/0.5 mg crossover group, the median DR severity remainedconstant. As described below, sham-treated eyes crossing over to0.5 mg ranibizumab beneted from anti-VEGF therapy withrespect to DR severity level; however, the median DR severitylevel remained constant through 36 months, perhaps because many

Table 1. Participants Progressing to Proliferative Diabetic RProgression Category

Baseline to Year 2

Sham(n 257)

Ranibi

0.3 mg(n 250)

Progression from NPDR to PDRy 18 3Received PRP laser 31/21 2/2Reported vitreous hemorrhage 41/23 13/13Progression from NPDR to PDRidentied by ophthalmoscopy

33/9 6/4

Underwent vitrectomy 16/3 0/0Reported iris neovascularization 2/0 1/0Reported retinal neovascularization 23/0 1/0Total with progression to PDR 74 22

Data shown are total/additional numbers of patients (not counted in precedingNPDR nonproliferative diabetic retinopathy; PDR proliferative diabetic r*Patients randomized to sham therapy were eligible for crossover to monthly 0yDocumented on fundus photographs.

3702-year delay in treatment, during which these patients had beenassigned to sham therapy.

Proliferative Diabetic Retinopathy

More patients in the sham/0.5 mg crossover group progressed toPDR compared with those receiving 36 months of ranibizumab(Table 1). From baseline to 24 months (720 days), a total of 74 of257 eyes in the sham treatment group progressed to PDR comparedwith only 22 of 250 and 26 of 252 eyes in the 0.3 and 0.5 mgranibizumab groups, respectively.3 At month 36 (1090 days), 87 of257 eyes in the sham/0.5 mg crossover group progressed to PDRcompared with only 32 of 250 eyes and 38 of 252 eyes in therespective ranibizumab groups. Vitreous hemorrhage was the mostfrequent manifestation of proliferative disease in ranibizumab-treated eyes. In the sham and sham/0.5 mg crossover groups, vit-reous hemorrhage and need for panretinal laser were the mostcommon evidence of progression to PDR.

opathy During the 36-Month Controlled Treatment Period

Baseline to Year 3

zumab

Sham/XO*(n 257)

Ranibizumab

0.5 mg(n 252)

0.3 mg(n 250)

0.5 mg(n 252)

4 29 7 93/2 34/18 4/4 7/512/11 45/26 18/16 18/1311/8 40/12 9/5 16/9

3/0 16/2 0/0 4/11/0 3/0 2/0 1/06/1 26/0 1/0 7/126 87 32 38

rows).etinopathy; PRP panretinal photocoagulation; XO crossover..5 mg ranibizumab at month 25.

An important outcome of this analysis was that treatment with

ranibizumab signicantly reduced the risk of progression to PDRcompared with sham therapy during the rst 2 years, when patientsin the sham group crossed over to 0.5 mg ranibizumab during thethird year, their rate of PDR development was attenuated andcomparable to the rates of progression seen in the active-treatmentarms in the rst 24 months; 13 eyes in the sham/0.5 mg crossovergroup progressed to PDR during year 3 compared with 10 and 12eyes in the 0.3 and 0.5 mg ranibizumab groups, respectively.

Baseline Predictive Factors of Proliferative DiabeticRetinopathy Progression

Approximately 9% to 10% of eyes receiving monthly ranibizumabtherapy still developed PDR at 24 months, and for this reason wesought to determine whether there were any systemic or ocularcharacteristics that might be predictive of progression to PDR,especially in the setting of chronic intravitreal anti-VEGF therapy.To facilitate comparison of predictive risk factors in sham- andranibizumab-treated eyes, 24-month data were used. This evalua-tion followed a sequential stepwise analysis; baseline characteris-tics emerging as signicant in univariate analyses were carriedforward into multivariate analysis.

Using this methodology, baseline factors identied as not beingpredictive of progression to PDR for both the sham and ranibizu-mab-treated groups included duration of diabetes (years), HbA1clevel, proteinuria, renal failure, hypertension, and smoking status.Several baseline characteristics were identied through univariateanalysis to be predictive of progression to PDR (Table 2).

Figure 4. KaplaneMeier analysis of time to rst proliferative diabeticretinopathy (PDR) progression from baseline in the pooled RIDE and RISEpopulation. Cumulative probabilities were calculated using theKaplaneMeier method. Progression was dened by (1) progression fromnonproliferative diabetic retinopathy (DR) (severity level 0 at a later time point; (4) case identied byophthalmoscopy; (5) vitrectomy; (6) iris neovascularization adverse event;and (7) retinal neovascularization adverse event. *P < 0.0001. Dashed ver-tical line indicates the sham crossover to 0.5 mg ranibizumab at month 25.

Ip et al Long-term Effects of Ranibizumab on DR Severityranibizumab signicantly reduced the rate of developing PDR(Fig 4). By month 36 (1090 days), the cumulative probability ofDR progression using the composite analysis was 39.1% of eyes inthe sham/0.5 mg crossover group versus 18.3% and 17.1% of eyestreated with 0.3 and 0.5 mg ranibizumab, respectively(P < 0.0001). From as early as month 6 of the studies, separationin the rate of PDR incidence was evident between the ranibizumaband sham groups. The rates of developing PDR were linear in all 3treatment arms in the rst 24 months, and were consistently lowerin eyes receiving ranibizumab therapy. Although treatment withTable 2. Factors Identied with Univariate Analyses as Predi

Ocular Features Comparator Group

Sham-treated patients*Central subeld thickness (mm) Each 15-mm increaseTotal retinal volume (mm3) Each unit increaseFocal or diffuse edema Diffuse vs. focalSubretinal uid presence on OCT Yes vs. noBilateral DME involvement Yes vs. noBCVA Each unit increaseCapillary loss within grid Yes vs. noIOP Each unit increaseDR severity Each step increaseDR severity 47 vs. 53 ETDRDR severity 53 vs. 60 ETDRCentral foveal thickness (mm) Each 15-mm increaseContrast sensitivity Each unit increaseRetinal thickening at center of macula

were identied with univariate analysis: diffuse-type edema on

this study because of the study design. However, in the

that cohort will be the subject of a future report.

Table 3. Factors Identied with Multivariate Analyses as Predic-tive of Progression to Proliferative Diabetic Retinopathy

ComparatorGroups

Hazard Ratio(95% CI)

PValue

Sham-treated patients*DR severity (ETDRS level) 53 vs. 47 4.23 (2.24e7.98)

severity and prevent disease progression. These results add3

matrix metalloproteinases, and a diversity of responses that

photographsean extension of the modied Airlie Houseclassication: ETDRS report number 10. Ophthalmology

intravitreal implant for diabetic macular edema: a 3-year

Ip et al Long-term Effects of Ranibizumab on DR Severityfurther weight to the conclusions of our prior report,namely that ranibizumab or a combination of ranibizumabplus focal laser treatment should be considered in eyes inwhich either therapy (ranibizumab or focal laser alone) maybe appropriate, in order to take full advantage of the addi-tional benecial effect of ranibizumab therapy on reducingDR severity.

In the pooled RIDE and RISE studies, systemic factorssuch as baseline duration of diabetes, HbA1c, proteinuria,renal failure, hypertension, and smoking status were notsignicantly associated with the development of PDR. Whileother studies have demonstrated these factors to be associatedwith DR development and worsening, they may not havebeen prognostic in the RIDE/RISE studies, because in thispatient population with DME, a substantial disease burdenwas already present at baseline (i.e., late-stage DR compli-cations had already developed). We speculate that in thesepatients, intraocular conditions such as the degree of macularcapillary nonperfusion may contribute more to DR wors-ening than systemic factors, especially over a relatively short(2-year) time period. Of note, in these data the only baselinefactor in ranibizumab-treated patients that was predictive ofprogression to PDR was the presence of macular capillarynonperfusion on uorescein angiography. Although macularcapillary nonperfusion was predictive of progression to PDR,it was not predictive of gains in visual acuity achieved withranibizumab therapy. One might speculate that eyes withbaseline capillary nonperfusion would have both worsevision and worse outcomes with treatment, but this was notobserved; similar gains in BCVA after ranibizumab therapywere seen whether macular capillary nonperfusion was pre-sent or absent. This may be related to the effects of VEGFinhibition on retinal nonperfusion in the setting of diabetes,as recently discussed by Campochiaro et al.22

In a small proportion of eyes (w10%), anti-VEGFtherapy was not effective for the prevention of progressionto PDR. In these eyes, disease progression may haveresulted from VEGF-independent pathways. The therapeuticintraocular concentration achieved with monthly ranibizu-mab is several orders of magnitude above the VEGFreceptor half maximal inhibitory concentration. The ranibi-zumab half-maximal in vitro inhibitory concentration forthe VEGF receptor has been estimated at approximately3 ng/ml,23 whereas the trough steady-state intravitreal con-centration of ranibizumab achieved with monthly dosingranges from 12 000 to 20 000 ng/ml (0.3 and 0.5 mg doses,respectively).24 Therefore, pathways other than that medi-ated by VEGF likely contribute to or are the cause of diseaseprogression in these eyes, perhaps particularly so in thegroup of patients identied as having macular capillarynonperfusion. Combined with the hypoxia/ischemia result-ing from degradation of retinal vessels, the ongoing biologicinsults associated with metabolic dysfunction (hyperglyce-mia, hypertension, and dyslipidemia) can compromise vi-sual function through several VEGF-independent signalingcascades. For example, Gologorsky et al25 reviewed amultitude of other pathways that may contribute to DRpathophysiology, including growth factor modulation (e.g.,insulin-like growth factors, angiopoietin), activation ofmulticenter, randomized, controlled clinical trial. Ophthal-mology 2011;118:15807.

5. Chaturvedi N, Porta M, Klein R, et al; DIRECT ProgrammeStudy Group. Effect of candesartan on prevention (DIRECT-Prevent 1) and progression (DIRECT-Protect 1) of retinopathyin type 1 diabetes: randomised, placebo-controlled trials.Lancet 2008;372:1394402.

6. Sjolie AK, Klein R, Porta M, et al; DIRECT Programme StudyGroup. Effect of candesartan on progression and regression ofretinopathy in type 2 diabetes (DIRECT-Protect 2): a rando-mised placebo-controlled trial. Lancet 2008;372:138593.

7. Diabetic Retinopathy Clinical Research Network, Elman MJ,Qin H, Aiello LP, et al. Intravitreal ranibizumab for diabeticmacular edema with prompt versus deferred laser treatment:three-year randomized trial results. Ophthalmology 2012;119:23128.

8. Diabetic Retinopathy Clinical Research Network (DRCR.net).Three-year follow-up of a randomized trial comparing focal/grid photocoagulation and intravitreal triamcinolone for dia-betic macular edema. Arch Ophthalmol 2009;127:24551.

3731991;98(suppl):786806.3. Ip MS, Domalpally A, Hopkins JJ, et al. Long-term effects of

ranibizumab on diabetic retinopathy severity and progression.Arch Ophthalmol 2012;130:114552.

4. Pearson PA, Comstock TL, Ip M, et al. Fluocinolone acetonidelie downstream of inammatory cytokines that may beelevated in patients with DR (e.g., transforming growthfactor-a and -b2 and interleukin-6 and -8).25 Therefore, theobservation in the RIDE/RISE studies that DR severitycontinued to progress in a small proportion of eyes treatedwith ranibizumab likely reects the complex pathophysi-ology of the disease. As new treatments targeting otherpathophysiologic mechanisms continue to be explored, wemight speculate that future therapies could use combinationregimens to arrest DR progression on multiple fronts.

In conclusion, these data demonstrate that modicationof the natural course of DR can be achieved with intravitrealranibizumab. The data also highlight the importance of earlyintervention in the DME patient population, in order to takefull advantage of the ancillary effect on DR severity level.Lastly, this analysis identied macular nonperfusion as abaseline risk factor for progression to PDR, and such eyesmay be considered for closer monitoring and/or ranibizumabtherapy as soon as clinically indicated.

Acknowledgments. Support for third-party writing assistanceby Michael P. Bennett of Envision Pharma Group was provided byGenentech, Inc. The authors thank Pin-Wen Wang, PhD, andJiameng Zhang, PhD, of Genentech, Inc. for statistical support.

References

1. Eye Diseases Prevalence Research Group. The prevalence ofdiabetic retinopathy among adults in the United States. ArchOphthalmol 2004;122:55263.

2. Early Treatment Diabetic Retinopathy Study Research Group.Grading diabetic retinopathy from stereoscopic color fundus

9. Aiello LP, Avery RL, Arrigg PG, et al. Vascular endothelialgrowth factor in ocular uid of patients with diabetic retinopathyand other retinal disorders. N Engl J Med 1994;331:14807.

10. Qaum T, Xu Q, Joussen AM, et al. VEGF-initiated blood-retinal barrier breakdown in early diabetes. Invest OphthalmolVis Sci 2001;42:240813.

11. Tolentino MJ, Miller JW, Gragoudas ES, et al. Intravitreousinjections of vascular endothelial growth factor produce retinalischemia and microangiopathy in an adult primate. Ophthal-

18. Zhang L, Krzentowski G, Albert A, Lefebvre PJ. Risk ofdeveloping retinopathy in Diabetes Control and ComplicationsTrial type 1 diabetic patients with good or poor metaboliccontrol. Diabetes Care 2001;24:12759.

19. Leese G. Longitudinal study examining the risk factorsfor proliferative retinopathy and maculopathy in type-Idiabetes: the Royal College of Physicians of Edin-burgh Diabetes Register Group. Eye (Lond) 2004;18:81420.

Ophthalmology Volume 122, Number 2, February 20152001;44:22039.

Footnotes and Financial Disclosures

Originally received: March 27, 2014.Final revision: May 5, 2014.Accepted: August 18, 2014.Available online: November 18, 2014. Manuscript no. 2014-453.1 Department of Ophthalmology and Visual Sciences, University of Wis-consin Medical School, Madison, Wisconsin.2 Joslin Diabetes Center, Harvard Department of Ophthalmology, Boston,Massachusetts.3 Genentech, Inc., South San Francisco, California.

Portions of this work have been presented at: the Joint American Academyof Ophthalmology/Asia-Pacic Academy of Ophthalmology Meeting,November 9e13, 2012, Chicago, Illinois; the 2013 Macula Society annualmeeting, February 27 to March 2, 2013, Dana Point, California; and the2013 Association for Research in Vision and Ophthalmology AnnualMeeting, May 5e9, 2013, Seattle, Washington.

Financial Disclosure(s):The author(s) have made the following disclosure(s): M.S.I.: Consultant/advisor Eye Technology Ltd, Genentech, Inc., NicOx, Notal Vision, QLTPhototherapeutics Inc., Regeneron, Sirion; Grant support Allergan Inc.

3742012;2012:629452.

J.S.: Consultant e Novartis; Scientic Advisory Board Abbott Labora-tories; Research supportOptovue, Genentech, Inc., BostonMicromachines.J.S.E.: Employee Genentech, Inc. (a member of the Roche Group);Equity and/or options Roche.Supported by Genentech, Inc. Support for third-party writing assistance wasprovided by Genentech, Inc. The sponsor participated in the design andconduct of the study; collection, management, analysis, and interpretationof the data; and preparation and review of the manuscript.

Abbreviations and Acronyms:BCVA best-corrected visual acuity; CI condence interval;DME diabetic macular edema; DR diabetic retinopathy;ETDRS Early Treatment Diabetic Retinopathy Study;HbA1c hemoglobin A1c; HR hazard ratio; NPDR nonproliferativediabetic retinopathy; OCT optical coherence tomography;PDR proliferative diabetic retinopathy; VEGF vascular endothelialgrowth factor.

Correspondence:Michael S. Ip, MD, University of Wisconsin Fundus Photograph ReadingCenter, 8010 Excelsior Drive, Suite 100, Madison, WI 53717. E-mail:[email protected] Complications Study Group. Risk factors forprogression to proliferative diabetic retinopathy in theEURODIAB Prospective Complications Study. Diabetologia

25. Gologorsky D, Thanos A, Vavvas D. Therapeutic in-terventions against inammatory and angiogenic mediators inproliferative diabetic retinopathy. Mediators Inammmology 1996;103:18208.12. Adamis AP, Miller JW, Bernal MT, et al. Increased vascular

endothelial growth factor levels in the vitreous of eyes withproliferative diabetic retinopathy. Am J Ophthalmol 1994;118:44550.

13. Groop LC, Teir H, Koskimies S, et al. Risk factors andmarkers associated with proliferative retinopathy in patientswith insulin-dependent diabetes. Diabetes 1986;35:1397403.

14. Klein R, Moss SE, Klein BE. Is gross proteinuria a risk factorfor the incidence of proliferative diabetic retinopathy?Ophthalmology 1993;100:11406.

15. Davis MD, Fisher MR, Gangnon RE, et al; Early TreatmentDiabetic Retinopathy Study Research Group. Risk factors forhigh-risk proliferative diabetic retinopathy and severe visualloss: Early Treatment Diabetic Retinopathy Study report #18.Invest Ophthalmol Vis Sci 1998;39:23352.

16. Janghorbani M, Jones RB, Allison SP. Incidence of and riskfactors for proliferative retinopathy and its association withblindness among diabetes clinic attenders. Ophthalmic Epi-demiol 2000;7:22541.

17. Porta M, Sjoelie AK, Chaturvedi N, et al; EURODIAB20. Nguyen QD, Brown DM, Marcus DM, et al; RISE and RIDEResearch Group. Ranibizumab for diabetic macular edema:results from 2 phase III randomized trials: RISE and RIDE.Ophthalmology 2012;119:789801.

21. Bressler NM, Edwards AR, Beck RW, et al; Diabetic Reti-nopathy Clinical Research Network. Exploratory analysis ofdiabetic retinopathy progression through 3 years in a ran-domized clinical trial that compares intravitreal triamcinoloneacetonide with focal/grid photocoagulation. Arch Ophthalmol2009;127:156671.

22. Campochiaro PA, Wykoff CC, Shapiro H, et al. Neutralizationof vascular endothelial growth factor slows progression ofretinal nonperfusion in patients with diabetic macular edema.Ophthalmology 2014;121:17839.

23. Yu L, Liang XH, Ferrara N. Comparing protein VEGF in-hibitors: In vitro biological studies. Biochem Biophys ResCommun 2011;408:27681.

24. Xu L, Lu T, Tuomi L, et al. Pharmacokinetics of ranibizumabin patients with neovascular age-related macular degeneration:a population approach. Invest Ophthalmol Vis Sci 2013;54:161624.

Long-term Effects of Therapy with Ranibizumab on Diabetic Retinopathy Severity and Baseline Risk Factors for Worsening Reti ...MethodsClinical Trial DesignPatients and TreatmentGrading Protocol and Clinical Assessment of Diabetic Retinopathy ProgressionStatistical Analyses

ResultsDiabetic Retinopathy Severity in Patients with Delayed Treatment (Sham/0.5 mg Crossover)Diabetic Retinopathy Severity in Patients Treated with Ranibizumab from Study EntryProliferative Diabetic RetinopathyBaseline Predictive Factors of Proliferative Diabetic Retinopathy Progression

DiscussionAcknowledgmentsReferences