P values for comparison with sham with prompt laser for ranibizumab with prompt laser, ranibizumab with deferred laser (Def), and triamcinolone acetonide with prompt laser were .04, .04, and .04 at 1 year; .01, .005, and .64 at 2 years, and .25, .001, and .10 at 3 years, respectively. Each visit week included visits that were ±14 days except the 52-, 68-, 84-, 120-, and 136-week visits, which were ±8 weeks, and the 104- and 156-week visits, which were ±16 weeks.
P value for comparison with sham with prompt laser for ranibizumab with prompt laser, ranibizumab with deferred laser (Def), and triamcinolone acetonide with prompt laser throughout the 3 years of follow-up were .05, .02, and <.001, respectively. Each visit week included visits that were ±14 days except the 52-, 68-, 84-, 120-, and 136-week visits, which were ±8 weeks, and the 104- and 156-week visits, which were ±16 weeks.
eTable 1. Selected baseline characteristics by treatment group for eyes without proliferative diabetic retinopathy
eTable 2. Selected baseline characteristics by treatment group for eyes with proliferative diabetic retinopathy
eFigure. Cumulative probability of worsening of retinopathy among study participants with two study eyes without proliferative diabetic retinopathy at baseline in both eyes by treatment group using life-table method
Bressler SB, Qin H, Melia M, Bressler NM, Beck RW, Chan CK, Grover S, Miller DG, . Exploratory Analysis of the Effect of Intravitreal Ranibizumab or Triamcinolone on Worsening of Diabetic Retinopathy in a Randomized Clinical Trial. JAMA Ophthalmol. 2013;131(8):1033-1040. doi:10.1001/jamaophthalmol.2013.4154
Copyright 2013 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
The standard care for proliferative diabetic retinopathy (PDR) usually is panretinal photocoagulation, an inherently destructive treatment that can cause iatrogenic vision loss. Therefore, evaluating the effects of therapies for diabetic macular edema on development or worsening of PDR might lead to new therapies for PDR.
To evaluate the effects of intravitreal ranibizumab or triamcinolone acetonide, administered to treat diabetic macular edema, on worsening of diabetic retinopathy.
Exploratory analysis was performed on worsening of retinopathy, defined as 1 or more of the following: (1) worsening from no PDR to PDR, (2) worsening of 2 or more severity levels on reading center assessment of fundus photographs in eyes without PDR at baseline, (3) having panretinal photocoagulation, (4) experiencing vitreous hemorrhage, or (5) undergoing vitrectomy for the treatment of PDR.
Community- and university-based ophthalmology practices.
Individuals with central-involved diabetic macular edema causing visual acuity impairment.
Eyes were assigned randomly to sham with prompt focal/grid laser, 0.5 mg of intravitreal ranibizumab with prompt or deferred (≥24 weeks) laser, or 4 mg of intravitreal triamcinolone acetonide with prompt laser.
Main Outcomes and Measures
Three-year cumulative probabilities for retinopathy worsening.
For eyes without PDR at baseline, the 3-year cumulative probabilities for retinopathy worsening (P value comparison with sham with prompt laser) were 23% using sham with prompt laser, 18% with ranibizumab with prompt laser (P = .25), 7% with ranibizumab with deferred laser (P = .001), and 37% with triamcinolone with prompt laser (P = .10). For eyes with PDR at baseline, the 3-year cumulative probabilities for retinopathy worsening were 40%, 21% (P = .05), 18% (P = .02), and 12% (P < .001), respectively.
Conclusions and Relevance
Intravitreal ranibizumab appears to be associated with a reduced risk of diabetic retinopathy worsening in eyes with or without PDR. Intravitreal triamcinolone also appears to be associated with a reduced risk of PDR worsening. These findings suggest that use of these drugs to prevent worsening of diabetic retinopathy may be feasible. Given the exploratory nature of these analyses, the risk of endophthalmitis following intravitreal injections, and the fact that intravitreal triamcinolone can cause cataract or glaucoma, use of these treatments to reduce the rates of worsening of retinopathy, with or without PDR, does not seem warranted at this time.
Diabetic retinopathy is characterized, in part, by development and worsening of retinal ischemia. The anatomic sequel of this pathophysiologic process, retinal neovascularization or proliferative diabetic retinopathy (PDR), can lead to vitreous hemorrhage, traction detachment from fibrous proliferation, or neovascular glaucoma. Irreversible vision loss can result from complications of PDR, although the frequency of severe vision loss is markedly reduced if panretinal photocoagulation (PRP) is performed. In the Early Treatment Diabetic Retinopathy Study (ETDRS),1 approximately 70% of participants with moderately severe to severe nonproliferative diabetic retinopathy at baseline progressed to PDR within 5 years when the participants were assigned to deferral of photocoagulation (ie, initiation of PRP only if high-risk proliferative retinopathy developed). In another study by the Diabetic Retinopathy Study Group,2 nearly half the eyes in which PDR developed experienced profound visual acuity loss (Snellen visual acuity worse than 5/200) from vitreous hemorrhage or traction retinal detachment by 5 years.
Quiz Ref IDThe current standard treatment for PDR is PRP, but this treatment is inherently destructive and has several potential adverse effects on aspects of visual function, including constriction of the peripheral visual field and reductions in night vision, contrast sensitivity, and color perception. Furthermore, in eyes with diabetic macular edema (DME), PRP in the absence of intravitreal administration of ranibizumab or triamcinolone acetonide can negatively affect vision and macular thickness, at least in the short term.3 Thus, therapeutic alternatives that might safely obviate or substantially delay the need for PRP are desirable.
Several clinical reports have suggested that anti–vascular endothelial growth factor (anti-VEGF) therapies, including pegaptanib (Macugen; Eyetech Pharmaceuticals), ranibizumab (Lucentis; Genentech Inc), and bevacizumab (Avastin; Genentech Inc), can cause transient regression of PDR.4- 11 Published reports suggest a consistent and rapid response in the majority of eyes with retinal or anterior segment neovascularization managed with intravitreal anti-VEGF agents. For example, one publication12 reported resolution of angiographic leakage of PDR-associated neovascularization of the disc in 19 of 26 eyes (73%) that were treated with intravitreal bevacizumab. Biological effects of regression of neovascularization were seen with bevacizumab doses as low as 6.2 µg, 200-fold less than the standard clinically used dose of 1.25 mg. Another small, prospective, open-label exploratory study9 randomized 20 individuals with PDR to receive intravitreal pegaptanib or PRP. By week 12, all pegaptanib-treated eyes demonstrated regression of neovascularization; this finding was maintained through the week 36 study visit. A secondary outcome was reported by the Diabetic Retinopathy Clinical Research Network (DRCR.net) from a trial evaluating laser, ranibizumab, and triamcinolone for DME.10 In that study, eyes assigned to sham with prompt laser were more likely to develop a vitreous hemorrhage or receive PRP than were eyes in the ranibizumab groups (8% vs 3%) and were more likely to show worsening from baseline to the 1-year visit as graded on fundus photographs, even though ranibizumab was not given every 4 weeks to all study participants assigned to the ranibizumab groups following the 12-week visit. These latter results suggest a beneficial effect of ranibizumab on the severity level of diabetic retinopathy and on PDR activity, which may be obtained without administering monthly treatments indefinitely. More recently, Ip and colleagues13 reported that, by the month 24 visit in the RIDE and RISE (Ranibizumab Injection in Subjects With Clinically Significant Macular Edema With Center Involvement Secondary to Diabetes Mellitus) trials, following every-4-week administration of sham or 0.3-mg or 0.5-mg of ranibizumab, the cumulative probability of worsening of diabetic retinopathy using a composite outcome was 34% in the sham-treated participants and 11% in the participants randomized to ranibizumab. With respect to intravitreal corticosteroids, the DRCR.net also reported that intravitreal triamcinolone acetonide (4 mg) appeared to reduce the risk of worsening of diabetic retinopathy.10,14 To further evaluate the effect of intravitreal ranibizumab or triamcinolone on worsening of diabetic retinopathy, an exploratory analysis of eyes in the DRCR.net trial evaluating laser, ranibizumab, and intravitreal triamcinolone for DME was performed to compare the effects of these treatments with those of a sham intravitreal injection with prompt focal/grid laser treatment on the risk of worsening of diabetic retinopathy for study participants monitored for up to 3 years.
The methods for the DRCR.net trial that compared sham with prompt focal/grid laser, intravitreal ranibizumab with prompt laser, intravitreal ranibizumab with deferred (≥24 weeks) laser, and intravitreal triamcinolone with prompt laser have been published in detail elsewhere,10 with the complete protocol available online (http://www.drcr.net). In brief, principal eligibility criteria included study eyes with DME confirmed by optical coherence tomography (Stratus; Carl Zeiss Meditec) central subfield thickness of at least 250 μm and a best-corrected visual acuity letter score of 73 through 24 (approximate Snellen equivalent, 20/32 to 20/320) after a protocol refraction and visual acuity measurement by means of an electronic visual acuity protocol.15 Study eyes were not permitted to have intravitreal anti-VEGF agents or intravitreal triamcinolone within 4 months prior to randomization. In addition, study eyes were not permitted to have PRP within 4 months prior to randomization or an anticipated need for PRP in the 6 months after randomization. Eligible eyes were randomized to sham intravitreal injection with prompt focal/grid laser treatment, 0.5 mg of intravitreal ranibizumab with prompt laser treatment, 0.5 mg of intravitreal ranibizumab with deferred (≥24 weeks) laser treatment, or 4 mg of intravitreal preservative-free triamcinolone acetonide with prompt laser treatment. If both eyes of a study participant were entered into the trial, at least 1 eye had to be assigned to sham with prompt laser, resulting in more eyes in that group compared with the other treatment arms. Visits were every 4 weeks throughout the 48-week visit for all treatment groups. Injections (intravitreal or sham) were required every 4 weeks for the initial 12 weeks of the study and then continued every 4 weeks through the 20-week visit if the visual acuity letter score was worse than 84 (20/20 Snellen equivalent) and the optical coherence tomography central subfield thickness was 250 µm or more. Otherwise, reinjection at 16 and 20 weeks was at the investigators’ discretion. Starting at the 24-week visit, injections were repeated every 4 weeks if there was improvement in visual acuity or optical coherence tomography findings from the previous visit and vision remained worse than 20/20 and central subfield thickness was 250 µm or more. Otherwise, reinjection was at the investigators’ discretion. Starting at the 52-week visit, follow-up was extended to 16-week intervals for the sham with prompt laser and triamcinolone with prompt laser groups, and sham injections were discontinued. For the ranibizumab groups, the follow-up interval could be extended to 8 weeks if injections were deferred at 3 consecutive visits and then again to 16 weeks if the injection could still be deferred. Any time ranibizumab injections were resumed, follow-up returned to 4-week intervals.
Standard 7-field or 4-field wide digital color fundus photographs used to assess retinopathy severity level by masked graders at an independent reading center were obtained at baseline, at 1 year, and either at 3 years or (following a protocol amendment) 1 additional time point between 1 and 3 years before investigators were unmasked to the primary outcome results.
Among the 4 treatment arms without PDR at baseline—sham with prompt focal/grid laser treatment, ranibizumab with prompt laser treatment, ranibizumab with deferred laser treatment, and triamcinolone with prompt laser treatment—1-year photographs were available for 93%, 94%, 93%, and 96% of eyes, respectively. Additional fundus photographs were scheduled for the 3-year visit. However, because ranibizumab was demonstrated to be safe and effective for the treatment of DME, the treatment protocol was amended to allow ranibizumab treatment for all study participants. The protocol amendment required fundus photographs at the next study visit, which was labeled as the “change-of-protocol” visit. Because most participants had completed their 2-year visit by the change-of-protocol visit, relatively few had photographs during the second study year, while approximately 50% of participants who completed a visit in the third study year before unmasking of the primary study outcome occurred had fundus photographs (Table 1). For the large number of participants without photographs at 3 years, the actual protocol 3-year visit was completed after the protocol change and ineligible for inclusion in analysis, and an earlier visit that occurred prior to the protocol change and falling within the predefined analysis window for the 3-year visit was substituted for the 3-year visit. Most of these visits were completed before institutional review board approval was obtained for the protocol amendment that allowed photographs after 1 year and before 3 years.
Worsening of ETDRS diabetic retinopathy severity level16 in eyes without PDR at baseline included any of the following occurrences during the course of the study: (1) worsening from no PDR (level 53 or better) to PDR (level 60 or worse) as determined by reading the center assessment of standard fundus photographs, (2) worsening by 2 or more levels on the ETDRS diabetic retinopathy severity scale on reading center assessment of fundus photographs, (3) having PRP, (4) experiencing vitreous hemorrhage, or (5) undergoing vitrectomy for treatment of PDR. Worsening of diabetic retinopathy during the study in eyes with PDR at baseline included having PRP, occurrence of vitreous hemorrhage, or vitrectomy for PDR.
Cumulative probabilities of worsening of retinopathy and corresponding 95% CIs at each study visit up to the 3-year visit (between 140 and 172 weeks from randomization) were calculated using the life-table method.17 Data were included only up to the change-of-protocol visit when all participants were permitted to receive ranibizumab.10 Data on study participants who had not completed the 3-year visit and did not meet the definition of worsening of retinopathy were considered to be censored following that visit.
Diabetic retinopathy was evaluated in 2 distinct subgroups based on the absence or presence of PDR at baseline. For the subgroup with PDR at baseline, a proportional hazards model was used to compute P values for the treatment group comparisons, adjusting for baseline covariates that were imbalanced by treatment or associated with worsening of retinopathy, including visual acuity, retinopathy severity, and whether the participant had 1 or 2 study eyes. A robust sandwich estimate of the covariance matrix was used to account for correlation within participants who had both eyes studied. For the subgroup without PDR at baseline, the proportional hazards model was not used to compute P values because of a significant violation of the proportional hazards assumption. Instead, a weighted combination of the life-table survival estimates, stratifying by baseline covariates (retinopathy severity and whether the participant had 1 or 2 study eyes), was used to obtain an overall estimate of survival and standard error for each treatment group at each annual time point that was adjusted for the covariates. Weights were equal to the proportion of participants in each of the covariate strata, with all treatment groups combined. The adjusted estimates were compared between treatment groups using the Z test. All P values are 2-sided. Statistical analyses were conducted using commercial software (SAS, version 9.3; SAS Institute, Inc).
Based on reading center assessment of baseline fundus photographs among the 4 treatment arms, 538 of 792 eyes (67.9%) did not have PDR (level 53 or better). Among the remaining 254 eyes with PDR (diabetic retinopathy severity level 60 or worse), 78.0% had evidence of PRP graded on fundus photographs or a history of PRP. Relevant baseline characteristics by PDR subgroup appeared similar among the treatment groups (eTable 1 and eTable 2 in the Supplement), with the following exceptions: the number of study eyes (because the study design required 1 eye of all participants with 2 eligible eyes to be randomized to sham with prompt laser treatment), the diabetic retinopathy severity level in both PDR subgroups (without PDR and with PDR), prior laser treatment for DME in eyes without PDR, race, insulin use, prior PRP, and optical coherence tomography central subfield thickness in eyes with PDR.
The median and mean number of injections from baseline to 1 year, from 1 year to 2 years, and from 2 years to 3 years by treatment group among eyes without or with PDR at baseline, limited to eyes eligible for analysis of worsening of diabetic retinopathy, are reported in Table 2. Among eyes assigned to ranibizumab with prompt laser treatment or ranibizumab with deferred laser treatment, the median number of injections in the first year was greater than in the second or third year when a median of 0 to 5 injections per year were given, depending on the treatment arm, both in eyes without or with PDR at baseline.
Figure 1 shows the cumulative probabilities of worsening of retinopathy by treatment group among eyes without PDR at baseline. Quiz Ref IDAt the 1-year visit, both ranibizumab groups and the triamcinolone group appeared less likely to have worsening. This result was not sustained for the triamcinolone group at the 2-year or 3-year visit, whereas it appeared sustained for the ranibizumab groups at the 2-year visit. By the 3-year visit, the ranibizumab with deferred focal/grid laser treatment group still appeared to be less likely to have worsening, with 7% (95% CI, 3%-15%) worsening compared with 23% (95% CI, 17%-32%) in the sham with prompt focal/grid laser treatment group. The cumulative probability of worsening for the ranibizumab with prompt focal/grid laser treatment group was 18% (95% CI, 10%-30%). These life-table results were similar when only worsening from no PDR to PDR or worsening of 2 steps or more on fundus photographs was used as the outcome (data not shown).
Figure 2 shows the Quiz Ref IDcumulative probabilities of worsening of retinopathy by treatment group among eyes with PDR at baseline. Both ranibizumab groups and the triamcinolone group were less likely to have worsening throughout the 3 years of follow-up, including 21% (95% CI, 11%-36%) for the ranibizumab with prompt laser treatment group, 18% (95% CI, 8%-37%) for the ranibizumab with deferred laser treatment group, and 12% (95% CI, 6%-23%) for the triamcinolone with prompt laser treatment group compared with 40% (95% CI, 29%-54%) for the sham with prompt laser group at the 3-year visit. Using the proportional hazards models, the relative risks for worsening compared with sham with prompt laser for each of the 3 treatment groups were 0.43 (95% CI, 0.19-0.98), 0.42 (95% CI, 0.20-0.89), and 0.23 (95% CI, 0.097-0.54) for ranibizumab with prompt laser, ranibizumab with deferred laser, and triamcinolone with prompt laser, respectively.
Only 26 study participants had 2 study eyes with PDR in both eyes at baseline, too few to determine whether the cumulative probabilities of worsening of retinopathy in this subgroup were similar to the results noted for the entire study group with PDR at baseline. There were 100 study participants without PDR in both eyes at baseline who had 1 eye assigned to the sham with prompt laser group and 1 eye to 1 of the 3 other intravitreal injection groups. In this subgroup of participants with 2 study eyes without PDR at baseline, patient-level confounders, such as systemic control of diabetes mellitus and all other systemic (nonstudy eye) and environmental variables that might affect worsening of retinopathy, were the same among eyes assigned to sham with prompt laser treatment and eyes assigned to intravitreal injections. The cumulative probabilities of worsening of retinopathy in each of the treatment arms in this subgroup of study participants without PDR at baseline (eFigure in the Supplement) appeared similar to results noted for the entire study group without PDR at baseline (Figure 1).
This exploratory analysis suggests that 0.5 mg of intravitreal ranibizumab, when used to treat DME involving the center of the macula in eyes with vision impairment, is associated with a reduced probability of retinopathy worsening through 3 years in eyes without or with PDR. These findings are consistent with an association of a reduced risk of worsening of retinopathy reported in other trials evaluating ranibizumab in the treatment of DME.13,18,19 Among 148 eyes without PDR at baseline that completed the 3-year visit, the treatment effect appears to be of greater magnitude through 3 years in the ranibizumab with deferred laser treatment group (n = 30) relative to the ranibizumab with prompt laser treatment group (n = 31). This observed difference could be related to the larger number of ranibizumab injections administered in the ranibizumab with deferred laser treatment group (mean, 5 total injections during years 2 and 3) compared with the ranibizumab with prompt laser treatment group (mean, 3 total injections during years 2 and 3). As such, the number and regularity of exposure to anti-VEGF therapy that may be needed to indefinitely retard worsening of diabetic retinopathy cannot be determined from the results of this analysis.
One major limitation of this analysis is that the investigators were aware of the randomization assignment for each study eye; therefore, investigator bias could have affected the assessment of 2 of the 5 components used in the composite primary outcome measure. However, these DRCR.net results are similar to those in another report13 in which ranibizumab was administered every 4 weeks continuously for 2 years to manage DME, and rates of retinopathy worsening also were reduced when compared with rates in eyes managed with sham injection. In addition, results were similar to those in a secondary analysis among DRCR.net study eyes without PDR at baseline in which retinopathy worsening was based solely on the event of at least a 2-step worsening of diabetic retinopathy severity level determined by masked photographic grader assessment (data not shown).
Another potential limitation of this study relates to the focus of this trial, which was to determine the effect of intravitreal ranibizumab or triamcinolone on DME rather than the prevention of retinopathy worsening. Had the trial focused on retinopathy worsening, different and more frequent photographic assessment of retinopathy worsening likely would have been incorporated. Nevertheless, our findings were similar to those reported by Ip et al,13 whose study included more frequent photographic outcome assessment.
In the present study, fundus photographs after 1 year were obtained for a very limited group, including 6% of eyes with a 2-year visit completed during the second year of follow-up and 50% of eyes with a 3-year visit. The lack of photographic assessments was because of the change of treatment protocol related to the beneficial effect of ranibizumab on DME. The overall proportions of individuals with worsening of diabetic retinopathy in the second and third years of follow-up likely are underestimated in the absence of these additional photographs. This would not be expected to bias the treatment group comparisons, and the results were in the same direction when using worsening from no PDR to PDR or worsening of 2 steps or more on fundus photographs only as the outcome (data not shown). Another potential limitation of our findings is that the results were limited to eyes with DME at study entry.
Interpretation of the intravitreal triamcinolone results is difficult because the 4-mg triamcinolone group was more likely to have cataract formation and cataract extraction during the 3-year follow-up period. Cataract and cataract surgery may have confounded retinopathy worsening in this group, with issues of ascertainment bias as well as worsening potentially related to the surgery.
The effects of triamcinolone with prompt laser on worsening of retinopathy appeared less in eyes without PDR at baseline compared with those with PDR. Although the triamcinolone with prompt laser group appeared to have a greater risk of worsening of retinopathy than the laser group at the 3-year visit (Figure 1), there were very few eyes (35 of 111) remaining at risk in the triamcinolone with prompt laser group for worsening of retinopathy at this 3-year visit, and the difference vs the laser group was not statistically significant. One might consider that at all other time points the triamcinolone with prompt laser group was not associated with a greater chance of worsening of retinopathy compared with the laser group. In a previous report14 from DRCR.net, intravitreal triamcinolone without prompt laser administered to manage DME appeared to reduce the risk of retinopathy worsening through 3 years compared with laser alone, although that analysis did not involve separating eyes without PDR from those with PDR at baseline.
Strengths of the current investigation include the prospective collection of fundus photographs and treatment data using standardized data collection and treatment protocols, randomization by treatment group, and a cohort of several hundred individuals. Masked assessment of fundus photographs provided uniform assessment of change in diabetic retinopathy severity across 52 clinical sites. Furthermore, outcomes were supplemented by information obtained at the time of clinical examination to capture worsening that may not be apparent on fundus images.
Building on these observations, the DRCR.net is presently conducting a trial to determine whether outcomes at 2 years in eyes with PDR that receive anti-VEGF therapy while deferring PRP for at least 4 weeks, and potentially indefinitely, result in visual acuity outcomes equivalent to those in eyes that receive standard prompt PRP therapy (NCT01489189, Prompt Panretinal Photocoagulation vs Ranibizumab + Deferred Panretinal Photocoagulation for Proliferative Diabetic Retinopathy). If equivalency in visual acuity outcomes is demonstrated, the study will be able to determine whether other visual function outcomes are superior (including visual field and self-reported visual function) in eyes receiving anti-VEGF therapy with deferred PRP compared with those receiving prompt PRP. Additional planned analyses include safety outcomes (eg, the risk of endophthalmitis following intravitreal injections), other administered treatments (eg, vitrectomy), and cost-effectiveness.
Quiz Ref IDUse of intravitreal corticosteroid preparations to reduce the likelihood of retinopathy worsening does not seem warranted because of the potential increased risk of glaucoma and cataract associated with use of intravitreal corticosteroids. In addition, the chance of endophthalmitis following intravitreal injections would have to be considered when weighing the risks and benefits of using intravitreal corticosteroids to reduce the risk of worsening of diabetic retinopathy.
Quiz Ref IDRoutine use of intravitreal anti-VEGF agents to reduce the likelihood of worsening of retinopathy seems premature, pending the results of an ongoing trial designed to specifically compare prompt PRP with intravitreal anti-VEGF with deferred PRP. Proliferative diabetic retinopathy frequently has been treated successfully and safely with PRP and vitrectomy for decades; definitive safety and efficacy of anti-VEGF therapy for worsening of diabetic retinopathy remains to be demonstrated.
Submitted for Publication: October 30, 2012; final revision received February 13, 2013; accepted February 17, 2012.
Corresponding Author: Haijing Qin, MS, Jaeb Center for Health Research, 15310 Amberly Dr, Ste 350, Tampa, FL 33647 (email@example.com).
Published Online: June 27, 2013. doi:10.1001/jamaophthalmol.2013.4154.
Author Contributions:Study concept and design: N. M. Bressler.
Acquisition of data: S. B. Bressler and Grover.
Analysis and interpretation of data: S. B. Bressler, Qin, Melia, N. M. Bressler, Beck, and Chan.
Drafting of the manuscript: S. B. Bressler, Qin, and N. M. Bressler.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Qin and Melia.
Obtained funding: N. M. Bressler.
Administrative, technical, and material support: N. M. Bressler, Chan, and Grover.
Study supervision: S. B. Bressler, Melia, N. M. Bressler, Beck, and Grover.
Conflict of Interest Disclosures: A complete list of all DRCR.net investigator financial disclosures can be found at http://www.drcr.net.
Funding/Support: This study was 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, US Department of Health and Human Services grants EY14231, EY14229, and EY18817.
Role of the Sponsors: The funding organization (National Institutes of Health) participated in oversight of the conduct of the study and review of the manuscript but not directly in the design or conduct of the study; in the collection, management, analysis, or interpretation of the data; or in the preparation of the manuscript. Allergan Inc provided the triamcinolone and Genentech Inc provided the ranibizumab. As per the DRCR.net Industry Collaboration Guidelines (available at http://www.drcr.net), 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. Allergan Inc has provided unrestricted funds to DRCR.net for its discretionary use, and Genentech Inc has provided funds restricted to DRCR.net clinical sites.