|Home | About | Journals | Submit | Contact Us | Français|
To evaluate 14-week effects of intravitreal ranibizumab or triamcinolone in eyes receiving focal/grid laser for diabetic macular edema (DME) and panretinal photocoagulation (PRP).
Three hundred and forty-five eyes with a visual acuity of 20/320 or better, center-involved DME receiving focal/grid laser, and diabetic retinopathy receiving prompt PRP were randomly assigned to sham (n=123), 0.5-mg ranibizumab (n=113) at baseline and 4 weeks, or 4-mg triamcinolone at baseline and sham at 4 weeks (n=109). Treatment was at investigator discretion from 14 to 56 weeks.
Mean changes (±standard deviation) in visual acuity letter score from baseline were significantly better in the ranibizumab (+1±11, P<0.001) and triamcinolone (+2±11, P<0.001) groups compared with the sham group (-4±14) at the 14-week visit, mirroring retinal thickening results. These differences were not maintained when study participants were followed for 56 weeks for safety outcomes. One eye (0.9%, 95% CI: 0.02% to 4.7%) developed endophthalmitis after receiving ranibizumab. Cerebrovascular/cardiovascular events occurred in 4%, 7%, and 3% of the sham, ranibizumab, and triamcinolone groups, respectively.
The addition of 1 intravitreal triamcinolone or 2 ranibizumab injections in eyes receiving focal/grid laser for DME and PRP is associated with better visual acuity and decreased macular edema by 14 weeks. Whether continued long-term intravitreal treatment is beneficial cannot be determined from this study.
Scatter photocoagulation (also referred to as panretinal photocoagulation or PRP) has been the standard treatment for proliferative diabetic retinopathy (PDR) since the Diabetic Retinopathy Study demonstrated that PRP should be considered when an eye approaches or has high risk PDR.1 The 2-year risk of severe visual acuity loss (<5/200 at 2 consecutive visits 4 months apart) without treatment in the Diabetic Retinopathy Study was reduced by approximately 60%. Data from the Early Treatment Diabetic Retinopathy Study (ETDRS) Group suggest that the 5-year risk of severe visual acuity loss for those with PDR could be reduced to approximately 1% with careful follow-up, prompt PRP, and vitrectomy when necessary.2
Although PRP is remarkably effective at reducing visual loss if applied in a timely and appropriate manner, worsening of existing macular edema, often accompanied by visual acuity loss, is a recognized side effect of PRP. Documentation of this side effect is limited and consists mainly of case reports and case series.3–6 In the ETDRS among eyes with center-involved macular edema (as graded on stereoscopic fundus photographs) at baseline, 19% lost ≥10 letters, including 11% that lost ≥15 letters, 4 months following baseline PRP (unpublished data from the ETDRS dataset analyzed by the Jaeb Center for Health Research). However, in a study conducted by the Diabetic Retinopathy Clinical Research Network (DRCR.net), eyes without center-involved diabetic macular edema (DME) at the time of PRP that underwent PRP in a single session did not develop an increase in edema or a reduction of visual acuity that was judged to be clinically meaningful.7 Specifically, in eyes without DME involving the center of the macula, median increases in optical coherence tomography (OCT) central subfield thickness 17 weeks after initiating PRP in 1 or 4 sittings was +14 or +15 μm (25th percentile = +5 or +6, 75th percentile = +20 or +34 μm), respectively, with little decrease in visual acuity (median −1 letter, 25th and 75th percentile = −4 and +2 letters).7
If vascular endothelial growth factor (VEGF) has a role in the development or exacerbation of DME,8, 9 then anti-VEGF drugs or corticosteroids or both might have a role in reducing PRP-induced exacerbation of pre-existing DME in the setting of severe non-proliferative diabetic retinopathy (NPDR) or PDR. Small randomized trials and case reports have suggested such a benefit for intravitreal triamcinolone given as an adjunct to PRP in patients with DME.10, 11 Since the start of this current trial, several additional small randomized trials and retrospective studies have been published that suggest a benefit of intravitreal anti-VEGF drugs or corticosteroids for DME.12–16 In addition, the DRCR.net has reported benefits for at least 1 year of intravitreal ranibizumab treatment of DME in the absence of diabetic retinopathy requiring simultaneous PRP, and has reported exploratory analyses suggesting that triamcinolone can reduce the risk of diabetic retinopathy progression.17, 18
In 2007, the DRCR.net began this randomized trial of 364 eyes with center-involved DME to evaluate the short term effects of intravitreal ranibizumab or intravitreal triamcinolone on pre-existing DME and visual acuity in eyes receiving PRP for severe NPDR or non-high risk PDR and also receiving focal/grid laser for DME. Given that PRP was to be completed within 49 days, and it was desired to have the intravitreal ranibizumab or triamcinolone present while the acute effects of PRP on macular edema could occur, the treatment protocol included intravitreal ranibizumab injections at the baseline and the 4 week visit, and intravitreal triamcinolone injection at the baseline visit. There were no restrictions or study guidelines on treatment for DME or diabetic retinopathy after 14 weeks and the study was not designed to determine if there was a long-term benefit of the initial intravitreal treatment. The 56-week follow-up was collected for safety outcomes only.
This phase 3 randomized, multi-center clinical trial was conducted by the DRCR.net at 48 clinical sites in the United States. The study adhered to the tenets of the Declaration of Helsinki. The protocol and Health Insurance Portability and Accountability Act compliant informed consent forms were approved by multiple institutional review boards. Each study participant gave written informed consent to participate in the study following an informed consent process. Independent study oversight was provided by a data and safety monitoring committee. The study is listed on www.clinicaltrials.gov, under identifier NCT00445003 (website registration date March 6, 2007) and the protocol is available on the DRCR.net website (www.drcr.net, accessed October 1, 2010). Key aspects of the protocol pertinent to this report are summarized below.
Eligible study participants were at least 18 years old with type 1 or type 2 diabetes and without substantial renal disease or uncontrolled hypertension. The major eligibility criteria for a study eye included : (1) presence of severe NPDR or PDR, (2) presence of center-involved DME on clinical exam and central subfield thickness on time domain optical coherence tomography (OCT) (Stratus, Carl Zeiss Meditec, Dublin, CA) ≥250 μm, and (3) best-corrected Electronic-Early Treatment Diabetic Retinopathy Study (E-ETDRS Visual Acuity Test©19) visual acuity letter score ≥24 (approximate Snellen equivalent 20/320 or better). Principal exclusion criteria included: (1) prior PRP that was sufficiently extensive that the investigator did not believe that ≥ 1200 additional burns were needed or possible, (2) treatment for DME within the prior 4 months, (3) history of open-angle glaucoma or steroid-induced intraocular pressure (IOP) elevation that required IOP-lowering treatment, and (4) IOP ≥25 mmHg. At the onset of the study, a study participant could contribute only 1 study eye. However, during the course of the study the protocol was modified to allow participants to have 2 study eyes, provided both were eligible at the time of study entry, with random assignment to different treatments.
After eligibility was determined at the clinical center and informed consent was obtained, study participants with 1 study eye were assigned randomly on the DRCR.net website (using a permuted blocks design stratified by visual acuity and the number of sittings planned to complete the PRP) with equal probability to one of 3 treatment groups: (1) sham injection at baseline and 4 weeks, (2) intravitreal injections of 0.5-mg ranibizumab (Lucentis™, Genentech, Inc., South San Francisco, CA) at baseline and 4 weeks, or (3) intravitreal injection of 4-mg triamcinolone acetonide (Trivaris®, Allergan, Inc.) at baseline and sham injection at 4 weeks. Study participants with 2 study eyes were randomized with equal probability to receive one of the 3 treatment scenarios: sham in the eye with a greater visual acuity score and ranibizumab or triamcinolone in the eye with a lower visual acuity score OR ranibizumab or triamcinolone acetonide in the eye with a greater visual acuity score and sham in the eye with a lower visual acuity score (if both eyes had the same visual acuity letter score, the right eye was considered the eye with the greater visual acuity score).
The initial sham or intravitreal injection was given on the day of randomization. Focal/grid laser for DME was performed 3 to 10 days after the injection for all treatment groups. PRP could be initiated immediately after the focal/grid laser or on a subsequent day, but was to be initiated within 14 days of the baseline injection and fully completed within 49 days of randomization. Additional PRP was performed only if the size or amount of neovascularization increased following completion of the study-required PRP. Follow-up visits were conducted at 1, 4, 14 (primary outcome), 34, and 56 weeks after randomization. Eyes in the ranibizumab group received a second injection and eyes in the sham or triamcinolone groups received a sham injection at the 4-week visit. After the 14-week visit, additional treatment for DME and diabetic retinopathy could be given at investigator discretion as part of standard care (i.e. there were no requirements for repeated ranibizumab, triamcinolone or focal/grid laser treatments). Study participants were masked to treatment assignments. The visual acuity examiner and OCT technician at the primary outcome visit (14 weeks) were masked to treatment groups. All adverse events were recorded, irrespective of whether the event was considered treatment-related.
Sham and intravitreal injections were preceded by a povidone iodine prep of the conjunctiva. For a sham injection, the hub of a syringe (without a needle) was pressed against the conjunctival surface to simulate the force of an actual injection. Use of antibiotics in the pre-injection, peri-injection, or post-injection period was at investigator discretion.
The focal/grid laser technique was modified from the original ETDRS protocol as described previously and used in prior DRCR.net protocols.20, 21 PRP consisted of 1200 to 1600 burns given over 1 to 3 sittings, as detailed in the protocol (www.drcr.net, accessed October 1, 2010) with completion of the regimen within 49 days of randomization. To avoid bias from knowledge of treatment group assignment, the investigator declared prior to randomization the number of sittings planned to complete the PRP and the approximate number of burns planned for each sitting. Additional anesthesia in the form of retrobulbar, peribulbar or sub-Tenon’s injection could be used at investigator discretion. Slit lamp or an indirect laser delivery system could be used. Lasers with the capability of producing automated patterns (e.g. the PASCAL laser) could be used according to guidelines designed to create equivalent burn characteristics to standard laser. Before the administration of each required PRP sitting after the initial application of PRP, visual acuity was measured. If best-corrected visual acuity decreased from baseline by 10 or more letters (2 or more lines) and the investigator believed the decrease was due to exacerbation of macular edema, the investigator could choose to defer additional PRP for 2 weeks or longer until it was deemed that the risk of adding PRP no longer outweighed the benefits.
At baseline and at each follow-up visit except the 1-week visit, best-corrected visual acuity letter score was measured in the study eye at 3 meters by a certified tester using the E-ETDRS visual acuity test©.19 Visual acuity letter scores were measured at the 1-week visit using the baseline refraction. Following pupil dilation, OCT images were obtained at baseline and at each follow-up visit by a certified operator using a standardized protocol as done in a previous DRCR.net protocol on a time domain OCT. If the automated thickness measurements were judged by the Reading Center to be inaccurate on any submitted image, center point thickness was measured manually and this value was used to impute a value for the central subfield based on a correlation of the 2 measures of 0.98 as published previously. 17 At baseline, 29% of 362 baseline scans (2 were lost by the sites) had a central subfield thickness measurement imputed whereas the quality of 2 other scans did not permit automatic or manual grading of the central subfield thickness. During follow-up, 7% of 1,673 follow-up scans had a central subfield thickness measurement imputed and 8 (<1%) had compromised quality that precluded manual grading. Although an imputed thickness <250 μm does not necessarily mean that the true thickness measurement is <250 μm if it had been measureable, manual grading of the baseline scans resulted in an imputed baseline central subfield thickness of <250 μm for 49 eyes (14%). Of note, 12 (24%) of the 49 scans with imputed central subfield thickness <250 μm were from one clinical site, representing 63% of the 19 baseline scans from that site. All data except safety data are presented with exclusion of eyes from that clinical site (19 eyes from 14 subjects), although results were similar when evaluated with inclusion of eyes from that clinical site (data not shown). Results also were similar when evaluated with exclusion of all eyes with a baseline central subfield thickness <250 μm from any clinical site (data not shown). Baseline OCT images also were assessed by the University of Wisconsin-Madison Fundus Photographic Reading Center for cystoid abnormalities and subretinal fluid.
Standard ETDRS 7-field color stereoscopic fundus photographs were obtained at baseline, 14 weeks, and 56 weeks by a certified photographer and graded at the reading center using validated procedures.22 Additional testing included measurement of hemoglobin A1c and blood pressure at baseline and the following procedures at baseline and each follow-up visit: (1) slit lamp examination, (2) fundus examination following pupil dilation, and (3) measurement of IOP.
The primary outcome was the mean change in visual acuity from baseline to 14 weeks. A sample size of 364 eyes was planned to have 90% power to detect a difference in the change in the visual acuity letter score from baseline to 14 weeks in two 2-group comparisons (ranibizumab group versus sham group and triamcinolone group versus sham group) assuming a population difference of 6.0, standard deviation of 16, correlation between the baseline and 14-week visual acuities of 0.61, a type 1 error rate of 0.0245 (adjusted for multiple comparisons and for alpha spending in interim analyses to maintain an overall type 1 error rate of 0.05), and no more than 10% loss to follow up.
The primary analysis followed the intent-to-treat principle and included all randomized eyes with the exception of 19 eyes randomized from one clinical site where 63% of eyes had baseline imputed central subfield thickness <250 μm. Data were included in the 14-week analysis for all examinations performed between 70 and 153 days (10 to 22 weeks) from randomization. For the eyes without 14-week data, the last-observation-carried forward method was used to impute data for the primary analysis. Similar results were produced when analyses (1) used Monte Carlo Markov Chains23 to impute missing data, (2) included only eyes with a completed 14-week examination, (3) were performed with truncation of outlier values to be at most 3 standard deviations from the mean, and (4) were performed using ranks of the visual acuity letter scores (instead of the actual scores) transformed to have normal distributions using van der Waerden scores (data not shown).
For analyses other than the primary analysis, only data from completed visits were used with no imputation for missing data. For some results, medians and interquartile ranges have been reported instead of or in addition to means and standard deviations to describe the distribution of the data.
Two pairwise comparisons were made for all analyses. The alpha level was set to 0.02 for the primary outcome comparison using a Bonferroni adjustment for multiple comparisons to preserve the overall type I error rate ≤0.05, while adjusting for alpha spending of 0.01 for 2 interim data reviews (that used alpha spending functions f(t) = min (αt3, α) and f*(t) = min (αt2, α), respectively), and was set to 0.025 for all other outcome comparisons. For all continuous outcomes, the treatment group comparisons were made using analysis of covariance (ANCOVA) models with generalized estimating equations (GEE) to account for correlated data from subjects with 2 study eyes. The interactions between various subgroups and the treatment group were also tested. In view of the large number of factors evaluated, only interactions with P values <0.01 were considered unlikely to be due to chance. For binary outcomes, GEE was also used to account for correlated data from subjects with 2 study eyes, with differences in proportions between treatment groups estimated using binomial regression and relative risks estimated using Poisson regression with robust variance estimation.24 All analyses, unless otherwise specified, included adjustments for the 2 randomization stratification variables; baseline visual acuity and number of sittings to complete the PRP. In addition, models with central subfield thickness as the outcome included baseline central subfield thickness as a covariate and models with retinal volume as the outcome included baseline retinal volume as a covariate. All P values are 2-sided. SAS version 9.1 (SAS Institute, Cary, NC) was used for all analyses.
Between March 2007 and June 2009, 319 study participants (mean age 55±12 years; 40% women) were enrolled, 26 (8%) with 2 study eyes. The 345 study eyes with DME were randomly assigned to either the sham injection group (N=123), ranibizumab injection group (N=113), or triamcinolone injection group (N=109). At baseline, the mean visual acuity letter score in study eyes was 64±15 (approximately 20/50) and the mean OCT central subfield retinal thickness was 392±151 μm. Based on investigator assessment, severe NPDR was present in 18% of eyes and PDR in the other 82%. Based on reading center assessment, moderately severe NPDR or less severe retinopathy was present in 20% of eyes, severe NPDR was present in 5%, and PDR was present in the other 75%, including 35% with high risk PDR (level 71 or 75). The baseline characteristics of the 3 groups were similar (Table 1).
Visit completion rates are shown in Figure 1 (available at ###). Four (1%) study participants died prior to the 14-week primary outcome visit and 4 died subsequently; all from causes apparently unrelated to study treatment. For the remaining study participants, the 14-week visit was completed for 118 eyes (96%) in the sham group, 103 eyes (91%) in the ranibizumab group, and 105 eyes (96%) in the triamcinolone group. The 56-week visit was completed for 111 eyes (90%) in the sham group, 95 eyes (84%) in the ranibizumab group, and 93 eyes (85%) in the triamcinolone group.
Except for 1 eye in the triamcinolone group that did not receive an injection, all other eyes received the randomization-assigned sham or intravitreal injection at baseline. A second ranibizumab injection was given at the 4-week study visit in 108 eyes (96%) of the 113 eyes in the ranibizumab group (the 4-week visit was missed for 5 eyes). One eye in the triamcinolone group received triamcinolone at the 4-week visit when sham should have been given. Topical antibiotics appeared to be given less frequently before or after sham injections than before or after ranibizumab or triamcinolone injections (data not shown).
Focal/grid laser was performed in all eyes except 2 eyes (2%) and 3 eyes (3%) in the ranibizumab and triamcinolone groups, respectively. Of those eyes with focal/grid laser, 9 (7%), 7 (6%) and 10 (9%) were performed outside of the 3 to 10 day window from randomization in the sham, ranibizumab, and triamcinolone groups, respectively. Prior to the 14-week visit, no additional (alternative) treatment for DME was given.
PRP was not completed in 1 eye (1%), 2 eyes (2%), and 2 eyes (2%) in the sham, ranibizumab, and triamcinolone groups, respectively. Of those eyes with PRP, PRP was completed within 49 days of randomization in 108 eyes (89%), 97 eyes (87%), and 87 eyes (81%) in the sham, ranibizumab, and triamcinolone groups, respectively. PRP was completed in 1 sitting in 49 eyes (40%), 38 eyes (34%), and 41 eyes (38%) in the sham, ranibizumab, and triamcinolone groups, respectively. PRP characteristics did not differ appreciably by treatment group (Table 2).
After completion of the study-required full PRP, additional PRP prior to the 14-week visit was given in 1 eye in the ranibizumab group. There was little to no difference identified among the 3 groups with respect to additional PRP given after 14-weeks, including 21 eyes (19%) in the ranibizumab group (P = 0.39) and 24 eyes (23%) in the triamcinolone group (P=0.77) compared with the sham group (29 eyes [24%]).
Treatment for DME at or after the 14-week visit, which was part of standard care at investigator discretion, was administered to fewer eyes in the ranibizumab group (48 eyes [44%]; P = 0.04) and triamcinolone group (45 eyes [42%]; P = 0.004) compared with the sham group (71 eyes [59%]) (Table 3).
As shown in Table 4 and Figure 2, for the 14-week primary outcome, the mean change ±standard deviation (SD) in the visual acuity letter score from baseline was −4±14 in the sham group, +1±11 in the ranibizumab group (P<0.001 compared with the sham group), and +2±11 in the triamcinolone group (P<0.001 compared with the sham group). The results (Figure 3) reflected both a greater proportion of eyes with an improvement of ≥10 letters (19%, P=0.02 and 22%, P=0.002) and a lower proportion of eyes with a worsening of ≥10 letters (9%, P=0.004 and 10%, P=0.005) in the ranibizumab and triamcinolone groups, respectively, compared with the sham group (8% for ≥10 letter gain and 23% for ≥10 letter loss). Most of the overall improvement in visual acuity from baseline in the ranibizumab and triamcinolone treated groups occurred by the 4-week study visit (Figure 2).
There were no obvious clinically important differential treatment effects (interactions) at the 14-week primary outcome visit for any of the following subgroups: prior treatment for DME, baseline visual acuity, baseline OCT-measured central subfield thickness, baseline level of diabetic retinopathy, description of edema by the treating ophthalmologist as predominantly focal or predominantly diffuse, PRP in single sitting vs. multiple sittings, or baseline hemoglobin A1c level (Table 5). Fourteen-week primary outcome results were similar to the overall results when excluding all eyes with an imputed baseline central subfield thickness <250 μm (data not shown). There were only 26 study participants with 2 study eyes, precluding any adequate comparison of results in these eyes compared with study participants with only 1 study eye.
In general, treatment group comparisons for mean central subfield thickness paralleled the visual acuity comparisons. Specifically, mean central subfield thickness increased slightly from baseline to the 4-week visit in the sham group (+10 ± 97 μm) and decreased from baseline to the 4-week visit in both the ranibizumab group (−91 ± 161 μm) and triamcinolone group (−106 ± 132 μm) (Figure 4). At the 14-week primary outcome visit, the mean central subfield thickness change from baseline was similar to the 4-week change in the sham (+10 vs. −5 μm) and triamcinolone (−106 μm vs. −92 μm) groups, but the 4-week change in central subfield thickness in the ranibizumab group (−91 μm) apparently worsened 10 weeks after the 4 week injection (−39 μm). Nevertheless, the difference between sham and the 2 treatment groups remained statistically significant for both the ranibizumab (P = 0.01) and the triamcinolone (P < 0.001) groups (Table 6, Figures 4 and and5).5). More eyes in the sham group (44 eyes [38%]) at 14-weeks had an increase in central subfield thickness ≥ 10% with at least a 25 μm increase from baseline compared with those in the ranibizumab and triamcinolone groups (17 eyes [17%] and 10 eyes [10%], respectively). Of the 44 eyes in the sham group exhibiting a central subfield thickness increase ≥ 10% with at least a 25 μm increase, 15 (34%) had concordant decrease in visual acuity of ≥ 10 letters at 14 weeks, and represented approximately half of the eyes in this group with this amount of visual acuity loss. Similarly, of the 17 eyes in the ranibizumab group and the 10 eyes in the triamcinolone group with a central subfield thickness increase ≥10% and at least a 25 μm increase at 14 weeks, 2 eyes (12%) and 1 eye (10%), respectively, had concordant decreases in visual acuity of ≥10 letters at 14 weeks. OCT retinal volume results at the 14-week visit were similar to those of OCT central subfield thickness (Table 7, available at ###).
The study was not designed to evaluate effectiveness of either intravitreal ranibizumab or intravitreal triamcinolone on visual acuity or retinal thickening beyond the 14-week study visit. The originally randomly assigned treatments were specified only during the first 14 weeks with other treatments commonly observed thereafter, especially in the sham group. Evaluations at the 34- and 56-week study visit are provided only asfor the purposes of longer-term safety information. By the 56-week study visit (Table 8, Figure 2), the mean change ± SD in the visual acuity letter score from baseline was −6±17 in the sham group, −4±21 in the ranibizumab group (P = 0.44 compared with the sham group), and −5±16 in the triamcinolone group (P=0.63 compared with the sham group). By the 56-week study visit (Table 9, Figures 4 and and5),5), the mean change ± SD in the central subfield thickness from baseline was −71±156 in the sham group, −52±227 in the ranibizumab group (P = 0.25 compared with the sham group), and −40±138 in the triamcinolone group (P = 0.45 compared with the sham group). Mean change in volume from baseline paralleled the mean change in central subfield thickness from baseline to the 56-week study visit (data not shown).
Major ocular adverse events are summarized in Table 10. One (0.9%, 95% CI: 0.02% to 4.7%) of 116 study eyes (0.4% of 227 intravitreal injections) in the ranibizumab group had endophthalmitis with a visual acuity letter score (approximate Snellen equivalent) of 39 (20/160) at the 56-week visit. Three eyes (2%) in the sham group, 1 eye (<1%) in the ranibizumab group, and 1 eye (<1%) in the triamcinolone group had a tractional retinal detachment by the 14-week visit. Four eyes (3%) in the sham group, 5 eyes (5%) in the ranibizumab group, and 1 eye (<1%) in the triamcinolone group developed a tractional retinal detachment afterwards.
Up to the 14-week visit, 16 eyes (12%) in the sham group, compared with 6 eyes (5%) and 7 eyes (6%) in the ranibizumab and triamcinolone groups, respectively, had a vitreous hemorrhage. Among eyes with vitreous hemorrhage, 9 eyes (56%), 2 eyes (40%), and 1 eye (14%) worsened ≥10 letters from baseline to 14-week visit in the sham, ranibizumab, and triamcinolone groups, respectively. Vitrectomy for PDR (which included 5 eyes which developed a tractional retinal detachment) was performed by the 56-week visit in 18 eyes (15%) in the sham group, 8 eyes (7%) in the ranibizumab group, and 7 eyes (6%) in the triamcinolone group, including 1 eye (<1%) in the sham group, no eyes in the ranibizumab group, and 1 eye (<1%) in the triamcinolone group prior to the 14-week visit. The occurrence of adverse events related to IOP was more frequent in eyes in the triamcinolone group than in the ranibizumab or sham groups (Table 10).
There were no systemic adverse events with a difference in frequency among the 3 groups that could not be attributed to chance however, there were 4 (4%) cardiovascular or cerebrovascular events, as defined by the Antiplatelet Trialists’ Collaboration,25 in the sham group compared with 8 (7%) in the ranibizumab group (P=0.33) and 4 (3%) in the triamcinolone group (P=0.86). In the ranibizumab group, 1 event occurred approximately 5 weeks after randomization, between baseline and the second ranibizumab injection, 1 event occurred approximately 3 weeks after the 4 week injection, and the events for the remaining 6 study participants occurred more than 4 weeks after the 4-week injection (Table 11 and Figure 6). There were no differences in frequency of Antiplatelet Trialists’ Collaboration events between ranibizumab and sham or triamcinolone groups when stratified by prior history of cardiovascular events. All reported systemic adverse events and study eye ocular adverse events are shown in Tables 12 (available at ###) and Table 13 (available at ###).
In this randomized clinical trial of eyes receiving PRP and concurrent focal/grid laser for DME, 2 ranibizumab injections or 1 triamcinolone injection more likely led to statistically significant improvements in visual acuity from baseline to 14 weeks compared with 2 sham injections (Table 4, Figure 2, and Figure 3). The study was not designed to evaluate effectiveness of either intravitreal ranibizumab or intravitreal triamcinolone on visual acuity or retinal thickening beyond the 14-week study visit, so that evaluations at the 34- or 56-week study visit are provided only asfor the purpose of longer-term safety information. As expected, differences from the sham group did not persist through the 34- or 56-week follow-up visits (Table 8 and Figure 2). Changes in retinal thickness parallel the visual acuity outcomes (Tables 6, ,77 and and9,9, and Figure 4). Since the treating ophthalmologists in this study were unmasked to the treatment assignment, one important potential weakness of this study design includes the possibility of bias introduced if the ophthalmologist were to change the way she or he applied PRP (e.g., 1 vs. more than 1 sitting, more vs. fewer total spots, or timing relative to injection). However, a bias could not be detected with respect to these parameters, nor with respect to how focal/grid laser was applied (Table 2). Furthermore, prior to the 14-week visit, no additional (alternative) treatment for DME was given in any of the 3 treatment groups which could have biased the results (Table 3).
There was little change in the median visual acuity or OCT central subfield thickness from baseline to 14 weeks in the group of eyes receiving sham treatment in this study, however, these eyes did received focal/grid laser for DME involving the central macula at time of entry. In 3 previous studies17, 21, 26 by the DRCR.net in which focal/grid laser was applied to eyes with central DME and similar levels of visual acuity in the absence of concomitant PRP, the OCT central subfield thickness had favorably changed at 16 weeks by medians (25th, 75th percentiles) of −33 (−90, +13), −27 (−61, +13), and −34 (−101, +10) μm, respectively. The median (25th, 75th percentile) change in visual acuity was also favorable at +2 (−4, +7), +1 (−3, +6), and +2 (−3, +8), respectively (unpublished data). Thus, the control arm of this study indicates that eyes with central DME receiving prompt PRP at time of focal/grid laser for DME appear more likely to have increased macular edema and visual acuity loss in the short term than eyes without central DME receiving prompt PRP (in the absence of focal/grid laser)7 as well as eyes with central DME receiving focal/grid laser but not prompt PRP. Also consistent with a previous DRCR.net study, eyes in both the ranibizumab and triamcinolone groups appeared less likely to undergo additional PRP, develop vitreous hemorrhage, develop tractional retinal detachment, or undergo vitrectomy.17
Despite continued meticulous attention to use of a lid speculum and antiseptic to the injection site, there still is a risk of endophthalmitis when using intravitreal injections to treat diabetic retinopathy, as occurred in 1 eye in this study receiving ranibizumab. A previous DRCR.net study evaluating intravitreal anti-VEGF drugs in patients with diabetes, only some of whom had previously treated PDR, did not identify an increased risk of tractional retinal detachments, cerebrovascular accidents, or cardiovascular events.17 This study expands on the previously published findings in a cohort of study participants who underwent PRP, and again did not identify an increased risk of tractional retinal detachments, cerebrovascular accidents, or cardiovascular events following 1 (triamcinolone) or 2 (ranibizumab) intravitreal injections beyond that which could be attributed to chance alone. The safety of ranibizumab injections, continued for a longer period of time in persons with diabetes, remains largely unknown; although, there are several additional randomized clinical trials underway.
In a trial reported previously by the DRCR.net, which evaluated ranibizumab for center-involved DME when PRP was not performed, there were no differences in systemic adverse events that could not be attributed to chance alone when comparing ranibizumab and sham groups.17 There were more cardiovascular or cerebrovascular events as defined by the Antiplatelet Trialists’ Collaboration reported in the ranibizumab group compared with the sham group (Table 11) however, these occurred only once between the baseline injection and the second injection and once approximately 3 weeks after the 4-week study visit injection. The remaining events occurred at some point beyond 4 weeks after the final 4-week study visit injection (Figure 6) when it is assumed that all study drug was cleared from the study participant’s body.
As has been reported in prior studies,13, 14, 27 triamcinolone was associated in this study with an increased risk of elevated IOP. Unlike in prior studies, triamcinolone was not associated with a higher incidence of cataract surgery through 1 year. Whether this was because of the younger age of this cohort compared with prior studies, a lower enthusiasm to operate on cataracts in this cohort with more advanced levels of diabetic retinopathy, or other factors, is unknown.
In summary, the risk of short-term exacerbation of macular edema and associated visual acuity loss following PRP in eyes also receiving focal/grid laser for DME can be reduced by intravitreal triamcinolone or ranibizumab. These results are not maintained by the 34- or 56-week visit with discontinuation of intravitreal treatments after 1 (triamcinolone at baseline) or 2 (ranibizumab at baseline and 4-week) injections however, this study was not designed to evaluate effectiveness of either intravitreal ranibizumab or intravitreal triamcinolone on visual acuity or retinal thickening beyond the 14-week study visit. Evaluations at the 34- or 56-week study visit are provided only for longer-term safety information. The impact and clinical implications of continuing these treatments beyond 14 weeks cannot be determined from this study. While there likely is no increased risk of tractional retinal detachments or cerebrovascular accidents beyond that of chance alone for patients similar to those in this trial and receiving ranibizumab as given in this trial, the side effects of long term intravitreal ranibizumab or steroid use have to be balanced against potential benefits. the benefits and risks of long-term intravitreal ranibizumab or triamcinolone use in eyes receiving focal/grid laser for DME when also receiving PRP remains largely unknown.
Financial Support: 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
The most recently published list of the Diabetic Retinopathy Clinical Research Network investigators and staff who participated in this study can be found at www.drcr.net.
An address for reprints will not be provided.
Financial Conflicts of Interest: A complete list of all DRCR.net investigator financial disclosures can be found at www.drcr.net
Financial Disclosure: 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, nor in the collection, management, analysis, or interpretation of the data, or in the preparation of the manuscript. Genentech provided the ranibizumab for the study and Allergan, Inc. provided the triamcinolone for the study. In addition, Genentech and Allergan, Inc. provided funds to DRCR.net to defray the study’s clinical site costs. As described in the Diabetic Retinopathy Clinical Research Network (DRCR.net) Industry Collaboration Guidelines (available at 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. A complete list of all DRCR.net investigator financial disclosures can be found at www.drcr.net.
Financial disclosures: Alexander Brucker: Genentech (C, G). Dr. Brucker recused himself from his role as Editor-in-Chief of RETINA when this manuscript was submitted to RETINA and Dr. Brucker was not involved in the DRCR.net decision to submit the manuscript to RETINA. Lloyd P. Aiello: Genentech (O) Andrew Antoszyk: Genentech (C, G) Allergan (G) Susan S. Bressler: Genentech (G) Allergan (G) Dennis Marcus: Genentech (C, G, T, B) Neil Bressler: Grants to investigators at The Johns Hopkins University are negotiated and administered by the institution (such as the School of Medicine) which receives the grants, typically through the Office of Research Administration. Individual investigators who participate in the sponsored project(s) are not directly compensated by the sponsor, but may receive salary or other support from the institution to support their effort on the projects(s). Dr. Neil Bressler is Principal Investigator of grants at The Johns Hopkins University sponsored by the following entities (not including the National Institutes of Health): Allergan, Bausch & Lomb, Carl Zeiss Meditec, EMMES Corporation, Genentech, Lumenis, Notal Vision*, Novartis, QLT, Regeneron, Steba Biotech, Abbott Medical Optics, ForSight Labs, LLC and Genzyme Corporation. Dr. Susan Bressler’s consulting arrangement with Notal Vision in connection with Dr. Neil Bressler’s role as principal investigator on a Notal Vision-sponsored research grant has been reviewed and managed by The Johns Hopkins University School of Medicine in accordance with its conflict of interest policy.